Methylglyoxal-Induced Dysfunction in Brain Endothelial Cells via the Suppression of Akt/HIF-1α Pathway and Activation of Mitophagy Associated with Increased Reactive Oxygen Species

Abstract

Methylglyoxal (MG) is a dicarbonyl compound, the level of which is increased in the blood of diabetes patients. MG is reported to be involved in the development of cerebrovascular complications in diabetes, but the exact mechanisms need to be elucidated. Here, we investigated the possible roles of oxidative stress and mitophagy in MG-induced functional damage in brain endothelial cells (ECs). Treatment of MG significantly altered metabolic stress as observed by the oxygen-consumption rate and barrier-integrity as found in impaired trans-endothelial electrical resistance in brain ECs. The accumulation of MG adducts and the disturbance of the glyoxalase system, which are major detoxification enzymes of MG, occurred concurrently. Reactive oxygen species (ROS)-triggered oxidative damage was observed with increased mitochondrial ROS production and the suppressed Akt/hypoxia-inducible factor 1 alpha (HIF-1α) pathway. Along with the disturbance of mitochondrial bioenergetic function, parkin-1-mediated mitophagy was increased by MG. Treatment of N-acetyl cysteine significantly reversed mitochondrial damage and mitophagy. Notably, MG induced dysregulation of tight junction proteins including occludin, claudin-5, and zonula occluden-1 in brain ECs. Here, we propose that diabetic metabolite MG-associated oxidative stress may contribute to mitochondrial damage and autophagy in brain ECs, resulting in the dysregulation of tight junction proteins and the impairment of permeability.

Keywords: brain endothelial cells; hypoxia-inducible factor 1 (HIF-1α); methylglyoxal; mitophagy; oxidative stress.

Scheme 1

Calculation of cellular bioenergetic profile in bEND.3 cell

Figure 1

Formation of Methylglyoxal (MG)-adducts and alteration of cell viability in bEnd.3 cells after MG treatment. (a–d) bEnd.3 cells were incubated with MG at different concentrations (0–1000 μM) for 24 h. MG-adduct formation in (a) the cell lysate or (b) the conditioned medium was determined by western blotting and Coomassie Brilliant Blue staining. Representative images were shown. The expression levels of (c) glyoxalase (Glo)-1 and (d) Glo-2 were examined after cell treatment with 1000 μM MG for 24 h (n = 3~4). (e) The changes in cell viability after MG exposure (6 h or 24 h) were determined by (e) PI staining or (f) MTT assay (n = 3). Data are presented as the mean ± SEM. * p < 0.05, ** p < 0.01 vs. control (CON).

Figure 2

Total cellular and mitochondrial reactive oxygen species (ROS) production and suppression of the Akt/HIF-1α pathway after MG exposure. (a) Total ROS formation was detected in DCF-DA-stained bEND.3 cells at the indicated times (n = 3). (b) Confocal microscopy images of DCF-DA-stained bEND.3 cells 24 h after MG treatment. Scale bar: 20 μm. (n = 3). (c) Mitochondrial ROS formation was measured at 0, 1, 3, and 6 h after MG treatment by confocal microscopy. Scale bar: 20 μm. Fluorescence was quantified from three independent experiments. (n = 3). Representative images are shown. (d,e) The protein levels of (d) phosphorylated and total Akt (p-Akt and t-Akt) and (e) HIF-1α were determined by western blotting. Protein levels were normalized to β-actin levels (n = 3). Data are presented as the mean ± SEM. * p < 0.05, ** p < 0.01 vs. control (CON).

Figure 3

Mitochondrial bioenergetic disturbance and activation of mitochondrial autophagy following MG treatment in brain endothelial cells (ECs). (a,b) bEnd.3 cells were incubated with MG for 6 h and applied to Seahorse MitoStress Assay (n = 3). (a) The profile of the oxygen consumption rate was plotted. (b) The parameters for mitochondrial respiration were calculated. (c) The localization of LC3B and mitochondria was examined at indicated times after MG treatment by confocal microscopy. Mitochondria were labeled by MitoTracker. Pearson’s correlation coefficients were calculated from three independent experiments (n = 3). Representative images are shown. Scale bar: 20 μm. (d,e) The protein level of (d) Parkin-1 and (e) LC3B-II in the mitochondrial fraction was determined by western blotting. Protein levels were normalized to that of cytochrome C oxidase subunit 4 (COXIV) (n = 3). (f) Mitochondrial mass was determined by staining with NAO (n = 3). Data are presented as the mean ± SEM. * p < 0.05, ** p < 0.01 vs. control (CON).

Figure 4

Effects of NAC and autophagy inhibitors on mitochondrial damage induced by MG in brain ECs. (a,b) bEnd.3 cells were treated with NAC (5 mM) for 1 h before and during MG stimulation (1000 μM) for 6 h. Cells were applied to Seahorse MitoStress Assay (n = 3). (a) The profile of the oxygen consumption rate was plotted. (b) The parameters for mitochondrial respiration were calculated. (c,d) NAC or autophagy inhibitors (3-MA or Baf A) were treated for 1 h before MG treatment, and maintained for 6 h during MG treatment. (c) The localization of LC3B and mitochondria was examined by confocal microscopy. Mitochondria were labeled by MitoTracker. Representative images are shown. Scale bar: 20 μm. (d) The effects of NAC or autophagy inhibitors (3-MA or Baf A) on MG-attenuated MTT reduction were determined (n = 3~4). Data are presented as the mean ± SEM. * p < 0.05, ** p < 0.01 vs. control (CON). # p < 0.05, ## p < 0.01 vs. MG-treated cells.

Figure 5

Autophagy activity and co-localization of occludin with LC3B 24 h after MG exposure in brain ECs. (a,b) The protein levels of LC3B-I/II were determined (a) at 0, 1, 3, or 6 h after treatment with 1000 μM MG (n = 3) or (b) at 24 h after treatment with 0–1000 μM MG by western blotting (n = 4). (c) MG-exposed bEnd.3 cells were stained with an antibody against LC3B and visualized by confocal microscopy. Scale bar: 50 μm. (n = 4) (d) Co-localization of occludin with LC3B was analyzed after MG exposure using confocal microscopy. Scale bar: 20 μm. Pearson’s correlation coefficients were calculated from three independent experiments (n = 3). Representative images are shown. Data are presented as mean ± SEM. * p < 0.05 vs. control (CON).

Figure 6

MG induces endothelial dysfunction through tight junction degradation and permeability destruction in bEND.3 cells. The protein expression levels of (a) occludin (n = 5), (b) claudin-5 (n = 4), and (c) ZO-1 (n = 3) were determined 24 h after treatment with 100–1000 μM MG by western blotting. (d) MG-exposed bEnd.3 cells were stained with an antibody against occludin, claudin-5, or ZO-1 and visualized by confocal microscopy. Scale bar: 50 μm. (e,f) Functional changes in endothelial permeability were measured by (e) an in vitro FITC-dextran permeability assay (n = 3–7) and (f) TEER measurements (n = 4) after MG treatment. Representative images are shown. Data are presented as mean ± SEM. * p < 0.05, ** p < 0.01 vs. control (CON).