Fig 2: Requirement of caspase-3 for Wnt2 activation and neurogenesis in ischemic cortex.a Combination of TUNEL staining and Wnt2 immunostaining (left panels), and Western-blotting of Wnt2 in the ischemic cortex of WT and caspase-3-/- mice (right panels). Expression of Wnt2 in the injured WT cortex is significantly increased, but remains unchanged in caspase-3-/- cortex. Dashed lines indicate lesion border. Bar = 30 µm. **P = 0.0061 (left panel), **P = 0.0019 (right panel). One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 3 mice per group. b Representative images of Nestin/ß-gal double immunostaining (left panels), and Western-blotting of ß-gal in the ischemic cortex of Topgal mice with or without caspase-3 mutation at 5 dpi (right panels). The expression of ß-gal in the ischemic cortex of Topgal:caspase-3+/+ mice is greatly enhanced, whereas it shows no notable change in Topgal:caspase-3-/- mice. Bar = 30 µm. **P = 0.0096 (left panel), **P = 0.0015 (right panel). One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 3–4 mice per group. c Double staining of BrdU with Nestin, and Western-blotting of Nestin in the ischemic WT and caspase-3-/- cortex. Notice the lower level of Nestin in the ischemic caspase-3-/- cortex. **P = 0.001 (left panel), **P = 0.0012 (right panel). One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 3–4 mice per group. d Double staining and quantification of BrdU/DCX. The numbers of BrdU/DCX-positive cells is significantly decreased in the caspase-3-/- cortex. Bars = 30 µm. **P = 0.0072 (left panel), **P = 0.0021 (right panel). Two-tailed Student’s t test. N = 3–4 per group. Arrows in (b, d) point to double-positive cells and arrowheads in (d) point to magnified double-positive cells. Mean ± standard error (SE). Asterisks with bars connecting two groups indicate difference between these two groups.

Fig 3: Effects of Wnt2 knockdown on astrocyte dedifferentiation and cortical neurogenesis.a Representative images of Nestin/GFAP double-immunostaining in the bilateral cortice with two ischemic injures infected by lentivirus expressing control shRNA (right) and Wnt2 shRNA (left), respectively. Bars = 200 µm in upper panel. Bar = 100 µm in enlarged view. b Western-blotting of Nestin and Sox2 in Wnt2 shRNA treated mice at 7 dpi. Wnt2 shRNA dramatically suppressed the expression of Nestin and Sox2 without obviously affecting the expression of GFAP. N = 5 mice per group. c Double-immunostaining and quantification of DCX/BrdU in ischemic cortices treated by control shRNA or Wnt2 shRNA. Arrowheads point to double-positive cells. Arrows point to magnified cells. Bar = 20 µm. **P = 0.0087. Two-tailed Student’s t test. N = 4 mice per group. d–f Forelimb activity assay of mice treated with control shRNA or Wnt2 shRNA. From 2 weeks on, the sliding scores and the asymmetric index of forelimb activity in Wnt2 shRNA treated group were much higher than control group, indicating worse functional recovery. N = 5 mice per group. P < 0.0001 (2w, 3w, 4w and 5w). Two-way RM ANOVA. a–c, asterisks with bars connecting two groups indicate difference between these two groups. e, f, P values indicate comparison of experimental group with control group of same time point. Mean ± standard error (SE).

Fig 4: Schematic graph of the major finding.Following cerebral ischemia, apoptotic neurons up-regulate Wnt2 protein via IRES mediated protein translation. Released Wnt2 activates Wnt/ß-catenin signaling in reactive astrocyte and triggers astrocytic dedifferentiation, which facilitated cortical neurogenesis.

Fig 5: Up-regulation of Wnt2 by apoptotic neurons and its involvement in Wnt//ß-catenin signaling activation.a Double-immunostaining of NeuN/Wnt2 in normal cortex. Bar = 50 µm. b, c Western-blotting of Wnt2 at different time points in ischemic cortex. Notice that Wnt2 protein increases quickly after ischemia. *P = 0.026 (1 h), 0.028 (2 h), 0.024 (4 h), 0.048 (8 h), 0.043 (12 h). One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 4 mice per group. d Combination of TUNEL staining and Wnt2 immunostaining at 24 hpi. Wnt2high immunoreactivity overlapped well with TUNEL-staining (arrowheads). Dashed line indicates lesion border and asterisk indicates lesion area. Bar = 30 µm. e Representative image of TUNEL/Wnt2/NeuN-triple immunostaining in the lesion area. Bar = 10 µm. f, g Western-blotting of Wnt2 in conditioned medium of normal cultured (NNCM), apoptotic (ANCM), Wnt2 shRNA pretreated apoptotic (shRNA ANCM) wild type and caspase-3-/- neurons. Apoptotic neurons increase Wnt2 release into the culture medium, and this increase is compromised by caspase-3 ablation and Wnt2 shRNA. BSA was used as internal control. *PNNCM vs ANCM = 0.012, *PANCM vs ShRNA-ANCM = 0.014. One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 3 batches of cells. h Western-blotting of active ß-catenin in the cytoplasmic and nuclear protein of astrocytes treated by NNCM, ANCM, AICM or shRNA ANCM. ANCM greatly increases the level of ß-catenin in nuclear protein. LMNB1 was used as nuclear protein control. i Double-immunostaining of GFAP/ß-gal in Topgal mice pretreated with Wnt2 shRNA. Arrows point to representative double-positive cells. Bar = 50 µm. j Western-blotting of ß-gal and Axin2 in ischemic cortex of Topgal mice treated with Wnt2 shRNA or control. Left: *Pipsi vs ipsi+Wnt2-shRNA = 0.038, *Pipsi+Con-shRNA vs ipsi+Wnt2-shRNA = 0.041, Right: *Pipsi vs ipsi+Wnt2-shRNA = 0.013, *Pipsi+Con-shRNA vs ipsi+Wnt2-shRNA = 0.016. One-way ANOVA followed by Bonferroni’s post hoc comparisons tests. N = 3 mice per group. Notice the less induction of ß-gal and Axin2 in Wnt2 shRNA pretreated cortex. Mean ± standard error (SE). Asterisks indicate the comparison with control group. Asterisks with bars connecting two groups indicate difference between these two groups.