Endothelial activating transcription factor 3 promotes angiogenesis and vascular repair in the mouse retina

Figure 1. ATF3 is localized at vascular endothelial cells (ECs) in the developing mouse retina (A) Retinal section staining of IB4 (green), ERG (magenta), and ATF3 (red) in WT mice at P5. Scale bars: 50 μm. (B) Schematic illustration of fluorescence-activated cell sorting (FACS) of mouse retinal ECs. (C) RT-qPCR analysis of mRNA in WT murine retinal non-EC (CD45-/CD31--) and EC (CD45-/CD31+) at P5. Error bars represent mean ± SEM. ∗p < 0.05. See also Figures S1 and S2.

Figure 2. Deficiency of ATF3 upregulated by VEGFA inhibits angiogenesis in vitro (A and B) RT-qPCR analysis of ATF3 mRNA in (A) HUVECs and (B) HRMECs stimulated with VEGFA (0, 10, 20, 40, and 80 ng/mL) for 3 h after serum starvation for 6 h. (C) HRMECs were stimulated with VEGFA (20 ng/mL) for 3 h after serum starvation for 6 h and stained with ATF3 (green) and DAPI (blue). Scale bars: 50 μm. (D) Tube formation assay on Matrigel using HRMECs transfected with negative control siRNAs (siNC) or ATF3 siRNAs (siATF3). The vascular density (left) and vascular length density (right) were measured using the ImageJ Vessel Analysis plugin. (E) Tube formation assay on Matrigel using ATF3-overexpressed (ATF3OE) HUVECs infected with lentiviral vectors of human ATF3. The vascular density (left) and vascular length density (right) were measured using the ImageJ Vessel Analysis plugin. Scale bars: 50 μm. Error bars represent mean ± SEM. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001. See also Figures S3 and S4.

Figure 3. Endothelial ATF3 is required for postnatal retinal angiogenesis in mice (A) Schematic illustration of tamoxifen administration for the generation of Cdh5-Cre Atf3fl/fl (Atf3iECKO) mice. (B) Retinal whole-mount staining of PECAM1 in Atf3fl/fl (control) and Atf3iECKO mice at P5. (C) Comparison of vascular progression lengths (control, n = 15 eyes; Atf3iECKO, n = 12 eyes). Scale bars: 500 μm. (D) Retinal whole-mount staining of IB4 (green), ESM1 (red), and ERG (white) in control and Atf3iECKO mice at P5. (E) Quantification of the proportion of the ESM1+ area relative to the ERG1 area in the vascular front (control, n = 7 eyes; Atf3iECKO, n = 8 eyes). (F) Retinal whole-mount staining of PECAM1 (green), ERG (blue), ki67 (red), and ERG (white) in control and Atf3iECKO mice at P5. Red color channel on the images was altered. (G) Number of ki67+/ERG+ cells/FOV in the vascular front (control, n = 10 eyes; Atf3iECKO, n = 13 eyes). Scale bars: 100 μm. Error bars represent mean ± SEM. ∗∗∗p < 0.001. See also Figure S5.

Figure 4. ATF3 expression is upregulated in endothelial cells of the OIR model (A) Schematic illustration of the mouse oxygen-induced retinopathy (OIR) model. (B) RT-qPCR analysis of ATF3 mRNA in normoxic (control) and OIR retinas at P17. Error bars represent mean ± SEM. ∗p < 0.05. (C) Retinal whole-mount staining of IB4 (green), ATF3 (red), and ERG (white) in control and OIR-WT mice at P12. (D) Retinal whole-mount staining of IB4 (green), ATF3 (red), and ERG (white) in control and OIR-WT mice at P17. (E) Retinal whole-mount staining of IB4 (green), ATF3 (red), and ERG (white) in control and OIR WT mice at P21. Scale bars: 100 μm. See also Figures S6 and S7.

Figure 5. Endothelial ATF3 deletion inhibits vascular remodeling of OIR retinas (A) Schematic illustration of the mouse OIR model and tamoxifen administration. (B–D) Retinal whole-mount staining of IB4 in Atf3fl/fl (control) and Cdh5-Cre Atf3fl/fl (Atf3iECKO) OIR mice at (B) P12, (C) P17, and (D) P21. (E–I) Quantification of the avascular area at (E) P12, (F) P17, and (G) P21, and neovascular tuft (NVT) areas at (H) P17 and (I) P21 (P12: control, n = 12 eyes; Atf3iECKO, n = 13 eyes) (P17: control, n = 7 eyes; Atf3iECKO, n = 8 eyes) (P21: control, n = 10 eyes; Atf3iECKO, n = 11 eyes). Scale bars: 500 μm. Error bars represent mean ± SEM. ∗∗p < 0.01.

Figure 6. scRNA-seq of retinal vascular ECs from the OIR mice (A) Schematic illustration of the scRNA-seq mouse OIR model. (B) Uniform Manifold Projection (UMAP) of CD45-/CD31+ cells from mouse retinas in OIR at P17. (B′) UMAP colored for expression of ATF3. (C) Gene Ontology (GO) analysis of highly expressed genes in ATF3-positive ECs.

Figure 7. RNA-seq analysis of HRMECs transfected with negative control or ATF3 siRNAs (A) Principal component analysis of HRMECs transfected with Silencer Select negative control siRNA (siNC) or ATF3 siRNA (siATF3). PC1, the first principal component explains 79% of the variance; PC2, the second principal component explains 9% of the variance. n = 4 replicates. (B) Heatmap highlighting differentially expressed top 20 genes between ATF3 KD HRMECs and control HRMECs. (C) Volcano plot highlighting differentially expressed top 20 genes in ATF3 KD HRMECs compared with control HRMECs. (D) Protein–protein interaction network of factors involved in angiogenesis that are commonly downregulated in ATF3-negative ECs from OIR mice and ATF3KD HRMECs. Red-colored factors are related to the VEGFA-VEGFR2 signaling pathway.

Figure 8. Expression of angiogenesis-related factors in ATF3 knockdown or -overexpressed ECs (A) RT-qPCR analysis of angiogenesis-related factors in HRMECs transfected with negative control siRNAs (siNC) or ATF3 siRNAs (siATF3). (B) RT-qPCR analysis of angiogenesis-related factors in control or ATF3-overexpressed (ATF3OE) HUVECs infected with lentiviral vectors of human ATF3. Error bars represent mean ± SEM. ∗∗p < 0.01; ∗∗∗p < 0.001.