Fig 2: Induced loss of endothelial Gas results in increased endothelial inflammation and atherosclerotic plaque formation.(A) Cross-sections of the inner curvatures of aortic arches of control or EC-Gas–KO mice stained with DAPI (blue) and antibodies against CCL2 or E-selectin (green) and CD31 (red). Bar diagrams show percentage of area stained by anti-CCL2 or anti–E-selectin antibodies of total endothelial cell area defined by anti-CD31 antibody staining (n = 6 animals per group; at least 3 sections were analyzed per animal). Scale bar: 25 µm. (B) Atherosclerosis-prone Ldlr-KO mice without (Ldlr-KO) or with endothelium-specific Gas deficiency (Ldlr-KO;EC-Gas–KO) were sham operated or underwent partial carotid artery ligation. Two weeks after the ligation, serial sections were made through the entire carotid arteries and stained with elastic stain (n = 6 mice per group). Scale bar: 100 µm. (C and D) Ldlr-KO or Ldlr-KO;EC-Gas–KO mice were fed a high-fat diet for 14 weeks. (C) En face view on whole aortae stained with Oil Red O. (D) Representative images of atherosclerotic plaques observed in brachiocephalic arteries. Scale bar: 50 µm. The bar diagrams show atherosclerotic lesion area as percentage of total aorta area (C) or total arterial area (D) (n = 8, Ldlr-KO; n = 9, Ldlr-KO;EC-Gas–KO [C]; n = 6 per genotype [D]). Data represent mean ± SEM; **P = 0.01, ***P = 0.001 (2-tailed Student’s t test).

Fig 3: Loss of endothelial ADM and its receptor CALCRL results in increased endothelial inflammation and atherosclerotic plaque formation.(A) Cross-sections of the inner curvatures of aortic arches of control and EC-CALCRL–KO mice stained with DAPI (blue) and antibodies against CCL2 (green), E-selectin (green), or CD31 (red). Bar diagrams show percentage of area stained by anti-CCL2 or anti–E-selectin antibodies of total endothelial cell area stained by anti-CD31 antibody (n = 6 mice per group). Scale bar: 25 µm. (B and C) Atherosclerosis-prone Ldlr-KO mice without (Ldlr-KO) or with induced endothelium-specific deficiency of CALCRL (Ldlr-KO;EC-CALCRL–KO) were fed a high-fat diet for 14 weeks. (B) Representative images of Oil Red O–stained brachiocephalic artery. Plaque area was quantified as percentage of total aortic area (n = 6 mice per group). (C) En face view on atherosclerotic lesions of aortae stained with Oil Red O (n = 7, Ldlr-KO; n = 8, Ldlr-KO;EC-CALCRL–KO). Scale bar: 200 µm. (D) Control or EC-Adm–KO mice were injected with AAV-PCSK9 or with AAV-Luc i.v. 1 week before partial carotid artery ligation. Two weeks after the ligation, serial sections were made through the entire carotid arteries and stained with elastic stain (n = 6, control; n = 8, EC-ADM KO). Scale bar: 100 µm. Data represent mean ± SEM; **P = 0.01, ***P = 0.001 (2-tailed Student’s t test).

Fig 4: ATF4 is an important mediator of ELAM-1 expression and IL-8 secretion in HTMCs. (A–C) HTMCs were exposed to media in the absence (Mock) or presence of 50 μM of TBHP for 12 h. Alternatively, cells were transfected with si-ATF4 or si-sc for 48 h, followed by incubation with 50 μM of TBHP for additional 12 h. (A) Representative images of subcellular expression of ELAM-1 as examined by indirect immunofluorescence (green, ELAM-1. blue, DAPI. scale bar, 40 μm). (B) Levels of cellular immunofluorescence of ELAM-1 were quantified and expressed as corrected total cell fluorescence (mean ± SEM, n = 3). (C) IL-8 secretion was assayed in culture supernatant of HTMCs using an IL-8 ELISA kit. Values were normalized for total protein at the respective treatment (mean ± SEM, n = 3). **P <0.01 vs. Mock, #P <0.05 and ##P <0.01 vs. TBHP+si-sc. TBHP+si-sc, exposure of cells transfected with si-sc to TBHP. TBHP+si-ATF4, exposure of cells transfected with si-ATF4 to TBHP. (D–F) A recombinant adenovirus coding ATF4 (Ad-GFP-ATF4) was constructed to express GFP protein as a marker for the identification of infected cells. Cultured HTMCs were infected with Ad-GFP-ATF4 or empty vector (Ad-GFP) for 72 h. (D) Western blot analysis for ATF4 and CHOP. Intensities of protein expression were quantified, normalized against the level of GAPDH and expressed as relative changes to Ad-GFP (mean ± SEM, n = 3). (E) Subcellular expression of ELAM-1 was detected by immunofluorescent staining (red, ELAM-1. blue, DAPI. green, GFP. scale bar, 20 μm). (F) IL-8 secretion was assayed using ELISA. Values were normalized for total protein at the respective treatment (mean ± SEM, n = 3). *P <0.05, **P <0.01 vs. Ad-GFP.

Fig 5: Up-regulation of ATF4, ELAM-1 and CHOP in the TM tissues of POAG patients. (A) Frozen sections of human TM tissues of POAG patients (n = 3) or normal controls (n = 4) were double stained with ATF4 (green) and ELAM-1 (red) antibodies. Note the partial colocalization of immunofluorescence in the TM of glaucomatous eyes. (B) Immunohistochemistry showing increased CHOP (green) expression in the TM of POAG patients. Arrows mark the TM. Blue: nuclear staining with DAPI. BF, bright field. Scale bar, 50 µm Supplementary Figure 2.