17β-Estradiol Ameliorates Tight Junction Disruption via Repression of MMP Transcription

Abstract

The blood-brain barrier (BBB) or blood-spinal cord barrier (BSCB) formed by capillary endothelial cells provides a physical wall between the central nervous system (CNS) and circulating blood with highly selective permeability. BBB/BSCB disruption by activation of matrix metalloproteinases (MMPs) has been shown to result in further neurological damage after CNS injury. Recently it has been discovered that estrogen attenuates BBB/BSCB disruption in in vitro and in vivo models. However, the molecular mechanism underlying the estrogen-mediated attenuation of BBB/BSCB disruption has not been elucidated fully. In the present study, we found that 17β-estradiol (E2) suppresses nuclear factor-κB-dependent MMP-1b, MMP-2, MMP-3, MMP-9, MMP-10, and MMP-13 gene activation in microvessel endothelial bEnd.3 cells subjected to oxygen and glucose deprivation/reperfusion injury. E2 induced the recruitment of ERα and nuclear receptor corepressor to the nuclear factor-κB binding site on the MMPs' gene promoters. Consistently, ER antagonist ICI 182.780 showed opposite effects of E2. We further found that E2 attenuates tight junction disruption through the decreased degradation of tight junction proteins in bEnd.3 cells subjected to oxygen and glucose deprivation-reperfusion injury. In addition, E2 suppressed the up-regulation of MMP expression, leading to a decreased BSCB disruption in the injured spinal cord. In conclusion, we discovered the molecular mechanism underlying the protective role of estrogenin BBB/BSCB disruption using an in vitro and in vivo model. Our study suggests that estrogens may provide a potential therapeutic intervention for preserving BBB/BSCB integrity after CNS injury.

Figure 1.. E₂ suppresses the up-regulation of MMP expression in brain microvessel endothelial bEnd.3 cells subjected to OGD/reperfusion injury.

The expression of MMPs was determined in bEnd.3 cells subjected to OGD/reperfusion injury with or without 10 nM E₂. A, The expression of MMP-1b, MMP-2, MMP-3, MMP-9, MMP-10, and MMP-13 gene is up-regulated in bEnd.3 cells subjected to OGD/reperfusion (−OGD vs + OGD/−E₂). However, E₂ treatment suppresses OGD/reperfusion-induced up-regulation of the MMP genes (+OGD/−E₂ vs +OGD/+E₂). Transcripts of MMPs and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. B, Lysates were immunoblotted with anti-MMP-1b, anti-MMP-2, anti-MMP-3, anti-MMP-9, anti-MMP-10, and anti-MMP-13 antibodies, respectively. The anti-β-actin antibody was used as a loading control. Western blots were analyzed quantitatively. All data represent mean ± SEM for three independent experiments. *, P ≤ .05; **, P ≤ .01; ***, P ≤ .005.

Figure 2.. ERα mediates E₂-induced suppression of MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury.

A and B, The expression of ERs was determined in bEnd.3 cells subjected to OGD/reperfusion with or without E₂. The expression levels of ERα and ERβ are not changed by E₂ treatment in bEnd.3 cells subjected to OGD/reperfusion injury. Transcripts of ERα, ERβ, and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. In addition, lysates were immunoblotted with anti-ERα and anti-ERβ antibodies, respectively. Western blots were analyzed quantitatively. C, Representative photomicrographs of ERα and ERβ in bEnd.3 cells. Cells were immunostained with anti-ERα and anti-ERβ antibodies, respectively (n = 3). E₂ induces ERα translocation into the nucleus in bEnd.3 cells subjected to OGD/reperfusion injury. However, ERβ localizes constantly in the nucleus with or without OGD/reperfusion injury. Nuclei were identified using DAPI staining. Scale bar, 20 μm. D, After control (siCont) or ER siRNA (siER) was transfected, the transcripts of ERα, ERβ, MMPs, and GAPDH were determined by quantitative PCR. The ERα and ERβ transcripts were efficiently depleted by ERα and ERβ siRNA, respectively. Depletion of ERα attenuates the E₂-induced suppression of MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/siCont/+E₂ vs +OGD/siERα/+E₂). However, the depletion of ERβ does not affect the E₂-induced suppression of MMP gene activation significantly except MMP-1b (+OGD/siCont/+E₂ vs +OGD/siERβ/+E₂). All data represent mean ± SEM for three independent experiments. **, P ≤ .01; ***, P ≤ .005.

Figure 3.. NF-κB p65 is required for both OGD/reperfusion injury-induced MMP gene activation and E₂-induced suppression of MMP gene expression in bEnd.3 cells.

A, Cells were transiently transfected with the wild or mutant type of mouse MMP gene promoter-driven firefly luciferase reporter vector in conjunction with a control renilla luciferase expression vector. Reporter activity is represented as fold activation relative to renilla luciferase activity. Activities of reporters containing MMP-1b, MMP-2, MMP-3, MMP-9, MMP-10, and MMP-13 gene promoter are elevated in bEnd.3 cells subjected to OGD/reperfusion injury (−OGD vs +OGD/−E₂). In addition, E₂ suppresses up-regulated activities of promoter reporters in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/−E₂ vs +OGD/+E₂). However, ERE mutations do not affect activities of MMP gene promoter in bEnd.3 cells subjected to OGD/reperfusion injury with or without E₂. B, Mutations of NF-κB binding site significantly attenuates up-regulated MMP gene promoter activities (−OGD vs +OGD/−E₂). However, marginal or no change of the mutated promoter activities was observed in bEnd.3 cells subjected to OGD/reperfusion injury with or without E₂ treatment (+OGD/−E₂ vs +OGD/+E₂). C, After control or NF-κB p65 siRNA was transfected, transcripts of p65, MMPs, and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. NF-κB p65 transcript was depleted efficiently by siRNA. Depletion of NF-κB p65 attenuates MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/siCont/−E₂ vs +OGD/sip65/−E₂). In addition, the depletion of p65 does not alter the MMP gene expression in bEnd.3 cells subjected to OGD/reperfusion injury with or without E₂ treatment (+OGD/sip65/−E₂ vs +OGD/sip65/+E₂). All data represent mean ± SEM for three independent experiments. *, P ≤ .05; **, P ≤ .01; ***, P ≤ .005.

Figure 4.. E₂ induces repressive transcription activity of NF-κB by forming an ERα/N-CoR/NF-κB association on the promoter of MMP genes in bEnd.3 cells subjected to OGD/reperfusion injury.

A, E₂ induces ERα interaction with NF-κB p65 and N-CoR in bEnd.3 cells subjected to OGD/reperfusion injury. Lysates were immunoprecipitated with anti-ERα antibody and immunoblotted with anti-ERα, p65, and N-CoR antibodies, respectively (n = 3). Immunoprecipitation with normal IgG was used for a negative control. B, After control or N-CoR siRNA was transfected, transcripts of N-CoR, MMPs, and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. N-CoR transcript was depleted efficiently by siRNA. The depletion of N-CoR attenuates E₂-induced suppression of MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/siCont/+E₂ vs +OGD/siN-CoR/+E₂). In addition, depletion of N-CoR elevates MMP gene expression in bEnd.3 cells subjected to OGD/reperfusion injury without E₂ treatment (+OGD/siCont/−E₂ vs +OGD/siN-CoR/−E₂). C, A ChIP assay was performed using the anti-ERα, p65, N-CoR, and phosphorylated RNA polymerase II antibodies, respectively. The occupancy of each protein was quantified with real-time PCR at the gene promoter region encompassing the NF-κB binding site. ChIP using normal IgG was performed as a negative control. Recruitment of NF-κB p65 and RNA pol II increases to the promoter of MMP genes in bEnd.3 cells subjected to OGD/reperfusion injury (−OGD vs +OGD/−E₂). E₂ induces the recruitment of ERα and N-CoR to the promoter of MMP genes in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/−E₂ vs +OGD/+E₂). Simultaneously, RNA polymerase II occupancy is decreased by E₂ treatment (+OGD/−E₂ vs +OGD/+E₂). However, NF-κB p65 occupancy is not changed in bEnd.3 cells subjected to OGD/reperfusion injury with or without E₂ treatment (+OGD/−E₂ vs +OGD/+E₂). All data represent mean ± SEM for three independent experiments. *, P ≤ .05; **, P ≤ .01; ***, P ≤ .005.

Figure 5.. ER antagonist ICI 182.780 attenuates the E₂-injuced suppression of MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury.

A, The expression of MMPs was determined in bEnd.3 cells subjected to OGD/reperfusion with or without E₂+ICI 182.780 treatment. Transcripts of MMPs and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. Inhibition of ERα by ICI 182.780 attenuates E₂-induced suppression of MMP gene activation in bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/−E₂/−ICI 182.780 vs +OGD/+E₂/+ICI 182.780). B, Recruitment of NF-κB p65 and RNA polymerase II increases to the promoter of MMP genes in bEnd.3 cells subjected to OGD/reperfusion injury (−OGD vs +OGD/−E₂). However, ICI 182.780 inhibits the recruitment of ERα and N-CoR on the promoter of MMP genes in E₂-treated bEnd.3 cells subjected to OGD/reperfusion injury (+OGD/−E₂/−ICI 182.780 vs +OGD/+E₂/+ICI 182.780). The occupancies of RNA polymerase II and NF-κB p65 are not changed in bEnd.3 cells subjected to OGD/reperfusion injury with or without E₂+ICI 182.780 treatment. (+OGD/−E₂/−ICI 182.780 vs +OGD/+E₂/+ICI 182.780). A ChIP assay was performed using the anti-ERα, p65, N-CoR, and phosphorylated RNA polymerase II antibodies, respectively. The occupancy of each protein was quantified with real-time PCR at the gene promoter region encompassing the NF-κB binding site. A ChIP using normal IgG was performed as a negative control. All data represent mean ± SEM for three independent experiments. ***, P ≤ .005.

Figure 6.. E₂ enhances tight junction integrity in bEnd.3 cells subjected to OGD/reperfusion injury.

A, Representative photomicrographs of Claudin 5 and Occludin in bEnd.3 cells. E₂ attenuates the degradation of Claudin 5 and Occludin in bEnd.3 cells subjected to OGD/reperfusion injury. Cells were immunostained with anti-Claudin 5 and Occludin antibodies, respectively (n = 3). Nuclei were identified using DAPI staining. Scale bar, 20 μm. Lysates were immunoblotted with anti-Claudin 5 and Occludin antibodies, respectively. Western blots were analyzed quantitatively. B, E₂ increases TEER in bEnd.3 cells subjected to OGD/reperfusion injury. TEER was determined using a voltohmmeter (Millipore) in bEnd.3 cells subjected to OGD/reperfusion with or without E₂. All data represent mean ± SEM for three independent experiments. ***, P ≤ .005.

Figure 7.. E₂ suppresses the up-regulation of MMP gene expression and BSCB disruption in an injured spinal cord.

A, The expression of MMPs was determined in the injured spinal cord of a male rat injected with E₂ (300 μg/kg). E₂ treatment suppresses the up-regulated MMP gene expression in injured spinal cord. Transcripts of MMPs and glyceraldehyde-3-phosphate dehydrogenase were determined by quantitative PCR. As controls, rats were sham operated (Sham) or injected with β-cyclodextrin (Veh) as described in Materials and Methods. B, Representative photographs of leakage of Evans blue dye after the spinal cord injury. E₂ treatment suppresses hemorrhage, showing decreased leakage of Evans blue in the injured spinal cord. Hemorrhage was measured at 1 day after the spinal cord injury using Evans blue dye. The quantification of Evans blue extravasation was performed using a fluorometer (excitation at 620 nm and emission at 680 nm). All data represent mean ± SEM for three independent experiments. ***, P ≤ .005.

Figure 8.. A proposed model.

Upon OGD/reperfusion injury (+OGD), NF-κB p65 translocates into the nucleus (Supplemental Figure 4). In turn, activated NF-κB p65 binds to MMP gene promoters thereby stimulates MMP gene expression in bEnd.3 cells. This event results in the degradation of tight junction proteins and BBB disruption (31, 49, 50). However, E₂ treatment (+OGD/+E₂) induces nuclear translocation of ERα and association of ERα with NF-κB p65 and N-CoR corepressor, which leads to the down-regulation of MMP gene expression. Therefore, E₂ has a protective effect on the BBB disruption via preservation of tight junction proteins.