Fig 2: Disruption of VETC formation by inhibiting the Angpt2/Tie2 signaling relieves the immunosuppression in the tumor microenvironment.(A) Disruption of the Angpt2/Tie2 axis by shAngpt2 or Rebastinib disrupted VETC formation in Hepa1-6 allografts. Rebastinib, Tie2 inhibitor. (B) The levels of TGF-β1 in TECs. (C) The number of Tregs. (D) The ratio of Treg/CD8+ T cells. (E and F) The proportions of TNFRSF4+ Tregs (E) and Ki67+ Tregs (F). (G) The proportion of PD1+CD8+ T cells. (H) The number of CD8+ T cells. (I) The proportion of GZMB+ or GZMK+CD8+ T cells. For (A–I), left panels: shNC, Hepa-shNC allografts (negative control, n = 11); shAngpt2, Hepa-shAngpt2 allografts with stable silencing of Angpt2 (n = 6). Right panels: Ctrl, Hepa1-6 allografts in mice treated with control solution (n = 6); Rebastinib, Hepa1-6 allografts in mice treated with Rebastinib (n = 8). Data are shown as mean ± SEM. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, by Mann-Whitney U test (A, left panel; B; E right panel; F left panel; H left panel), 2-tailed Student’s t test (A, right panel; C; D; E, left panel; F, right panel; G; H, right panel; I).

Fig 3: Effects of anti‐VEGFA, ‐ANG‐2, or bispecific antibody on pathological neovascularization in the mouse OIR model. (A) Schematic representation of the experimental design to evaluate the effects of intraocular injected antibodies in a mouse OIR model. The antibodies were intravitreally injected into OIR mice at P14. Retinal whole‐mounts were prepared at P17 and P19 and immunostained with anti‐CD31 antibody. (B) The areas of neovascularization (red) and vaso‐obliteration (non‐perfusion, yellow) in the OIR retina treated with control IgG, anti‐VEGFA antibody, anti‐ANG‐2 antibody, or anti‐VEGFA/ANG‐2 bispecific antibody are shown. Scale bar = 1 mm. (C and D) The area (% of total area) of neovascularization and the vaso‐obliteration area in the OIR mouse retina. The averages of the area with standard errors are shown in the bar graphs. (E and F) The antibodies were intravitreally injected into OIR mice at P14, and the retina was harvested at P17 and frozen sectioned. Immunostaining with anti‐CD31 and ‐Ki67 antibodies was performed to examine endothelial cell proliferation. Scale bar = 50 μm. Nuclei were visualized with DAPI staining. Numbers of CD31 and Ki67 double‐positive cells inner to the INL were counted (F). Data are the averages of 8 samples with standard deviation. (C, D, F) Tukey's HSD test was used for the statistical analysis. *p < 0.05; **p < 0.01.

Fig 4: Changes of IBA1‐positive cells and the effects of intraocular injected anti‐VEGFA, ‐ANG‐2, or bispecific antibody. (A) Retinal whole‐mounts of the OIR mice at P17 and P19 treated with control IgG, anti‐VEGFA, ‐ANG‐2, or bispecific antibody were immunostained with anti‐IBA1 antibody and observed by a confocal microscope. Scale bar = 1 mm. (B) The IBA1 signal‐positive area was examined using ImageJ software, and the percentage of the positive area to the whole area is shown in (B). The averages of 3 independent samples with standard deviation are shown. *p < 0.05; **p < 0.01. (C) Alexa Fluor (AF) 488‐conjugated VEGFA or ANG‐2 was intravitreally injected into OIR mice at P17. The eyes were harvested one hour after the injection, and whole‐mounts were immunostained with anti‐CD31 and anti‐IBA1 antibodies and observed by a confocal microscope. White arrowheads indicate VEGFA‐ or ANG‐2‐Alexa Fluor 488 signals in IBA1‐positive cells. Scale bar = 20 μm.