The Hedgehog transcription factor Gli3 modulates angiogenesis

Abstract

Rationale: The Gli transcription factors are mediators of Hedgehog (Hh) signaling and have been shown to play critical roles during embryogenesis. Previously, we have demonstrated that the Hh pathway is reactivated by ischemia in adult mammals, and that this pathway can be stimulated for therapeutic benefit; however, the specific roles of the Gli transcription factors during ischemia-induced Hh signaling have not been elucidated.

Objective: To investigate the role of Gli3 in ischemic tissue repair.

Methods and results: Gli3-haploinsufficient (Gli3(+/-)) mice and their wild-type littermates were physiologically similar in the absence of ischemia; however, histological assessments of capillary density and echocardiographic measurements of left ventricular ejection fractions were reduced in Gli3(+/-) mice compared to wild-type mice after surgically induced myocardial infarction, and fibrosis was increased. Gli3-deficient mice also displayed reduced capillary density after induction of hindlimb ischemia and an impaired angiogenic response to vascular endothelial growth factor in the corneal angiogenesis model. In endothelial cells, adenovirus-mediated overexpression of Gli3 promoted migration (modified Boyden chamber), small interfering RNA-mediated downregulation of Gli3 delayed tube formation (Matrigel), and Western analyses identified increases in Akt phosphorylation, extracellular signal-regulated kinase (ERK)1/2 activation, and c-Fos expression; however, promoter-reporter assays indicated that Gli3 overexpression does not modulate Gli-dependent transcription. Furthermore, the induction of endothelial cell migration by Gli3 was dependent on Akt and ERK1/2 activation.

Conclusions: Collectively, these observations indicate that Gli3 contributes to vessel growth under both ischemic and nonischemic conditions and provide the first evidence that Gli3 regulates angiogenesis and endothelial cell activity in adult mammals.

Figure 1. Cardiac function, structural integrity, and vascularity were more compromised in Gli3^+/− mice than in WT mice after MI

MI was surgically induced in Gli3^+/− mice and their WT littermates. (A) Gli3 mRNA expression was evaluated by quantitative RT-PCR and normalized to 18S rRNA expression in areas of ischemic and non-ischemic myocardial tissue 1, 5, and 14 days after surgically induced MI. (B) Heart cross-sections were stained with anti-Gli3 antibodies (green) (left panel) or triple stained with anti-Gli3 antibodies (green), anti-CD31 antibodies (red) to identify endothelial cells, and DAPI (blue) to identify nuclei (right panel). (C) Heart cross sections were stained with anti-Gli3 antibodies (red) (left panel) or triple-stained with anti-Gli3 antibodies (red), anti-desmin antibodies (green) to identify cardiac muscle, and DAPI (blue) to identify nuclei (right panel). (D) LV ejection fractions and (E) systolic and diastolic LV volumes were evaluated by echocardiography 7, 14, and 28 days after MI in Gli3^+/− (n=8) and WT (n=9) mice. (F) Representative images of Masson trichrome–stained sections from the hearts of WT and Gli3^+/− mice harvested 28 days after MI; regions of fibrosis appear blue. (G) Fibrosis area was reported as the ratio of the length of fibrosis to the LV circumference. (H) Representative images of anti-CD31–stained sections from the hearts of WT and Gli3^+/− mice harvested 28 days after MI; ECs (i.e., CD31+ cells) appear reddish-brown. (I) Capillary density was quantified as the number of CD31+ vessels per HPF in the ischemic region and in the ischemic border zone. ** p≤0.01; * p≤0.05.

Figure 2. Angiogenesis is impaired in the ischemic hind limbs of Gli3^+/− mice

HLI was surgically induced in Gli3^+/− mice and their WT littermates. Seven and 14 days after HLI, the mRNA expression of (A) Gli3, (B) Gli2, and (C) Gli1, and of the Hh receptor (D) Ptch1 was evaluated by quantitative RT-PCR and normalized to 18S rRNA expression in muscle harvested from the non-ischemic (NI) and ischemic (I) hind limbs of WT and Gli3^+/− mice. (E) Representative images of anti-CD31–stained sections from the ischemic limb muscle of WT and Gli3^+/− mice harvested 14 days after HLI; ECs (i.e., CD31+ cells) appear reddish-brown. (F) Capillary density was quantified as the number of CD31+ vessels per HPF; n=8 WT and 9 Gli3^+/− mice on day 7, n=7 in each group on day 14, and n=5 WT and 7 Gli3^+/− mice on day 28. (G) Representative images of anti-smooth-muscle α-actin (anti-SMA)–stained sections from the ischemic limb muscle of WT and Gli3^+/− mice harvested 14 days after HLI; smooth muscle cells (i.e., SMA+ cells) appear reddish-brown. Smooth muscle–containing vessels were quantified as (H) the number of SMA+ vessels per HPF and (I) the proportion of vessels that contained smooth muscle. ***p≤0.001; **p≤0.01.

Figure 3. VEGF-induced angiogenesis is impaired in Gli3^+/− mice

Pellets containing either PBS or VEGF were implanted in the corneas of Gli3^+/− mice and their WT littermates. (A) VEGF-induced angiogenesis was evaluated 8 days later by injecting mice with 50 μL fluorescien-BS1-Lectin I 15 minutes before sacrifice, then viewing the corneas under fluorescence. (B) Angiogenesis was quantified as described previously. (C) Corneal cross sections from WT mice implanted with VEGF-containing pellets were stained with anti-Gli3 antibodies (red) and with antibodies to the EC-specific marker vWF (green); nuclei were stained with DAPI (blue).

Figure 4. Gli3 regulates function and gene expression in ECs

(A) HUVECs were transfected with Gli3 siRNA or control GFP siRNA, cultured for 48 hours, then seeded on Matrigel™. Tube formation was assessed under a phase-contrast microscope 8 and 24 hours after seeding. (B-G) HUVECs were transduced with adenoviral vectors coding for Gli3 expression (AdGli3) or with control adenoviral vectors coding for LacZ expression (AdLacZ). (B) Forty-eight hours after transduction, 5×10⁴ cells were seeded in the upper chamber of a modified Boyden chamber, and the lower chamber was filled with EBM-2 medium containing 1% fetal-bovine serum. Migration was quantified 8 hours after seeding by calculating the number of cells per HPF that had migrated to the lower chamber. The mRNA expression of (C) CXCL1, (D) CXCL2, (E) Ccl2, (F) IL-8, and (G) PD-ECGF was evaluated via quantitative RT-PCR and normalized to 18S rRNA expression. (H) Seven days after surgically induced hind-limb ischemia, PD-ECGF mRNA expression in muscle harvested from the non-ischemic and ischemic hind limbs of WT and Gli3^+/− mice was evaluated via quantitative RT-PCR and normalized to 18S rRNA expression; n=8 WT mice and 9 Gli3^+/− mice. **p≤0.01. (I) HUVECs were co-transfected with 3 plasmids: 1) a plasmid coding for Gli-regulated BS luciferase expression (pGli-BS) or mutant-Gli–regulated BS luciferase expression (pmGli-BS), 2) a plasmid coding for human Gli3 expression (phGli3) or a control pcDNA3 plasmid, and 3) a plasmid coding for LacZ expression. Forty-eight hours after transfection, luciferase activity in cell lysates was assayed and normalized to β-galactosidase activity.

Figure 5. Gli3 over-expression in ECs promotes Akt and ERK1/2 phosphorylation

(A) HUVECs were transduced with adenoviral vectors coding for Gli3 expression (AdGli3) or with control adenoviral vectors coding for LacZ expression (AdLacZ), and ERK1/2 phosphorylation (ERK1/2-P) was evaluated 48 hours later by Western blot. (B) HUVECs were transfected with Gli3 siRNA or control GFP siRNA; 48 hours after transfection, cells were treated with or without 50 ng/mL VEGF for 5 minutes, then ERK1/2 phosphorylation was evaluated by Western blot. (C-D) HUVECs were transduced with AdGli3 or AdLacZ; (C) 24 hours after transduction, cells were incubated with 10 μmol/L of the MAPK-ERK1/2 inhibitor U0126 or dimethyl sulfoxide (vehicle) for 24 hours, then c-Fos mRNA expression was evaluated via quantitative RT-PCR and normalized to 18S rRNA expression. (D) Forty-eight hours after transduction, Akt phosphorylation (Akt-P) was evaluated by Western blot. (E) HUVECs were co-transduced with AdLacZ or AdGli3 and AdLacZ or an adenoviral vector coding for a dominant-negative Akt mutant (DN-Akt). ERK1/2 phosphorylation was evaluated 48 hours later by Western blot. (F) HUVECs were transduced with AdGli3 or AdLacZ; 24 hours after transduction, cells were incubated with 10 μmol/L U0126 or dimethyl sulfoxide (vehicle) for 24 hours, then Akt phosphorylation was evaluated by Western blot. (G-I). AdLacZ- and AdGli3-transduced HUVECs were incubated with or without 10 μmol/L U0126 for 24 hours or co-transduced with DN-Akt. (G) 5×10⁴ cells were seeded in the upper chamber of a modified Boyden chamber, and the lower chamber was filled with EBM-2 medium containing 1% fetal-bovine serum. Migration was quantified 8 hours after seeding by calculating the number of cells per HPF that had migrated to the lower chamber. (H-I) The mRNA expression of (H) CXCL1 and (I) PD-ECGF was evaluated via quantitative RT-PCR and normalized to 18S rRNA expression. ***p≤0.001.