Poldip2 mediates blood-brain barrier disruption in a model of sepsis-associated encephalopathy

Abstract

Background: Sepsis-associated encephalopathy (SAE), a diffuse cerebral dysfunction in the absence of direct CNS infection, is associated with increased rates of mortality and morbidity in patients with sepsis. Increased cytokine production and disruption of the blood-brain barrier (BBB) are implicated in the pathogenesis of SAE. The induction of pro-inflammatory mediators is driven, in part, by activation of NF-κΒ. Lipopolysaccharide (LPS), an endotoxin produced by gram-negative bacteria, potently activates NF-κΒ and its downstream targets, including cyclooxygenase-2 (Cox-2). Cox-2 catalyzes prostaglandin synthesis and in the brain prostaglandin, E2 is capable of inducing endothelial permeability. Depletion of polymerase δ-interacting protein 2 (Poldip2) has previously been reported to attenuate BBB disruption, possibly via regulation of NF-κΒ, in response to ischemic stroke. Here we investigated Poldip2 as a novel regulator of NF-κΒ/cyclooxygenase-2 signaling in an LPS model of SAE.

Methods: Intraperitoneal injections of LPS (18 mg/kg) were used to induce BBB disruption in Poldip2^+/+ and Poldip2^+/- mice. Changes in cerebral vascular permeability and the effect of meloxicam, a selective Cox-2 inhibitor, were assessed by Evans blue dye extravasation. Cerebral cortices of Poldip2^+/+ and Poldip2^+/- mice were further evaluated by immunoblotting and ELISA. To investigate the role of endothelial Poldip2, immunofluorescence microscopy and immunoblotting were performed to study the effect of siPoldip2 on LPS-mediated NF-κΒ subunit p65 translocation and Cox-2 induction in rat brain microvascular endothelial cells. Finally, FITC-dextran transwell assay was used to assess the effect of siPoldip2 on LPS-induced endothelial permeability.

Results: Heterozygous deletion of Poldip2 conferred protection against LPS-induced BBB permeability. Alterations in Poldip2^+/+ BBB integrity were preceded by induction of Poldip2, p65, and Cox-2, which was not observed in Poldip2^+/- mice. Consistent with these findings, prostaglandin E2 levels were significantly elevated in Poldip2^+/+ cerebral cortices compared to Poldip2^+/- cortices. Treatment with meloxicam attenuated LPS-induced BBB permeability in Poldip2^+/+ mice, while having no significant effect in Poldip2^+/- mice. Moreover, silencing of Poldip2 in vitro blocked LPS-induced p65 nuclear translocation, Cox-2 expression, and endothelial permeability.

Conclusions: These data suggest Poldip2 mediates LPS-induced BBB disruption by regulating NF-κΒ subunit p65 activation and Cox-2 and prostaglandin E2 induction. Consequently, targeted inhibition of Poldip2 may provide clinical benefit in the prevention of sepsis-induced BBB disruption.

Keywords: Blood-brain barrier; Brain microvascular endothelial cells; Cyclooxygenase-2; Lipopolysaccharide; Poldip2; Sepsis-associated encephalopathy.

Fig. 1

Heterozygous deletion of Poldip2 protects against LPS-induced BBB permeability. a Schematic of the experimental design. BBB disruption following 18 h LPS (18 mg/kg, IP) or PBS treatment was assessed by Evans blue dye extravasation in Poldip2^+/+ and Poldip2^+/− mice. Evans blue dye was administered by intravenous injection and allowed to circulate for 30 min before animals were sacrificed. b The graph depicts Evans blue dye concentration extracted from whole brains normalized to dry brain weight. Bars represent mean ± SEM. Two-way ANOVA ***p < 0.001 vs. Poldip2^+/+ + PBS, ^#p < 0.05 vs. Poldip2^+/+ + LPS, n = 4–7 mice/group. c Representative images of whole brains following Evans blue dye extravasation

Fig. 2

LPS induces Poldip2 expression in the brain. Poldip2 expression was measured in Poldip2^+/+ cerebral cortices after 6 h of LPS (18 mg/kg, IP) or PBS treatment. Immunoblotting for Poldip2 was performed and β-tubulin was used as a loading control. Representative blots are shown. The graph depicts Poldip2 expression normalized to β-tubulin. Bars represent mean ± SEM. Unpaired t test, **p < 0.01 vs. PBS, n = 5 mice/group

Fig. 3

Heterozygous deletion of Poldip2 abrogates LPS-induced NF-κΒ/Cox2 signaling in vivo. Poldip2^+/+ and Poldip2^+/− mice were treated with LPS (18 mg/kg, IP) or PBS to examine NF-κΒ/Cox2 signaling in vivo. a Immunoblotting for the NF-κΒ subunit p65 and Cox2 was performed on cerebral cortices of Poldip2^+/+ and Poldip2^+/− mice after 6 h of treatment. β-tubulin was used as a loading control. Representative blots are shown. b p65 expression was quantified by densitometry. The graph depicts p65 expression normalized to β-tubulin. Error bars represent mean ± SEM. Two-way ANOVA *p < 0.05 vs. Poldip2^+/+ + PBS, ^##p < 0.01 vs. Poldip2^+/+ + LPS, n = 6 mice/group. c Cox-2 expression was quantified by densitometry. The graph depicts Cox-2 expression normalized to β-tubulin. Bars represent mean ± SEM. Two-way ANOVA **p < 0.01 vs. Poldip2^+/+ + PBS, ^#p < 0.05 vs. Poldip2^+/+ + LPS, n = 6 mice/group. d PGE2 levels in the cerebral cortices of Poldip2^+/+ and Poldip2^+/− mice were assessed by ELISA after 18 h of LPS or PBS. The graph depicts PGE2 concentrations normalized by total protein concentration. Bars represent mean ± SEM. Two-way ANOVA, ***p < 0.001 vs. Poldip2^+/+ + PBS, ^#p < 0.05 vs. Poldip2^+/+ + LPS, n = 4–6 mice/group

Fig. 4

Poldip2 and Cox2 co-localize in brain endothelial cells. Immunofluorescence microscopy was used to examine Poldip2 and Cox-2 localization in vivo. Poldip2^+/+ mice were treated with LPS (18 mg/kg, IP) or PBS for 18 h before brains were isolated and prepared for staining. Antibodies against PECAM-1 (red), Cox-2 (green), and Poldip2 (pseudo-colored in blue) were used. Composite image depicts overlay of Cox-2, Poldip2, and PECAM-1 images. Scale bars equal 20 μm. Representative images are shown. n = 3–4 mice/group

Fig. 5

SiRNA against Poldip2 blocks LPS-induced p65 nuclear translocation. Confluent RBMVECs were transfected with siRNA against Poldip2 (siPoldip2) or control (siCtl) and treated with LPS (1 μg/mL) or PBS. a Following transfection and PBS or LPS treatment for 1 h, RBMVECs were fixed and incubated with antibodies against p65 (green) and DAPI (blue). White represents areas of colocalization. Scale bars equal 20 μm. Representative images are shown. b The areas of overlap between DAPI and p65 was measured by ImageJ. The graph depicts the average fold change in colocalization ± SEM. Bars represent mean ± SEM of 4 independent experiments. Two-way ANOVA ****p < 0.0001 vs. siCtl + PBS, ^####p < 0.0001 vs. siCtl + LPS. c After transfection and 1 h of treatment, RBMVECs were fractionated and the nuclear fraction was blotted for p65 and Lamin B1. Representative blots are shown. d The graph depicts nuclear p65 normalized to Lamin B1. Bars represent mean ± SEM of three independent experiments. Two-way ANOVA, ***p < 0.001 vs. siCtl, ^##p < 0.01 vs. siCtl + LPS

Fig. 6

SiRNA against Poldip2 blocks LPS-induced Cox-2 induction. a Confluent monolayers of RBMVECs were treated with LPS for 0, 30, 60, 120, and 180 min before cells were lysed and immunoblotting for Cox-2 was performed. β-tubulin was used as a loading control. Representative blots are shown. The graph depicts Cox-2 expression normalized to β-tubulin. Bars represent mean ± SEM of three independent experiments. Two-way ANOVA, ***p < 0.001 vs. 0 min, **p < 0.01 vs. 0 min, *p < 0.05 vs. 0 min. b Cells transfected with siRNA against Poldip2 or control were treated with LPS or PBS for 1 h before cells were lysed and immunoblotting for Cox-2 was performed. β-tubulin was used as a loading control. Representative blots are shown. The graph depicts Cox-2 expression normalized to β-tubulin. Bars represent mean ± SEM of three independent experiments. Two-way ANOVA, **p < 0.01 vs. siCtl, ^##p < 0.01 vs. siCtl + LPS

Fig. 7

Poldip2 silencing and selective inhibition of Cox2 block LPS-induced endothelial permeability. RBMVECs were grown on transwell inserts and FITC-dextran diffusion into the lower chamber was quantified spectrophotometrically to assess changes in permeability. a RBMVECs were transfected with siPoldip2 or siCtl, seeded on transwell inserts, and treated with LPS (1 μg/mL) or PBS for 3 h before incubation with FITC-dextran. The graph depicts FITC-dextran concentration expressed as fold change relative to siCtl + PBS. Bars represent mean ± SEM of six independent experiments. Two-way ANOVA, ****p < 0.0001 vs. siCtl + PBS, ^###p < 0.001 vs. siCtl + LPS. b RBMVECs were plated on transwell inserts and treated with LPS (1 μg/ml), meloxicam (10 μΜ), or both LPS and meloxicam simultaneously for 3 h before incubation with FITC-dextran. The graph depicts FITC-dextran concentration expressed as fold change relative to Ctl + PBS. Bars represent mean ± SEM of four independent experiments. Two-way ANOVA, ****p < 0.0001 vs. Ctl + PBS, ^###p < 0.001 vs. Ctl + LPS

Fig. 8

Poldip2 mediates LPS-induced BBB permeability via Cox-2. a Schematic of the experimental design. Evans blue dye was used to assess the effect of meloxicam on LPS-induced BBB permeability in Poldip2^+/+ and Poldip2^+/− mice. Animals received PBS or LPS (18 mg/kg, IP) for 18 h, meloxicam (5 mg/kg) subcutaneously at time 0 and 10 h later, or both LPS and meloxicam. Evans blue dye was administered by intravenous injection and allowed to circulate for 10 min before animals were sacrificed. b The graph depicts Evans blue dye concentration extracted from whole brains normalized to dry brain weight. Bars represent mean ± SEM. Two-way ANOVA, ****p < 0.0001 vs. Poldip2^+/+ + PBS, ^###p < 0.001 vs. Poldip2^+/+ + LPS, ^####p < 0.0001 vs. Poldip2^+/+ + LPS, n = 3–7 animals/group