aFGF alleviates diabetic endothelial dysfunction by decreasing oxidative stress via Wnt/β-catenin-mediated upregulation of HXK2

Abstract

Vascular complications of diabetes are a serious challenge in clinical practice, and effective treatments are an unmet clinical need. Acidic fibroblast growth factor (aFGF) has potent anti-oxidative properties and therefore has become a research focus for the treatment of diabetic vascular complications. However, the specific mechanisms by which aFGF regulates these processes remain unclear. The purpose of this study was to investigate whether aFGF alleviates diabetic endothelial dysfunction by suppressing mitochondrial oxidative stress. We found that aFGF markedly decreased mitochondrial superoxide generation in both db/db mice and endothelial cells incubated with high glucose (30 mM) plus palmitic acid (PA, 0.1 mM), and restored diabetes-impaired Wnt/β-catenin signaling. Pretreatment with the Wnt/β-catenin signaling inhibitors IWR-1-endo (IWR) and ICG-001 abolished aFGF-mediated attenuation of mitochondrial superoxide generation and endothelial protection. Furthermore, the effects of aFGF on endothelial protection under diabetic conditions were suppressed by c-Myc knockdown. Mechanistically, c-Myc knockdown triggered mitochondrial superoxide generation, which was related to decreased expression and subsequent impaired mitochondrial localization of hexokinase 2 (HXK2). The role of HXK2 in aFGF-mediated attenuation of mitochondrial superoxide levels and EC protection was further confirmed by si-Hxk2 and a cell-permeable form of hexokinase II VDAC binding domain (HXK2VBD) peptide, which inhibits mitochondrial localization of HXK2. Taken together, these findings suggest that the endothelial protective effect of aFGF under diabetic conditions could be partly attributed to its role in suppressing mitochondrial superoxide generation via HXK2, which is mediated by the Wnt/β-catenin/c-Myc axis.

Keywords: Diabetes; Endothelial dysfunction; HXK2; Mitochondrial superoxide; aFGF.

Fig. 1

aFGF attenuated diabetes-induced endothelial dysfunction both in vivo and in vitro. (A) Representative immunofluorescence with CD31 of aorta tissue sections, scale bars = 200 μm. (B) Representative confocal images of aortic rings sprouting, scale bars = 500 μm. (C) Concentration–response curves following Ach or SNP treatments of mice aortic rings. (D) Representative confocal images of TUNEL stained aorta tissue sections, scale bars = 20 μm, from db/m mice, db/db mice, and intraperitoneal aFGF (0.5 mg/kg) treated db/db mice aorta tissue sections. (E) Capillary-like tube formation of HUVECs, scale bars = 300 μm, HUVECs were cultured either in NG or HG + PA medium in the presence or absence of aFGF (100 ng/mL) for 72 h, MAN was served as the osmotic control for the HG + PA. (F) Representative confocal images of aortic rings sprouting, scale bars = 500 μm. (G) Cell lysates of HUVECs were used to detect the Bax, Bcl-2 as well as c-Caspase-3 protein levels by immunoblotting. β-Actin was served as the loading control. All values displayed are means ± SEM of 6 independent experiments. For (A)-(D), #p < 0.05 vs. db/m mice; *p < 0.05 vs. db/db mice or vehicle treated db/db mice; For (E)-(G), #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA.

Fig. 2

aFGF lowered the level of mitochondrial superoxide in vascular endothelial cells. (A) OCR was analysed using a Seahorse XF analyser. (B) ATP production in HUVEC. (C) Mitochondrial O₂^•− in HUVEC was measured by mitochondria targeted probe MitoSOX and UPLC after accumulation of O₂^•−-specific product 2-OH-Mito-E⁺. (D) mtROS of HUVECs was detected by MitoSOX staining assay, scale bars = 1000 μm. HUVECs were cultured either in NG or HG + PA medium in the presence or absence of aFGF (100 ng/mL) or MitoTEMPO (10 μM) for 72 h. MAN was served as the osmotic control for the HG + PA. (E) Mitochondrial membrane potential of HUVECs was detected by TMRM fluorescence staining, scale bars = 5 μm. (F) Representative confocal images of MitoSOX stained aorta tissue sections, scale bars = 20 μm, from db/m mice, db/db mice, and intraperitoneal aFGF (0.5 mg/kg) or MitoTEMPO (0.7 mg/kg) treated db/db mice. (G) Capillary-like tube formation of HUVECs, scale bars = 300 μm. All values displayed are means ± SEM of 6 independent experiments. For (A)-(E) and (G), #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA; For (F), #p < 0.05 vs. db/m mice; *p < 0.05 vs. db/db mice or vehicle treated db/db mice.

Fig. 3

aFGF restored HG + PA-reduced Wnt/β-catenin signaling pathway activity, and the endothelial protective action of aFGF against HG + PA is Wnt/β-catenin signaling pathway dependent, in vitro. (A) Representative immunofluorescence staining with β-catenin in HUVECs, scale bars = 5 μm, (B) Nuclear and cytosolic extracts from HUVECs were isolated to detect β-catenin protein level by immunoblotting. Lamin B1 and β-Actin were served as loading controls for nuclear and cytosolic fractions, respectively, HUVECs were cultured either in NG or HG + PA medium alone or with aFGF (100 ng/mL) for 72 h, MAN was served as the osmotic control for the HG + PA. For manipulation of Wnt/β-catenin pathway, IWR (5 μM) was pretreated for 2 h before aFGF administration. (C) OCR was analysed using a Seahorse XF analyser. (D) ATP production in HUVEC. (E) Mitochondrial O₂^•− in HUVEC was measured by mitochondria targeted probe MitoSOX and UPLC after accumulation of O₂^•−-specific product 2-OH-Mito-E⁺. (F) mtROS of HUVECs was detected by MitoSOX staining assay, scale bars = 1000 μm, (G) Mitochondrial membrane potential was detected by TMRM fluorescence staining, scale bars = 5 μm, (H) TUNEL assay of HUVECs, scale bars = 100 μm, (I) Capillary-like tube formation of HUVECs, scale bars = 300 μm. All values displayed are means ± SEM of 6 independent experiments. #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA; % p < 0.05 vs. HG + PA co-incubated with aFGF.

Fig. 4

aFGF activated Wnt/β-catenin signaling pathway to promote c-Myc expression and protect the endothelial function against HG + PA. (A) TEF/LEF-luciferase reporter activity in HUVECs. (B) Immunoblotting and sqRT-PCR analysis of c-Myc in HUVECs. HUVECs were cultured either in NG or HG + PA medium in the presence or absence of aFGF (100 ng/mL) for 72 h, MAN was served as the osmotic control for the HG + PA. For manipulation of Wnt/β-catenin pathway, IWR (5 μM) and ICG-001 (10 μM) was pretreated for 2 h before aFGF administration. (C) OCR was analysed using a Seahorse XF analyser. (D) ATP production in HUVEC. (E) Mitochondrial O₂^•− in HUVEC was measured by mitochondria targeted probe MitoSOX and UPLC after accumulation of O₂^•−-specific product 2-OH-Mito-E⁺. (F) mtROS of HUVECs was detected by MitoSOX staining assay, scale bars = 1000 μm, (G) Mitochondrial membrane potential was detected by TMRM fluorescence staining, scale bars = 5 μm, (H) TUNEL assay of HUVECs, scale bars = 100 μm, (I) Capillary-like tube formation of HUVECs, scale bars = 300 μm. All values displayed are means ± SEM of 6 independent experiments. #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA; % p < 0.05 vs. HG + PA co-incubated with aFGF.

Fig. 5

aFGF promoted c-Myc expression to increase HXK2 expression and its mitochondrial localization. (A) Mitochondrial and cytosolic extracts were isolated to detect the HXK2 and Cytochrome c protein levels in HUVECs. COX IV and β-Actin were served as loading controls for mitochondrial and cytosolic fractions, respectively. (B) Representative immunofluorescence analysis of HXK2 (red) in HUVECs. The COX IV immunostaining (green) highlights mitochondria, and nuclei were stained with DAPI (blue), scale bars = 5 μm, (C) mtROS of HUVECs was detected by MitoSOX staining assay, scale bars = 1000 μm, (D) Mitochondrial membrane potential was detected by TMRM fluorescence staining, scale bars = 5 μm, (E) TUNEL assay of HUVECs, scale bars = 100 μm, (F) Capillary-like tube formation of HUVECs, scale bars = 300 μm. All values displayed are means ± SEM of 6 independent experiments. #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA; % p < 0.05 vs. HG + PA co-incubated with aFGF. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 6

aFGF promoted the combination of HXK2 with mitochondria to protect the endothelial function against HG + PA. (A) Mitochondrial and cytosolic extracts were isolated to detect the HXK2 and Cytochrome c protein levels in HUVECs. HUVECs were cultured either in NG or HG + PA medium alone or with aFGF (100 ng/mL) for 72 h, MAN was served as the osmotic control for the HG + PA, cell-permeable form of HXK2VBD (100 μM) was pretreated for 1 h before aFGF administration. (B) OCR was analysed using a Seahorse XF analyser. (C) ATP production in HUVEC. (D) Mitochondrial O₂^•− in HUVEC was measured by mitochondria targeted probe MitoSOX and UPLC after accumulation of O₂^•−-specific product 2-OH-Mito-E⁺. (E) Representative immunofluorescence analysis of HXK2 (red) in HUVECs. The COX IV immunostaining (green) highlights mitochondria, and nuclei were stained with DAPI (blue), scale bars = 5 μm, (F) mtROS of HUVECs was detected by MitoSOX staining assay, scale bars = 1000 μm, (G) Mitochondrial membrane potential was detected by TMRM fluorescence staining, scale bars = 5 μm, (H) TUNEL assay of HUVECs, scale bars = 100 μm, (I) Capillary-like tube formation of HUVECs, scale bars = 300 μm. All values displayed are means ± SEM of 6 independent experiments. #p < 0.05 vs. NG or MAN; *p < 0.05 vs. HG + PA; % p < 0.05 vs. HG + PA co-incubated with aFGF. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 7

Inhibition of Wnt/β-catenin-c-Myc-HXK2 pathway and/or interference with mitochondrial location of HXK2 abrogated aFGF-prompted wound healing in T2DM. (A) Confocal immunofluorescence staining with CD31 of wounded skin tissue sections at 7 days post wounding, scale bars = 30 μm, and (B) Images of skin wounds from db/m mice, db/db mice and db/db mice receiving aFGF (100 ng/mL) treatment. For signaling pathway analysis, ICG-001 (10 μM) or 10058-F4 (50 μM) or BrPA (50 μM) was injected intradermally into the wound edges in the mice after aFGF treatment. (C) Confocal immunofluorescence staining with CD31 of wounded skin tissue sections at 7 days post wounding, scale bars = 30 μm, and (D) images of skin wounds from db/m mice, db/db mice and db/db mice receiving aFGF (100 ng/mL) treatment. Ad-sh-HXK2 was injected intradermally into the wound edges in the mice the day before wounding, HXK2VBD peptide (100 μM) was injected intradermally into the wound edges in the mice after aFGF treatment. (E) Schematic showing that aFGF alleviated diabetes-induced endothelial impairment by downregulating mtROS via Wnt/β-catenin pathway. All values displayed are means ± SEM of 6 independent experiments. #p < 0.05 vs. db/m mice; *p < 0.05 vs. db/db mice; % p < 0.05 vs. aFGF treated db/db mice.