Fig 2: ERK1/2 signaling upregulates CMA1 to increase Ang II in UCP2 knockout endothelial cells. A,B) Gene Ontology (GO) analysis of biological processes related to the upregulated and downregulated genes in the UCP2 knockout endothelial cells. C) Schematic diagram of the common up‐regulated genes from RNA‐Seq of E15 and E18. D) Heat map of 10 differentially expressed genes from RNA‐Seq of E15 and E18. E) RT‐PCR analysis of the mRNA expression level of CMA1 in the UCP2WT and UCP2ECKO isolated brain endothelial cells. ***P < 0.001 (mean ± SEM, unpaired two‐tailed Student's t test, UCP2WT n = 5mice; UCP2ECKO n = 3 mice). F) Western blot analysis of the expression levels of CMA1, p‐ERK1/2, and ERK1/2. β‐actin was detected as loading control. G) Statistics of relative intensity of CMA1 showing the increased expression of CMA1 in UCP2ECKO isolated brain endothelial cells. **P < 0.01 (mean ± SEM, unpaired two‐tailed Student's t test, n = 3 mice). H) Statistics showing the increased ratio of p‐ERK1/2 and ERK1/2 in UCP2ECKO isolated brain endothelial cells. *P < 0.05 (mean ± SEM, unpaired two‐tailed Student's t test, UCP2WT n = 3 mice; UCP2ECKO n = 4 mice). I) ELISA analysis by collecting supernatant in isolated UCP2WT and UCP2ECKO mice brain ECs showing increased Ang II in endothelial UCP2 knockout mice. ***P < 0.001 (mean ± SEM, unpaired two‐tailed Student's t test, UCP2WT n = 6; UCP2ECKO n = 7 from 3 independent experiments). J) Confocal immunofluorescence image of ROS level labeled by DHE in UCP2WT and UCP2ECKO isolated brain primary endothelial cells. Scale bar, 100 µm. K) Quantification showing the increased ROS level in UCP2ECKO isolated brain primary endothelial cells. ***P < 0.001(mean ± SEM, unpaired two‐tailed Student's t test, UCP2WT n = 10; UCP2ECKO n = 15 from 3 independent experiments). L) Western blot analysis of the expression levels of CMA1, p‐ERK1/2 and ERK1/2 in bEnd.3 treated with the inhibitor of ERK1/2 SCH772984. β‐actin was detected as loading control. M,N) Statistics showing reduced the ratio of p‐ERK1/2 and ERK1/2 (M) and CMA1 (N) in bEnd.3 treated with SCH772984 compared with control. ***P < 0.001(mean ± SEM, unpaired two‐tailed Student's t test, n = 3 independent experiments (M); n = 4 independent experiments (N)). O) ELISA analysis showing that knockdown of CMA1 significantly restored Ang II increase due to endothelial UCP2 knockout. **P < 0.01 (mean ± SEM, one‐way ANOVA with Dunnett's multiple‐comparison correction, n = 5 each group from 3 independent experiments). Data are represented as means ± SEM. unpaired two‐tailed Student's t test; one‐way ANOVA with Dunnett's multiple‐comparison correction, n.s., not significant, *P < 0.05, **P < 0.01, ***P < 0.001.

Fig 3: Knockdown of endothelial UCP2 accelerates human NPCs differentiation toward astrocytes. A) The diagram of human CMEC/D3 and human NPCs coculture system. B) Schematic of human embryonic stem cells differentiate into human NPCs, followed by neurons and astrocytes. C) RT-PCR analysis showing the knockdown efficiency of human UCP2-shRNA at the mRNA level. ***P < 0.001(mean ± SEM, one-way ANOVA with Dunnett's multiple-comparison correction, n = 5 each group from 3 independent experiments). D) Confocal immunofluorescence image of TUJ1+neurons at 12 days of differentiation from human ESCs. Scale bar, 100 µm. E) Quantification of the percent of TUJ1+ cells showing decreased TUJ1+cells in human NPCs cocultured with UCP2-shRNA lentivirus-infected with hCMEC/D3 cells line. *P < 0.05 (mean ± SEM, unpaired two-tailed Student's t test, n = 7 each group from 3 independent experiments). F) Confocal immunofluorescence image of MAP2+neurons and S100ß + astrocytes at 20 d of differentiation from human ESCs. Scale bars, 20 µm. G,H) Quantification of the percent of MAP2+ neurons and S100ß + astrocytes showing decreased MAP2+ neurons and increased S100ß+ astrocytes in human NPCs cocultured with UCP2-shRNA lentivirus-infected with hCMEC/D3 cells line. **P < 0.001, ***P < 0.0001 (mean ± SEM, unpaired two-tailed Student's t test, Control n = 8; UCP2-shRNA n = 7 from 3 independent experiments). I) Western blot analysis of the expression levels of ALDH1L1, GFAP, TUJ1 in human NPCs cocultured with UCP2-shRNA lentivirus-infected with hCMEC/D3 cells line. ß-actin was detected as loading control. J,K) Statistics of relative intensity of ALDH1L1 and GFAP showing the increased expression of ALDH1L1 and GFAP in human NPCs cocultured with UCP2-shRNA lentivirus-infected with hCMEC/D3 cells line. *P < 0.05 (mean ± SEM, unpaired two-tailed Student's t test, ALDH1L1: Control n = 3; UCP2-shRNA n = 4; GFAP: n = 4 each group from 3 independent experiments). L) Statistics of relative intensity of TUJ1 showing the decreased expression of TUJ1 in human NPCs cocultured with UCP2-shRNA lentivirus-infected with hCMEC/D3 cells line. *P < 0.05 (mean ± SEM, unpaired two-tailed Student's t test, n = 3 independent experiments). Data are represented as means ± SEM. unpaired two-tailed Student's t test; *P < 0.05, ***P < 0.001.

Fig 4: Loss of endothelial UCP2 leads to persistent increased astrocyte production. A) Confocal immunofluorescence image of S100ß + cells in P0 UCP2WT and UCP2ECKO mice. Scale bar, 100 µm. B) Quantification of the number of S100ß + cells showing increased S100ß + cells in UCP2ECKO mice. **P < 0.01 (mean ± SEM, unpaired two-tailed Student's t test, n = 5 mice each group). C) Confocal immunofluorescence image of GFAP+ cells in P0 UCP2WT and UCP2ECKO mice. The down is higher magnification images. Scale bars, 100 µm (up), 50 µm (down). D) Quantification of the number of GFAP+ cells showing increased GFAP+ cells in UCP2ECKO mice. ***P < 0.001 (mean ± SEM, unpaired two-tailed Student's t test, n = 7 each group from 3 independent experiments). E) Western blot analysis of the expression levels of astrocyte marker GFAP and ALDH1L1. ß-actin was detected as loading control. F) Statistics of relative intensity of GFAP and ALDH1L1 showing increased the expression of GFAP+ and ALDH1L1+ cells in P2 UCP2ECKO mice. *P < 0.05 (mean ± SEM, unpaired two-tailed Student's t test, GFAP n = 3 mice each group; ALDH1L1 n = 4 mice each group). G) Confocal immunofluorescence image of GFAP+ cells in P8 UCP2WT and UCP2ECKO mice. A1, A2, B1, and B2 are higher magnification images. Scale bars, 200 µm (up), 50 µm (down). H) Quantification of the number of GFAP+ cells showing increased GFAP+ cells in UCP2ECKO mice. *P < 0.05 (mean ± SEM, unpaired two-tailed Student's t test, n = 5 each group from 3 independent experiments). I) Confocal immunofluorescence image of S100ß + cells in P8 UCP2WT and UCP2ECKO mice. The down is higher magnification images. Scale bars, 200 µm (up), 50 µm (down). J) Quantification of the number of S100ß + cells showing increased S100ß + cells in UCP2ECKO mice. **P < 0.01 (mean ± SEM, unpaired two-tailed Student's t test, UCP2WT n = 5 each group, UCP2ECKO n = 6 each group from 3 independent experiments). Data are represented as means ± SEM. unpaired two-tailed Student's t test; At least three biological replicates are shown. *P < 0.05. **P < 0.01, ***P < 0.001.

Fig 5: Endotheliocyte‐driven AngII‐gp130‐JAK‐STAT pathway regulates initiation of gliogenesis. A) Western blot analysis of the expression levels of TUJ1, GFAP, STAT3, and p‐STAT3 in primary NPCs treated with Ang II. β‐actin was detected as loading control. B) Statistics of the relative intensity of TUJ1 showing the decreased expression of TUJ1 in primary NPCs treated with Ang II. *P < 0.05 (mean ± SEM, unpaired two‐tailed Student's t test, n = 3 independent experiments). C) Statistics of relative intensity of ALDH1L1 showing the increased expression of ALDH1L1 in primary NPCs treated with Ang II. *P < 0.05 (mean ± SEM, unpaired two‐tailed Student's t test, n = 4 independent experiments). D) Immunoprecipitation (IP) experiment showing the interaction between Ang II and AGTR1A. E,F) Co‐immunoprecipitation (CO‐IP) assays for the interaction between AGTR1A and gp130. G) Confocal immunofluorescence image of FLAG and HA showing the colocalization of AGTR1A and gp130. Scale bar, 5 µm. H) Western blot analysis of the expression levels of p‐JAK1/2, STAT3, and p‐STAT3 in UCP2WT and UCP2ECKO cerebral cortex. β‐actin was detected as loading control. I) Statistics showing the increased expression levels of p‐JAK1/2 in UCP2ECKO cerebral cortex. *P < 0.05 (mean ± SEM, unpaired two‐tailed Student's t test, n = 3 mice each group). J) Statistics showing the increased ratio of p‐STAT3 and STAT3 in UCP2ECKO cerebral cortex. *P < 0.05 (mean ± SEM, unpaired two‐tailed Student's t test, n = 4 mice each group). K) Confocal immunofluorescence image showed that endothelial CMA1 knockdown can rescue the developmental imbalance caused by endothelial UCP2 deletion. Scale bar, 50 µm. L,M) Quantification of the percent of MAP2+ cells and GFAP+ cells. *P < 0.05, ***P < 0.001 (mean ± SEM, one‐way ANOVA with Dunnett's multiple‐comparison correction, n = 5 each group from 3 independent experiments). Data are represented as mean ± SEM. unpaired two‐tailed Student's t test; one‐way ANOVA with Dunnett's multiple‐comparison correction, *P < 0.05, ***P < 0.001.