Fig 2: A model for Caveolae-mediated Tie2 signaling in CCM progression.CCM3 in brain ECs normally suppresses caveolae by yet unknown mechanism. CCM3 loss in ECs causes enhanced caveolae-mediated Tie2 surface expression and endocytosis, leading to EC junctional disruption, EC-pericyte dissociation, BBB dysfunction, and lumen enlargement. Non-receptor kinase Abl cooperatively regulates Tie2 expression and activity. Cav1 deficiency or Tie2 inhibition, like neutralization of Angpt2 ligand, ameliorates CCM lesion progression in the mouse CCM models, with normalization of EC-pericyte interaction and vascular integrity.

Fig 3: Dual inhibition of VEGF-A and ANG-2 reduced the number of TUNEL-positive cells and improved visual functionality, as revealed by ERG, in a model of sCNV in JR5558 mice ABar graph of TUNEL-positive cells in the retina of mice treated with 5 mg/kg anti-VEGF-A, anti-ANG-2, and 3, 5, and 10 mg/kg anti-VEGF-A/ANG-2. Error bars show SEM with n = 6 animals per group and * denotes all significant changes after one-sided ANOVA (P = 0.0041) and Tukey's multiple t-test. IgG control is significantly different from anti-VEGF-A/ANG-2 mid (**, P = 0.0077) and high (**, P = 0.0028).B–G(B) IgG control, (C) anti-VEGF-A, (D) anti-ANG-2, (E–G) anti-VEGF-A/ANG-2 low, mid, and high: representative micrographs of whole-mount retinae preparations display the reduction in TUNEL-positive photoreceptor cells (green) clustered around focal neovascular lesions in anti-VEGF-A/ANG-2 (E–G) and, to a lesser extent, anti-VEGF-A-treated (C) and anti-ANG-2-treated (D) groups compared with IgG-treated controls (B) (scale bar, 200 µm).HERG was used to assess retinal function in response to large-field flash stimuli under both scotopic and photopic conditions. Representative photopic (light-adapted) and scotopic (dark-adapted) flash response series show increased amplitude and reduced latency after treatment with bispecific anti-VEGF-A/ANG-2 compared with IgG control (both 3 mg/kg). ERG shows reduced depression of B-wave responses in JR5558 mice treated with anti-VEGF-A/ANG-2 compared to IgG control.IMaximum photopic ERG amplitude was increased using anti-VEGF-A/ANG-2 compared to single anti-VEGF-A or anti-ANG-2 treatments (all at 3 mg/kg). There was a significant increase in amplitude at stimuli > 2 log (cd*s/m2) between the bispecific anti-VEGF-A/ANG-2 and IgG control group under scoptopic and photopic conditions. Asterisk (*) denotes significance after ANOVA and Dunnett's multiple t-test compared to IgG control compared for each stimuli separately; error bars show SEM with n = 8 animals per group. Under photopic conditions, anti-VEGF-A/ANG-2 reached significance at 2.58 and 2.30 cd*s/m2 with *, P = 0.0498 and *, P = 0.0479, respectively. Under scotopic conditions, all stimuli reached significance for anti-VEGF-A/ANG-2 at -2.3 cd*s/m2 (*, P = 0.0498); -1.93 cd*s/m2 (**, P = 0.0035); -1.63 and -1.33 cd*s/m2 (****, P < 0.0001); -1.03 cd*s/m2 (***, P = 0.0002); -0.73 cd*s/m2, (**, P = 0.004); -0.43 cd*s/m2, (**, P = 0.0024); -0.13 cd*s/m2 (**, P = 0.0024) and 0.18 cd*s/m2 (**, P = 0.0031). In addition, anti-ANG-2 treatment reached significance at -0.13 cd*s/m2 (+ P = 0.011) and 0.18 cd*s/m2 (+ P = 0.005), respectively.Data information: ERG, electroretinogram; sCNV spontaneous choroidal neovascularization.

Fig 4: Inhibition of vessel leakiness and lesion number by combined inhibition of VEGF-A and ANG-2 in the model of sCNV in JR5558 mice using fluorescence angiography A, BReal-time qPCR analysis of ANG-2 levels in retina and RPE/choroid complexes of JR5558 and C57BL/6J (C57) mice. (A) Relative expression levels of ANG-2 were significantly increased in the RPE/choroid complexes of JR5558 mice in comparison with C57, at 50 and 62 days old (*, P = 0.022 at D50 and *, P = 0.042 at D62). (B) By contrast, retina levels of ANG-2 were not significantly different between C57 and JR5558 mice, indicating neovascularization is driven locally by ANG-2. Asterisk (*) denotes statistical significance of JR5558 mice compared to wild-type C57BL/6 using unpaired t-test for each time point analyzed separately.CSchematic presentation of experimental design. Mice received IP injections of CrossMAb anti-VEGF-A/ANG-2 (species crossreactive, B20-4.1 and LC10), anti-ANG-2 (LC10), an anti-VEGF-A IgG1 (B20-4.1), or IgG control at postnatal D14 and D19 followed by fluorescence angiography at D26 (early intervention), results demonstrated in (D–K). Alternatively, mice received antibody at postnatal D47 and D55 followed by fluorescence angiography at D60 (late intervention), results demonstrated in (L) and (M).D–IRepresentative examples of fluorescence angiograms of IgG control (D), anti-VEGF-A (E), anti-ANG-2 (F) (all at 5 mg/kg), and three doses of anti-VEGF-A/ANG-2 (G, H, and I) (at 3, 5, and 10 mg/kg).J, KBar graph of numbers of spontaneously occurring lesions (J) and area by fluorescence angiography (K) after two weekly doses of antibody (antibodies at 5 mg/kg IP and anti-VEGF-A/ANG-2 at 10 [high], 5 [mid], and 3 [low] mg/kg IP) followed by analysis a week later in the early intervention model.L, MBar graph of numbers of spontaneously occurring lesions (L) and area by fluorescence angiography (M) after two weekly doses of antibody (3 mg/kg IP) followed by analysis a week later in the late intervention model.Data information: SEM is shown as error bars with n = 4 (A, B) or n = 8 (J–M) animals per group and significance indicated by * using ANOVA (J, K: P < 0.0001; L: P < 0.0018; M: P < 0.031 followed by Tukey's multiple t-test in J–M). In (J), IgG control is significant against anti-VEGF-A (****, P < 0.0001), anti-ANG-2 (**, P = 0.0069), anti-VEGF-A/ANG-2 low (*, P = 0.0194), mid and high (****, P < 0.0001). Furthermore, anti-VEGF-A/ANG-2 high is significant against anti-VEGF-A (*, P = 0.0428), anti-ANG-2 (****, P < 0.0001), and anti-VEGF-A/ANG-2 low (****, P < 0.0001). Finally, anti-ANG-2 is significantly different from anti-VEGF-A/ANG-2 mid (**, P < 0.0041). In (K), IgG control is significant against anti-VEGF-A (***, P = 0.0003) and anti-VEGF-A/ANG-2 mid and high (both ****, P < 0.0001). Furthermore, anti-VEGF-A/ANG-2 high is significantly different from anti-VEGF-A (**, P = 0.0037), anti-ANG-2 (****, P < 0.0001), and anti-VEGF-A/ANG-2 low (***, P = 0.0001). Finally, anti-ANG-2 is significantly different from anti-VEGF-A/ANG-2 mid (**, P < 0.0022) and anti-VEGF-A/ANG-2 low against mid (*, P = 0.022). In (L), IgG control is significantly different from anti-ANG-2 (*, P = 0.023) and anti-VEGF-A/ANG-2 (*, P = 0.014). Vehicle is different from anti-ANG-2 (*, P = 0.024) and anti-VEGF-A/ANG-2 (*, P = 0.016). In (M), IgG control is significantly different from anti-ANG-2 (*, P = 0.044) and anti-VEGF-A/ANG-2 (*, P = 0.049). D, day; IP, intraperitoneal; sCNV, spontaneous choroidal neovascularization.

Fig 5: Vitreous concentrations of angiopoietins in patients newly diagnosed with retinal diseases and cell model of barrier breakdown testing the interaction of VEGF-A and angiopoietins A, BBox plots of vitreal ANG-1 (A) and ANG-2 (B) levels from newly diagnosed patients with wAMD, DR, proliferative DR and RVO compared to controls (macular hole). The interquartile range of the data is indicated by the box. A nonparametric Kruskal–Wallis analysis followed by Dunn's method for multiple comparisons was used to show significant differences of the groups to control which are indicated by asterisks. ANG-1 levels did not differ significantly, but ANG-2 levels were significantly different: control vs. AMD (*, P = 0.0451), vs. DR (****, P < 0.0001), pDR (****, P < 0.0001), and RVO (****, P < 0.0001). CSchematic of cellular experiments to measure transendothelial resistance in human endothelial cells using transwell filters and CellZscope technology.D24-h concentration of ANG-2 in supernatants of the basal side of the culture stimulated with 5 ng/ml of VEGF-A, 200 ng/ml of ANG-1, or the combination. Error bars show SEM with one-sided ANOVA (P < 0.0001) and Tukey's multiple t-test for five independent experiments indicating significance for control vs. VEGF-A (**, P = 0.0028); VEGF-A vs. ANG-1 (****, P < 0.0001); ANG-1 vs. VEGF-A and ANG1 (**, P = 0.0061); VEGF-A vs. VEGF-A and ANG-1 (ns, P = 0.0625).EHuman endothelial cells plated on filters were assessed for endothelial barrier function over time after the addition of 5 ng/ml VEGF-A alone or in combination with 200 ng/ml ANG-1 or 10 µg/ml anti-ANG-2 or all three test items. The final time point was used for statistical analysis. Error bars show SEM with one-sided ANOVA (P < 0.0001) and Tukey's multiple t-test for three independent experiments indicating significance of VEGF-A vs. untreated (****, P < 0.0001); vs. VEGF-A and ANG-1 and anti-ANG-2 (***, P = 0.0004); vs. VEGF-A and anti-ANG-2 (*, P = 0.038); vs. VEGF-A and ANG-1 (*, P = 0.040). Furthermore, VEGF-A and anti-ANG-2 and ANG-1 are significantly different vs. VEGF-A and ANG-1 (*, P = 0.0451); vs. VEGF-A and anti-ANG-2 (*, P = 0.0474); vs. untreated (*, P = 0.0492). Finally untreated is significantly different vs. VEGF-A and ANG-1 (***, P = 0.0009) and VEGF-A and anti-ANG-2 (***, P = 0.0008).Data information: ANOVA, analysis of variance; SEM, standard error of the mean; ANG-2, angiopoietin-2; DR, diabetic retinopathy; RVO, retinal vein occlusion; wAMD, wet age-related macular degeneration; pDR, proliferative diabetic retinopathy; vs., versus.