Primary human brain microvascular endothelial cells (HBMECs) from CSC systems were purchased from ScienCell Research Laboratories, Inc. They were maintained in EGM-2 Endothelial Cell Growth Medium-2 Bullel kit (Lonza Group, Ltd.) on a Collagen type 1 (Corning, Inc.)-coated 10 cm diameter plate in a 37°C humidified atmosphere of 95% air and 5% CO₂ in incubator and used for experiments at 80% confluency.

THP-1 was purchased from ZQ Cell Research and cultured in RPMI1640 (GE Healthcare) medium supplemented with 0.05 mM β-Mercaptoethanol (Sigma-Aldrich; Merck KGaA), 10% FBS (Gibco; Thermo Fisher Scientific, Inc.) and 1% penicillin streptomycin (Thermo Fisher Scientific, Inc.). Cells were grown in a humidified atmosphere of 5% CO₂ at 37°C and used for experiments at >80% confluency.

Resveratrol was purchased from Sigma-Aldrich; Merck KGaA and a 50 mM storage solution in DMSO was prepared. Per experimental needs, cells were treated with 20 µM Resveratrol for 48 h before further analysis. Sirtinol (Sigma-Aldrich; Merck KGaA) was made into 20 mM storage solution by DMSO and diluted into a 20 µM working solution for cell treatment. For tumor necrosis factor (TNF)-α treatment, cells were treated with 10 ng/ml TNF-α (eBioscience; Thermo Fisher Scientific, Inc.) for 6 h before different assays.

SiGENOME Human ABCD1 (cat. no. M009605010005) small interfering (si)RNA-SMARTpool and siGENOME non-targeting siRNA (cat. no. D0012100105) were purchased from GE Healthcare Dharmacon, Inc. HSD17B4siRNA-Homo-427/990/1870 and KLF4 siRNA were purchased from Shanghai GenePharma Co., Ltd. Silencing was performed according to the protocol provided by the manufacturer. Briefly, 1.5-2×10⁵ HBMECs were seeded in 6-well plates for 24 h. siRNA (5 µM) dissolved in serum free medium was mixed with DharmaFECT transfection reagent (GE Healthcare Dharmacon, Inc.) and incubated for 20 min at room temperature. Cells were then replaced with fresh medium containing 25 nM siRNA and then cultured for 48 h for further analysis.

Total RNA was extracted with Qiagen RNeasy kit. First-strand cDNA was synthesized (25°C, 10 min; 42°C, 30 min; 85°C, 5 min) using Hiscript QRT SuperMix (Thermo Fisher Scientific, Inc.) in 20 µl reaction. Thermocycling conditions were 95°C for 10 min and 40 PCR cycles (95°C, 20 sec; 55°C, 30 sec; 72°C, 20 sec). After the amplification reaction, the temperature was 95°C for 5 sec and 65°C for 60 sec. PCRs were performed using SYBR master mix (Thermo Fisher Scientific, Inc.) and the data were analyzed by calculating the ∆Cq value (24). The primers used were listed as below: Claudin 5 forward: 5′-CTC TGC TGG TTC GCC AAC AT-3′, and Claudin 5 reverse: 5′-CAG CTC GTA CTT CTG CGA CA-3′; β-actin forward: 5′-CGA GCA CAG AGC CTC GCC TTT GCC-3′ and β-actin reverse: 5′-TGT CGA CGA CGA GCG CGG CGA TAT-3′.

HBMEC protein was extracted with radioimmunoprecipitation assay (RIPA) buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate and 0.1% SDS). After lysis on an ice for 0.5 h, samples were collected and centrifuged at 4°C at 20,817 g for 0.5 h. The supernatant was transferred into fresh tubes and protein concentration was determined by the bicinchoninic acid (BCA) method. Nuclear Protein was extracted using Nucleoprotein Extraction kit from Beyotime Institute of Biotechnology according to the manufacturer's protocol. Briefly, cells were lysed with RIPA buffer on ice for 0.5 h, which were collected and centrifuged at 4°C at 20,817 × g for 0.5 h. The supernatant was removed. After adding 100 µl buffer C, the precipitation was centrifuged at 4°C at 20,817 × g for 0.5 h. The supernatant was transferred into fresh tubes for next step.

Equal amounts of protein (60 µg/lane) were separated by 10% SDS-PAGE and transferred onto nitrocellulose filter membrane (EMD Millipore). Membranes were blocked with 5% non-fat milk in phosphate-buffered saline containing 0.05% Tween-20 at room temperature for 2 h and probed with antibodies at 4°C for 14 h including Anti-ABCD1 (1:1,000; OriGene Technologies, Inc.; cat. no. TA803208), and anti-hydroxysteroid (17B4) dehydrogenase 4 (1:2,000; Abcam; cat. no. ab128565), anti-SIRT1 (1:1,000; Santa Cruz Biotechnology, Inc.; cat. no. sc15404), anti-intracellular adhesion molecular (ICAM1; 1:1,000; Santa Cruz Biotechnology, Inc.; cat. no. sc7891), anti-vascular cellular adhesion molecule (VCAM1; 1:1,000; Santa Cruz Biotechnology, Inc.; cat. no. sc13160), anti-Claudin 5 (1:2,000; Abcam; cat. no. ab15106), anti-KLF4 (1:2,000; Proteintech; cat. no. 11880), anti-phosphorylated NF-κB inhibitor (p-IκBα; 1:2,000; Abcam; cat. no. ab133462), anti-NF-κB (1:2,000; Abcam; cat. no. ab16502) and anti-Acetyl-Lysine (1:2,000; ABclonal; cat. no. A2391). Anti-β-actin (1:5,000; Santa Cruz Biotechnology, Inc.; cat. no. sc47778) was used as a protein loading control while anti-Lamin b (1:5,000; Santa Cruz Biotechnology, Inc. cat. no. sc56143) was elected as nuclear protein loading control. Membranes were developed with Odyssey Infrared laser scanning image-forming system after being washed and incubated with appropriate secondary antibodies (1:2,000; horseradish peroxidase, Goat anti mouse or Goat anti rabbit, Vicmed Company; cat. no. C30409) at room temperature for 2 h.

The protein lysates, which obtained as aforementioned, were incubated with 1 µg IgG (Beyotime Institute of Biotechnology) and 20 µl resuspended Protein A/G PLUS-Agarose (Santa Cruz Biotechnology, Inc.) for 30 min at 4°C, then centrifugation was used to remove the non-specific protein adhering to protein A/G at 4°C at 2,500 × g for 5 min. After collecting and measuring the concentration of protein by the bicinchoninic acid method, the supernatants were made up to a total amount of 500 µl with IP buffer. The corresponding primary antibody with 5 µl was added to the samples including anti-KLF4 (ProteinTech Group, Inc.; cat. no. 11880), or anti-SIRT1 (Santa Cruz Biotechnology, Inc.; cat. no. sc-15404) or IgG (Beyotime Institute of Biotechnology) which as a contrast, followed by incubation overnight at 4°C. Subsequently, the samples were incubated with 100 µl Protein A/G PLUS-Agarose at 4°C for 4 h. Finally, the result was produced by western blot analysis.

For the adhesion assay, HBMECs were plated in 6-well plates at a density of 5×10⁴ cells/well until 80-90% confluent, with certain groups treated with tumor necrosis factor (TNF)-α (10 ng/ml) for 6 h. Furthermore, 1×10⁵ THP-1 cells/well were incubated with 1 µl Calcein AM (Sigma-Aldrich; Merck KGaA; 4 mM) at a dilution of 1:1,000 for 1 h at 37°C. Subsequently, the Calcein AM-labeled THP-1 cells were seeded onto the endothelial monolayer at a density of 1×10⁵ cells/well and incubated with RPMI-1640 medium plus 10% FBS for 1 h at 37°C. The cells were washed with PBS and fixed in 4% paraformaldehyde at room temperature for 20 min prior to imaging. Images were captured in five randomly selected microscopic fields at a ×10 magnification using an inverted fluorescence microscope (TH4-200; Olympus Corp.) and fluorescence was quantified using ImageJ software (Version no. 1.4.3.67, National Institutes of Health).

The Amplite™ Fluorimetric NADP/NADPH Ratio Assay kit was purchased from AAT Bioquest. The assay was performed according to the manufacturer's protocol. In brief, HBMECs were plated at a density of 1×10⁵ cells/well in 6-well plates and allowed to attach for 24 h. After removing the medium, 100 µl NADP/NADPH lysis buffer (AAT Bioquest, Inc.) was added to each well, followed by incubation at room temperature for 15 min. The supernatant was collected by centrifugation for subsequent analysis at room temperature at 2,000 × g for 10 min. The fluorescence intensity was measured using a microplate reader (Gene Company, Ltd.) with excitation and emission wavelengths of 530 and 590 nm, respectively.

The MDA Assay kit and the SOD assay kit were purchased from NJJC Institute of Bioengineering. First, HBMECs were plated at a density of 1×10⁵ cells/well in 6-well plates and allowed to attach for 24 h, followed by treatment with or without TNF-α (10 ng/ml) for 6 h. Next, the cells were collected and the corresponding reagents were added according to the manufacturer's protocols. The absorbance was measured using a microplate reader (Gene Company, Ltd.) at 550 nm for SOD and 532 nm for MDA, respectively.

All the results in this study were analyzed by SPSS 16.0 statistical software; GraphPad Prism 6.0 software (GraphPad Software, Inc.) was used for the generation of graphs. Values are expressed as the mean ± standard deviation. Statistical analyses were performed using one-way ANOVA with Tukey multiple comparison test to determine statistical significance between the groups. P<0.05 was considered to indicate a statistically significance difference.

In the first set of experiments, the direct impact of ABCD1 and HSD17B4 deficiency in HBMECs was investigated. Several key protein markers involved in endothelial interactions and permeability to leukocytes, including tight-junction proteins and adhesion molecules, as well as changes in SIRT1 expression levels prior to and after activation with TNF-α (10 ng/ml), were measured. As presented in Fig. 1, silencing of ABCD1 led to a significant 35% reduction in Claudin 5 (P<0.05), a 38% reduction in SIRT1 (P<0.05) as well as a 41% increase in ICAM1 (P<0.05) and a 33% increase in VCAM1 (P<0.05), while stimulation with TNF-α lead to further decreases in Claudin 5 (53%, P<0.001) and SIRT1 (47%, P<0.001), as well as a 2 and 2.8-fold increase in ICAM1 (P<0.001) and VCAM1 (P<0.001), respectively. Similar results were observed following HSD17B4 silencing, which led to a ~25% reduction in Claudin 5 (P<0.05), a 36% increase in ICAM1 and a ~25% increase in VCAM1, accompanied by a 39% reduction in SIRT1 (P<0.05) at the basal level without TNF-α stimulation. Upon TNF-α stimulation, a 57% reduction of Claudin 5 (P<0.001), a 58% reduction of SIRT1 (P<0.001), a 2.1-fold increase of ICAM1 (P<0.001) and a 2.2-fold increase of VCAM1 (P<0.001) were observed. Of note, inhibition of SIRT1 by treatment with Sirtinol (20 µmol/l) led to comparable alterations of the expression of adhesion molecules and Claudin 5 prior to and after endothelial activation (39/58% reduction in Claudin 5; 43/94% increase in ICAM1 and 36/300% increase of VCAM1 expression prior to/after stimulation with TNF-α, respectively). A increased higher expression of ICAM1 and VCAM1 and decreased Claudin 5 levels were detected in the groups with ABCD1 silencing plus Sirtinol and HSD17B4 silencing plus Sirtinol, but this was not statistically significant (Fig. 1).

It was then investigated whether increasing SIRT1 function using the SIRT1 activator resveratrol (20 µmol/l) attenuates the protein alterations associated with peroxisomal dysfunction caused by ABCD1 or HSD17B4 silencing. The targeting association was demonstrated by significant increases in SIRT1 protein levels after resveratrol treatment at a baseline and following TNF-α treatment (P<0.05). Resveratrol also partially normalized SIRT1 levels after ABCD1 and HSD17B4 silencing, particularly when HBMECs were activated with TNF-α (P<0.01). HBMECs treated with resveratrol exhibited a reduction of ICAM1 and VCAM1 and increased Claudin 5 expression under basal conditions (P<0.05) and following TNF-α treatment (P<0.01). Of note, resveratrol treatment of HBMECs with silencing of ABCD1 or HSD17B4 attenuated the overexpression of ICAM1 and VCAM1 and significantly increased Claudin 5 expression after activation with TNF-α (Fig. 2).

After determining the changes of molecular biomarkers, the functional consequences of low SIRT1 levels in HBMECs achieved by inhibition with Sirtinol or silencing of ABCD1 or HSD17B4. As previously described, a significant increase in adhesion of THP-1 cells was identified after ABCD1 silencing under basal conditions (P<0.05) and following TNF-α treatment (P<0.01; Fig. 3). Similarly, a significant increase in adhesion was observed after HSD17B4 silencing and Sirtinol treatment under basal conditions (P<0.05) and upon stimulation with TNF-α (P<0.01; Fig. 3).

To further assess whether restoration of SIRT1 levels in ABCD1- and HSD17B4-silenced cells is able to attenuate their increased adhesion to monocytes that had been observed, silenced HBMECs were treated with resveratrol. First, resveratrol treatment alone significantly reduced monocyte adhesion to 48 and 41% under basal conditions (P<0.05) and following TNF-α treatment (P<0.01), respectively. Furthermore, in ABCD1-silenced HBMECs, treatment with resveratrol markedly inhibited monocyte adhesion by 32 and 41% under basal and TNF-α treatment conditions, respectively (P<0.05). Similarly, in HSD17B4-silenced HBMECs, treatment with resveratrol also markedly inhibited monocyte adhesion by 31 and 40% at basal and TNF-α treatment conditions, respectively (P<0.05), suggesting functional correction by resveratrol treatment (Fig. 3).

As a NAD⁺-dependent protein deacetylase and the key metabolic sensor, SIRT1 tightly regulates numerous cellular activities involved in systemic metabolic homeostasis (16). In the present study, it was revealed that the NADP/NADPH ratio in HBMECs was significantly increased after ABCD1 or HSD17B4 silencing (P<0.01; Table I), suggesting a disturbance of metabolic homeostasis. By measuring oxidative stress makers MDA and SOD, it was unveiled that ABCD1 and HSD17B4 silencing led to significantly increased MDA levels and decreased SOD levels in HBMECs, suggesting increased oxidative stress. This effect was mimicked by sirtinol treatment, while resveratrol treatment significantly reduced the oxidative stress under basal and TNF-α treatment conditions (P<0.01; Table II).

In order to investigate the molecular mechanisms that mediate the effect of SIRT1 upon upregulation of adhesion molecules, activation of NF-κB, a transcriptional factor known to regulate ICAM1 expression, was then assessed. Indeed, SIRT1 depletion by sirtinol or silencing of ABCD1 or HSD17B4 in HBMECs led to increased IκBα phosphorylation and subsequent NF-κB accumulation in the nucleus under basal conditions (P<0.05; Fig. 4A and B) and following TNF-α treatment (P<0.01; Fig. 5A and B), suggesting activation of the NF-κB signaling pathway. When SIRT1 activity was enhanced by resveratrol treatment, a significant reduction of IκBα phosphorylation and nuclear NF-κB levels was observed in the control as well as in the ABCD1 or HSD17B4 gene silencing groups under basal conditions (P<0.01; Fig. 4C and D) and following TNF-α treatment (P<0.01; Fig. 5C and D), indicating that restoration of the function of SIRT1 is able to reverse NF-κB activation and significantly inhibit the expression of adhesion molecules, including ICAM1. A recent study reported that SIRT1 deacetylated KLF4 to activate Claudin 5 transcription in ovarian cancer cells (25). The present results indicated that ABCD1 and HSD17B4 silencing significantly reduced nuclear KLF4 levels under basal (P<0.05) and TNF-α treatment conditions (P<0.001), and the same effect was observed following SIRT1 inhibition using sirtinol (Figs. 4A and B; 5A and C). Increased SIRT1 activation by resveratrol significantly increased nuclear KLF4 levels in the control group and also restored the levels of KLF4 that were reduced by ABCD1 and HSD17B4 silencing under basal (P<0.05; Fig. 4C and D) and TNF-α treatment conditions (P<0.01; Fig. 5C and D). The direct interaction between SIRT1 and KLF4 and subsequent regulation of Claudin 5 expression was assessed by IP and silencing of KLF4. Reduction of SIRT1 binding to KLF4 protein and increase in KLF4 protein acetylation following ABCD1 and HSD17B4 silencing were confirmed (P<0.01; Fig. 6A-D). Finally, KLF4 silencing significantly inhibited Claudin 5 transcription in HBMECs (P<0.05; Fig. 6E and F).