β2-Adrenergic Receptor-Mediated HIF-1α Upregulation Mediates Blood Brain Barrier Damage in Acute Cerebral Ischemia

Abstract

Disruption of the blood brain barrier (BBB) within the thrombolytic time window is an antecedent event to intracerebral hemorrhage in ischemic stroke. Our recent studies showed that 2-h cerebral ischemia induced BBB damage in non-infarcted area and secreted matrix metalloproteinase-2 (MMP-2) accounted for this disruption. However, the factors that affect MMP-2 secretion and regulate BBB damage remains unknown. Since hypoxia-inducible factor-1 alpha (HIF-1α) was discovered as a mater regulator in hypoxia, we sought to investigate the roles of HIF-1α in BBB damage as well as the factors regulating HIF-1α expression in the ischemic brain. in vivo rat middle cerebral artery occlusion (MCAO) and in vitro oxygen glucose deprivation (OGD) models were used to mimic ischemia. Pretreatment with HIF-1α inhibitor YC-1 significantly inhibited 2-h MCAO-induced BBB damage, which was accompanied by suppressed occludin degradation and vascular endothelial growth factor (VEGF) mRNA upregulation. Interestingly, β2-adrenergic receptor (β2-AR) antagonist ICI 118551 attenuated ischemia-induced BBB damage by regulating HIF-1α expression. Double immunostaining showed that HIF-1α was upregulated in ischemic neurons but not in astrocytes andendothelial cells. Of note, HIF-1α inhibition with inhibitor YC-1 or siRNA significantly prevented OGD-induced VEGF upregulation as well as the secretion of VEGF and MMP-2 in neurons. More importantly, blocking β2-AR with ICI 118551 suppressedHIF-1α upregulation in ischemic neurons and attenuated occludin degradation induced by the conditioned media of OGD-treatedneurons. Taken together, blockade of β2-AR-mediated HIF-1α upregulation mediates BBB damage during acute cerebral ischemia. These findings provide new mechanistic understanding of early BBB damage in ischemic stroke and may help reduce thrombolysis-related hemorrhagic complications.

Keywords: HIF-1α; blood brain barrier; cerebral ischemia; matrix metalloproteinase; tight junction proteins; β2-AR.

Figure 1

Effect of 2-h ischemia on hypoxia-inducible factor-1 alpha (HIF-1α) and vascular endothelial growth factor (VEGF) expression. (A) Diagram of the experimental procedure. (B) The black outlines showed the two regions of interest (ROI): 1, parietal and insular cortex and dorsal striatum; 2, ventral striatum and preoptic area. (C) Representative western blot images of HIF-1α protein in ROI 1 and 2 in ischemic (I) and nonischemic (NI) hemisphere (left panel). Ratios of HIF-1α (I/NI) were quantitated (right panel). (D) Representative western blot images of VEGF protein in region of ROI 1 and 2 in I and NI hemisphere (left panel). Ratios of VEGF (I/NI) were quantitated (right panel). After 2-h middle cerebral artery occlusion (MCAO), there was a significant increase of HIF-1α protein (C) but not VEGF protein (D) in ROI 2 but not ROI 1 and pretreatment with HIF-1α inhibitor YC-1 significantly prevented this increase. ^#P < 0.05 vs. Vehicle group in ROI 2, *P < 0.05 vs. Vehicle group in ROI 1. (E) Ratio of VEGF mRNA (I/NI) was quantitated. Two-hour MCAO induced significant upregulation of VEGF mRNA and pretreatment with YC-1 significantly prevented this increase (^#P < 0.05 vs. Vehicle). Data were expressed as mean ± SEM, n = 4/group.

Figure 2

Effect of YC-1 treatment on 2-h ischemia-induced Evan’s blue (EB) leakage and occludin expression. (A) Diagram of the experimental procedure. Rats received YC-1 at 24 h and 30 min before ischemia onset. (B) Ten consecutive sections showed EB leakage from vehicle or YC-1 treated rats. (C) EB leakage was quantitated and expressed as average area proportion of section measured (%). ^#P < 0.05 vs. Vehicle group. n = 6/group. (D) EB leakage in the brain tissue was quantitated according to the external EB standard curve. EB leakage was expressed as per gram of brain tissue (μg/g). ^#P < 0.05 vs. Vehicle group. Data were expressed as mean ± SEM, n = 5/group. (E) Western blot was conducted to detect occludin protein levels in the NI and ischemic (I) hemispheric tissue. A representative western blot revealed occludin protein expression in ROI 1 and ROI 2 treated with YC-1 or vehicle (left panel). The band intensity of occludin was quantitated after normalization to the Ponceau S (right panel). Two-hour MCAO induced a significant reduction in occludin protein level in ROI 2 (E, right panel. *P < 0.05 vs. Vehicle group in ROI 1), but not in ROI 1 (P > 0.05). Pretreatment with YC-1 attenuated occludin degradation (^#P < 0.05 vs. Vehicle group). Data were expressed as mean ± SEM, n = 5/group. (F) EB leakage (red) was seen in the ischemic ROI 2 after 2-h ischemia and IHC showed occludin degradation in the area where EB leakage occurred. Pretreatment with YC-1 significantly attenuated occludin degradation as well as EB leakage. n = 3/group. Data were expressed as mean ± SEM.

Figure 3

Effect of β2 adrenergic receptor (β2-AR) antagonist ICI118551 on 2-h ischemia-induced HIF-1α expression and blood brain barrier (BBB) disruption. (A) Diagram of the experimental procedure. Rats received ICI118551 10 min before the onset of ischemia. (B,C) Western blot was conducted to detect HIF-1α (B) and occludin (C) protein levels in the NI and ischemic (I) hemispheric tissue. A representative western blot revealed HIF-1α (B) and occludin (C) protein expression treated with YC-1 or vehicle. The band intensity of HIF-1α and occludin was quantitated after normalization to the β-actin. Two-h MCAO induced a significant increase of HIF-1α (B) and reduction of occludin (C). Pretreatment with ICI118551 inhibited HIF-1α upregulation and attenuated occludin degradation (^#P < 0.05 vs. Vehicle group). Data were expressed as mean ± SEM, n = 5/group. (D) Ten consecutive sections showed EB leakage from vehicle or YC-1 treated rats. (E) EB leakage was quantitated and expressed as average area proportion of section measured (%). ^##P < 0.01 vs. Vehicle group, n = 6/group. (F) EB leakage in the brain tissue was quantitated according to the external EB standard curve. EB leakage was expressed as per gram of brain tissue (μg/g). ^#P < 0.05 vs. Vehicle group. Data were expressed as mean ± SEM, n = 5/group.

Figure 4

Effect of YC-1 on 2-h ischemia-induced HIF-1α expression in endothelial cells, neurons and astrocytes. (A) Diagram of the experimental procedure. Rats received YC-1 at 24 h and 30 min before ischemia onset. The cellular pattern of HIF-1α expression was analyzed by double-immunofluorescence labeling. Double immunostain of HIF-1α (green) and NeuN (marker of neurons, red) showed a good colocalization of HIF-1α and neurons (B, upper panel). Double immunostain of HIF-1α (green) and GFAP (marker of astrocytes, red) showed no co-localization of HIF-1α and astrocytes (B, middle panel). Double immunostain of HIF-1α (green) and RECA-1 (marker of endothelial cells, red) showed no co-localization of HIF-1α and endothelial cells (B, bottom panel). n = 3/group. Scale bar = 50 μm. (C) YC-1 significantly decreased the proportion of HIF-1α-positive neurons but not astrocytes or endothelial cells *P < 0.05 vs. Vehicle group. n = 3/group.

Figure 5

Effect of HIF-1α inhibition by YC-1 or HIF-1α siRNA on 2-h oxygen glucose deprivation (OGD)-induced VEGF secretion in neurons. The neurons were exposed to OGD for 2 h and VEGF secretion was assessed by measuring its content in cellular extracts (CE) and conditioned media (CM) using western blot and ELISA respectively. Two-hours OGD induced a significant increase in VEGF mRNA expression in neurons and pretreatment with YC-1 (A, left panel) or HIF-1α siR (A, right panel) significantly inhibited this effect. **P < 0.01 vs. control (Ctrl), ^#P < 0.05 vs. OGD, n = 4/group. (B) Representative blots for VEGF and β-actin (upper panel). The relative VEGF protein level was calculated after normalization to β-actin (bottom panel). Two-hours OGD had no effect on the VEGF protein levels in neurons but induced a significant increase of VEGF in cultured media and pretreatment with HIF-1α inhibitor YC-1 (C, left panel) or HIF-1α siR (C, right panel) attenuated this VEGF secretion induced by 2-h OGD. **P < 0.01 vs. Ctrl. ^#P < 0.05 vs. OGD. ^##P < 0.01 vs. OGD. HIF-1α siR represents HIF-1α siRNA. Data are expressed as mean ± SEM, n = 3/group.

Figure 6

Effect of YC-1 on ischemia-induced matrix metalloproteinase-2 (MMP-2) secretion in neurons and endothelial cells. Gelatin zymography results showed that 2-h OGD elevated MMP-2 levels in conditioned media (CM), which was accompanied by a significant reduction in their levels in whole cellular extracts (CE). (A) Representative gelatin zymogram (upper panel) showing that OGD decreased MMP-2 levels in CE of SH-SY5Y cells and YC-1 treatment prevents this decrease. The relative levels of MMP-2 were quantitated (bottom panel). (B) Representative gelatin zymogram (upper panel) showing that OGD increased MMP-2 levels in CM of SH-SY5Y cells and YC-1 treatment prevents this increase. The relative levels of MMP-2 were quantitated (bottom panel). (C) Representative gelatin zymogram (upper panel) showing that YC-1 treatment did not significantly inhibited OGD-induced MMP-2 decrease in CE of endothelial cells. The relative levels of MMP-2 were quantitated (bottom panel). (D) Representative gelatin zymogram (upper panel) showing that YC-1 treatment did not inhibit OGD-induced MMP-2 increase in CM of endothelial cells. The relative levels of MMP-2 were quantitated (bottom panel). (E) Pretreatment with ICI118551 did not prevent 2-h OGD-induced occludin degradation in endothelial cells. *P < 0.05 vs. control (Ctrl), ^#P < 0.05 vs. OGD. Data are expressed as mean ± SEM, n = 6/group.

Figure 7

Effect of ICI118551 on ischemia-induced HIF-1α expression and occludin degradation induced by the conditioned media of OGD-treated neurons. HIF-1α expression in neurons, endothelial cells and astrocytes were detected after 2-h OGD. Two-hour OGD induced a significant HIF-1α upregulation in neurons (A) but not endothelial cells (B) or astrocytes (C). Pretreatment with ICI118551 inhibited OGD-induced HIF-1α upregulation in neurons (A), but did not attenuated OGD-induced occludin degradation in endothelial cells (B) *P < 0.05 vs. control (Ctrl), **P < 0.01 vs. control (Ctrl), ^#P < 0.05 vs. OGD. (D) Western blot analysis showed that conditioned media (CM) from OGD-treated neurons induced occludin degradation in endothelial cells and pretreatment neurons with ICI 118551 attenuated this degradation. **P < 0.01 vs. neuron CM (Ctrl), ^#P < 0.05 vs. neuron CM.