Loss of PPARγ in endothelial cells leads to impaired angiogenesis

Abstract

Tie2-promoter-mediated loss of peroxisome proliferator-activated receptor gamma (PPARγ, also known as PPARG) in mice leads to osteopetrosis and pulmonary arterial hypertension. Vascular disease is associated with loss of PPARγ in pulmonary microvascular endothelial cells (PMVEC); we evaluated the role of PPARγ in PMVEC functions, such as angiogenesis and migration. The role of PPARγ in angiogenesis was evaluated in Tie2CrePPARγ(flox/flox) and wild-type mice, and in mouse and human PMVECs. RNA sequencing and bioinformatic approaches were utilized to reveal angiogenesis-associated targets for PPARγ. Tie2CrePPARγ(flox/flox) mice showed an impaired angiogenic capacity. Analysis of endothelial progenitor-like cells using bone marrow transplantation combined with evaluation of isolated PMVECs revealed that loss of PPARγ attenuates the migration and angiogenic capacity of mature PMVECs. PPARγ-deficient human PMVECs showed a similar migration defect in culture. Bioinformatic and experimental analyses newly revealed E2F1 as a target of PPARγ in the regulation of PMVEC migration. Disruption of the PPARγ-E2F1 axis was associated with a dysregulated Wnt pathway related to the GSK3B interacting protein (GSKIP). In conclusion, PPARγ plays an important role in sustaining angiogenic potential in mature PMVECs through E2F1-mediated gene regulation.

Keywords: Angiogenesis; E2F1; Endothelial cell; GSKIP; Osteopetrosis; PPARγ; Pulmonary hypertension; Wnt signaling.

Fig. 1.

Loss of PPARγ attenuates angiogenesis and impairs EPC-like cell mobilization from the bone marrow. (A) In vivo angiogenesis assay with subcutaneously placed matrigel plugs in wild-type (WT) and TIE2CrePPARγ^flox/flox (KO) mice. Arrows indicate blood vessels in matrigel plugs stimulated with vehicle (H₂O; Con) or BMP2 (10 ng/ml). Scale bar: 25 mm. (B) Number of vessels per field (20× magnification) was used for quantifying vessels. ‘C’ refers to control conditions. (C–E) Percentage of CD34⁺/VEGFR2⁺ cells in gated live cell population from blood (C), and CD34⁺/VEGFR2⁺/CD45⁻ live cell population from (D) spleen and (E) bone marrow of WT and KO mice was analyzed using flow cytometry. (F–H) Cross-transplantation of bone marrow between KO and WT mice rescued the cell mobilization defect of the CD34⁺/VEGFR2⁺ live cell population in samples from blood (F) and CD34⁺/VEGFR2⁺/CD45⁻ cells in live cell population in samples from (G) spleen and (H) bone marrow in KO mice. (I) In vivo angiogenesis in WT and KO mice was analyzed by matrigel plug assay after bone marrow (BM) transplantation. The angiogenic defect in KO mice was not rescued after bone marrow transplantation from WT mice. Error bars represent mean±s.e.m. from six matrigel plugs from three separate mice in B and I and from 3–5 mice per group in C–H. *P<0.05, **P<0.01, ***P<0.001 versus respective control; one-way ANOVA with Bonferroni's multiple comparison tests in B,I; unpaired two-tailed t-test in C–H.

Fig. 2.

Loss of PPARγ in mature endothelial cells attenuates the angiogenic response. (A) In vitro tube-formation with isolated PMVEC from wild-type (WT) and TIE2CrePPARγ^flox/flox (KO) mice. Fluorescence microscopy was used to detect calcein-incorporated live cells and their tube structures after 16 h on matrigel, when treated with vehicle (H₂O; Con) and BMP2 (10 ng/ml). (B) Tube formation was quantified by tube area per total field area (10× magnification) from three different fields per experiment. (C) Migration assay with PMVECs isolated from WT and KO mice and stimulated with vehicle (H₂O; Con), VEGF (50 ng/ml) and BMP2 (10 ng/ml) for 6 h. (D) Migration assay with Non-target (‘C’) and PPARγ (‘P’) siRNA-transfected human PMVECs stimulated with vehicle (H₂O; Con) and VEGF (50 ng/ml). The cells were counted from three different fields (20× magnification) at the center of each well to quantify the number of migrated cells in C and D. Error bars represent mean±s.e.m. from four separate experiments in B and from three separate experiments in C,D. **P<0.01, ***P<0.001, **** P<0.0001 versus respective control; one-way ANOVA with Bonferroni's multiple comparison tests.

Fig. 3.

Apelin can partially restore angiogenic defect both in vivo and in vitro after loss of PPARγ. (A) In vivo angiogenesis capacity was analyzed by matrigel plug assay. Matrigel plugs were placed subcutaneously into wild-type (WT) and TIE2CrePPARγ^flox/flox (KO) mice. The number of blood vessels from matrigel plugs treated with vehicle (H₂O; Con), BMP2 (10 ng/ml), apelin (100 nM), and combination of BMP2 and apelin was counted. (B) Migration assay with human PMVEC treated with either Non-target siRNAs (Con) or with PPARγ-siRNA. Cells were stimulated with vehicle (H₂O; Con), VEGF (50 ng/ml), apelin (100 nM), and combination of VEGF and apelin. Error bars represent mean±s.e.m. from six matrigel plugs from three separate mice per group in A and from three separate experiments in B. *P<0.05, **P<0.01, ***P<0.001 versus respective control; one-way ANOVA with Bonferroni's multiple comparison tests.

Fig. 4.

Downregulation of PPARγ leads to a significant attenuation of E2F1. (A,B) Analysis of (A) E2F1 mRNA expression (B) E2F1 protein expression in human PMVEC transfected with Non-target (Con) and PPARγ siRNAs. (C) E2F1 expression in wild-type (WT) and Tie2CrePPARγ^flox/flox (KO) mice. Densitometry analysis was used to quantify the amount of protein per sample in B,C. (D) Migration assay in human PMVECs transfected with Non-target (‘C’) and E2F1 (‘E’) siRNA, stimulated with vehicle (H₂O; Con) and VEGF (50 ng/ml) for 6 h. Error bars represent mean±s.e.m. from three separate experiments. *P<0.05, ***P<0.001 versus respective control; unpaired two-tailed t-test in A–C; one-way ANOVA with Bonferroni's multiple comparison tests in D.

Fig. 5.

Loss of PPARγ leads to similar dysregulation of Wnt signaling as downregulation of E2F1. Human PMVECs were transfected with Non-target, PPARγ, or E2F1 siRNAs. Samples were collected 48 h after initiation of transfections. (A,B) Protein expression levels of (A) GSKIP and (B) phosphorylated GSK3β (Ser9), in PMVECs transfected with Non-target (Con) and PPARγ siRNA. (C,D) β-catenin protein expression in PMVECs transfected with Non-target (Con) and PPARγ siRNA measured in the (C) cytoplasm and (D) nucleus. (E–G) Protein expression levels of (E) GSKIP, (F) phosphorylated GSK3β (Ser9), and (G) β-catenin, in PMVECs transfected with Non-target (Con) and E2F1 siRNA. Densitometry analysis was used to quantify the amount of protein per sample in A–C,E–G. Error bars represent mean±s.e.m. from three separate experiments. *P<0.05, **P<0.01, ***P<0.001 versus respective control; unpaired two-tailed t-test.

Fig. 6.

Inhibition of GSKβ is able to restore β-catenin levels and rescue migration defect. (A) β-catenin protein levels of control (Non-target siRNA; ‘C’) and PPARγ-silenced (PPARγ siRNA; ‘P’) human PMVECs. PMVECs were treated with vehicle (DMSO; Con) and BIO-Acetoxime; 6-Bromoindirubin-3′-acetoxime, (BIA; 20 nM) for 6 h. Densitometry analysis was used to quantify the amount of protein per sample. (B,C) Migration assay with human PMVECs treated with PPARγ (‘P’; B), E2F1 (‘E’; C) and Non-target (‘C’; B,C) siRNAs. Cells were stimulated with vehicle (Con), VEGF (50 ng/ml) and VEGF in combination with BIA for 6 h. Bars represent mean±s.e.m. from three separate experiments. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 versus respective control; one-way ANOVA with Bonferroni's multiple comparison tests in A-C.

Fig. 7.

Loss of GSKIP leads to severe angiogenic defect in vitro. (A,B) Analysis of GSKIP (A) mRNA and (B) protein expression levels 48 h after initiation of siRNA transfections. (C) In vitro angiogenesis assay of human PMVECs transfected with Non-target (Con) and GSKIP siRNA. Cells were stimulated with vehicle (H₂O; Con) and VEGF (50 ng/ml) for 16 h. (D) Tube formation was quantified by tube area per total field area (10× magnification) from three different fields per experiment. (E,F) Protein expression levels of (E) GSKIP and (F) β-catenin in PMVECs isolated from wild-type (WT) and Tie2CrePPARγ^flox/flox (KO) mice. Densitometry analysis was used to quantify the amount of protein per sample. Error bars represent mean±s.e.m. from three separate experiments. *P<0.05, **P<0.01, ****P<0.0001 versus respective control; unpaired two-tailed t-test in A,E,F; one-way ANOVA with Bonferroni's multiple comparison tests in D.

Fig. 8.

Hypothetical model explaining the dysfunctional Wnt–β-catenin signaling in PPARγ deficiency. We propose a model where loss of PPARγ leads to impaired angiogenesis through a dysfunctional Wnt–β-catenin signaling pathway. PPARγ contributes to the activity of the Wnt–β-catenin signaling pathway by regulating E2F1 expression, which is essential in sustaining critical levels of GSKIP in PMVECs and contributes to the migration capacity of PMVECs. GSKIP inhibits GSK3β to inactivate the degradation complex of β-catenin (β-C). This allows β-catenin to accumulate and form complexes with its co-factors such as PPARγ (Alastalo et al., 2011). The PPARγ–β-catenin complex translocates into the nucleus and regulates genes important for endothelial cell homeostasis. APLN, apelin.