1α,25-Dihydroxyvitamin D3 targeting of NF-κB suppresses TNF-α induced adhesion molecules expression in human endothelial cells

Background Vitamin D and its analogues have been documented to be associated with endothelial dysfunction in various diseases. However, the underlying mechanism remains unknown. Here, we conducted an in vitro study to evaluate the effect of 1α,25-dihydroxyvitamin D3, the active form of vitamin D, on adhesion molecules expression in human endothelial cells. The possible mechanism involved in this process was also explored. Methods Human umbilical vein cells (HUVECs) were cultured and treated according to the experiment requirement. Western Blot and RT-PCR were used to evaluate the expression of vascular cell adhesion molecule-1 (VCAM-1) and E-selectin. ChIP assay, immunouorescence, Western Blot and co-immunoprecipitation were used to assess the effect of 1α,25-dihydroxyvitamin D3 on NF-κB signaling. VCAM-1 and E-selectin and protein after TNF-α stimulation. ChIP assay that increased the p65 to the of D3. TNF-α induced IκBα phosphorylation and p65 NF-κB activation, leading to an inhibition of p65 nuclear translocation. These effects were reversed by a specic vitamin D receptor siRNA (VDR-siRNA). Co-immunoprecipitation revealed that 1α,25-dihydroxyvitamin D3 induced an increased binding of VDR to p65, which inhibited the ability of p65 binding to target gene promoters.

Background Cardiovascular (CV) diseases remain the leading cause of death in modern societies with endothelial dysfunction being the common pathway of various risk factors. As an important non-traditional risk factor for CV events, vitamin D and its analogues have been documented to be associated with endothelial dysfunction in a diverse group of people, including healthy humans [1,2], patients with chronic kidney disease (CKD) [3][4][5][6][7] [8], diabetes [9], lupus [10], stroke [11], obesity [12], and rheumatoid arthritis [13]. Although the physiological mechanisms linking vitamin D to endothelial dysfunction in these settings have not been established, it may involve vascular in ammation. During in ammatory process, the phenotype of endothelial cells turns to be activated [14]. And endothelial activation induces an upregulated expression of adhesion molecules, such as E-selectin and vascular cell adhesion molecule-1 (VCAM-1), which play a pivotal role in leukocyte-endothelium interactions, eventually leading to atherosclerosis and CV diseases.
Tumor necrosis factor-α (TNF-α) is one of the primary mediators of endothelial activation, which contributes to the in ammatory endothelial cells response and is initiated through activation of the classical NF-B pathway [15]. Vitamin D has been documented to have anti-in ammatory effect in various diseases. It is reported that in healthy women, serum 25(OH)D concentrations are negatively correlated with TNF-α concentrations [16]. Supplementation of paricalcitol, a vitamin D analogue, was associated with a reduction of serum levels of TNF-α in CKD patients [17]. The relationship between vitamin D and NF-B signaling has also been reported. Suzuki et al. documented in their study that in human coronary arterial endothelial cells, vitamin D inhibits activation of NF-B signaling pathway as well as the expression of its downstream target E-selectin [18]. However, it is still unknown which stage of NF-B pathway is affected by vitamin D in endothelial cells. Here, we conducted an in vitro study to evaluate the effect of 1α,25-dihydroxyvitamin D3, the active form of vitamin D, on TNF-α induced adhesion molecules expression in human endothelial cells. We also explored the effect of 1α,25dihydroxyvitamin D3 on various stages of NF-B pathway, including early activation, nuclear translocation and its binding to VCAM-1 and E-selectin promoters, to provide additional insight into the molecular mechanisms linking vitamin D to endothelial function.

Results
Treatment of HUVECs with 1α,25-(OH) 2 D 3 inhibits TNFαinduced VCAM-1 and E-selectin release As shown in Figure 1, when HUVECs were stimulated with TNF-α 40ng/ml for 24 h, the expression of VCAM-1 and E-selectin were signi cantly increased. Pretreatment of HUVECs with various concentrations of 1α,25-(OH)2D3for 30min markedly reduced VCAM-1 and E-selectin mRNA and protein levels, with the maximum reduction observed in the middle concentration (10 -8 M 1α,25-(OH)2D3) state. Therefore, vitamin D is able to suppress the induction of adhesion molecules by TNF-α in HUVECs. After transfected with a speci c VDR-siRNA, the inhibitory effect of 1α,25-(OH)2D3 on expression of adhesion molecules was reversed, demonstrating that the inhibition is VDR-dependent. The degree of VDR knockdown by VDR-siRNA was assessed in these cells at both mRNA and protein level. As shown in Figure 2, a strong reduction of VDR transcripts and protein expression was observed by RT-PCR and western blots.

NF-κB signaling pathway is critical for mediating HUVECs activation
It is reported that NF-κB signaling pathway plays an important role in the process of endothelial activation. To con rm this, HUVECs were incubated with TNF-α 40ng/ml in the absence or presence of a speci c NF-κB inhibitor (NF-κB SN50). As shown in Figure 3, inhibition of NF-κB signaling reduced TNF-α induced VCAM-1 and E-selectin expression in these cells, indicating that an intact NF-κB signaling is required for the activation of HUVECs.
We rst investigated whether 1α,25-(OH)2D3 can modulate the signaling events after p65 NF-κB nuclear translocation, in other words, p65 NF-κB binding to VCAM-1 and E-selectin promoters. As shown in Figure 4, ChIP assay demonstrated that 1α,25-(OH)2D3 abrogated the binding of p65 NF-κB to its cognate cis-acting element in VCAM-1 and E-selectin promoters in HUVECs. After transfected with a speci c VDR-siRNA, the inhibitory effect of 1α,25-(OH)2D3 on p65 DNA binding was reversed. This result indicates that 1α,25(OH)2D3 blunts TNF-α induced adhesion molecules expression by blocking the NF-κB binding activity.
Since IκBα is the major inhibitor of NF-κB that binds to p65 NF-kB and blocks its activation and subsequent nuclear translocation, we explored further the effect of 1α,25(OH)2D3 on IκBα phosphorylation, as well as p65 NF-κB phosphorylation and activation. As shown in Figure 6, when HUVECs were incubated with TNF-α 40 ng/ml, IκBα and p65 NF-κB were rapidly phosphorylated.

Effect of 1α,25-(OH)2D3 on the interaction between VDR and p65 NF-κB
It is reported that 1α,25-(OH)2D3 can upregulate the expression of VDR. Meanwhile, in human proximal tubular cells, vitamin D abolishes p65 NF-κB binding to RANTES promoter by facilitating VDR/p65 interaction. Our aforementioned results that effect of 1α,25-(OH)2D3 can be reversed by a speci c VDR-siRNA suggests that vitamin D exerts its bene cial effect at least partly through VDR. Thus, we further explored the effect of 1α,25-(OH)2D3 on VDR expression and the interaction between VDR and p65 NF-κB in HUVECs. As presented in Figure 7, 1α,25-(OH)2D3 upregulated the expression of VDR in HUVECs. Meanwhile, in HUVECs that overexpressed p65 NF-κB, increased p65 was detected in the cell lysates precipitated by anti-VDR antibody after stimulation by TNF-α and/or 1α,25-(OH)2D3 ( Figure 8). Hence, it is possible that an increased formation of VDR/p65 complex after 1α,25-(OH)2D3 treatment reduced the level of free p65, thereby affected its binding to promoters of targeting genes, resulting in a inhibition of p65-mediated gene transcription.

Discussion
This study demonstrated that 1α,25-(OH)2D3, the active form of vitamin D, suppresses TNF-α induced upregulation of VCAM-1 and E-selectin expression in HUVECs. Mechanically, 1α,25-(OH)2D3 targets NF-κB, a principal signaling pathway involved in the regulation of in ammatory reaction in various circumstances, in a VDR-dependent manner. 1α,25-(OH)2D3 can also upregulate the expression of VDR, promote the binding of VDR to p65, and thus inhibit the ability of p65 binding to VCAM-1 and E-selectin gene promoters. Since endothelial activation is a critical process that contributes to the pathogenesis of atherosclerosis and CV diseases, inhibition of endothelial activation may be an important mechanism by which vitamin D exerts its bene cial activity in ameliorating CV events.
As a fat-soluble secosteroid, vitamin D is responsible for calcium and phosphorus metabolism, and multiple other biological effects, including regulation of immune system and endothelial function. We have shown in our previous study that in CKD patients, hypovitaminosis D was associated with decreased brachial artery ow-mediated dilation (FMD) and increased VCAM-1 and E-selectin [7], and vitamin D supplementation can improve endothelial function [8]. The present study revealed that 1α,25-(OH)2D3 suppresses TNF-α induced high expression of VCAM-1 and E-selectin in HUVECs, which provides additional evidence for the endothelial protective effect of vitamin D through in vitro experiment. In agreement with our ndings, Martinesi et al [19]. demonstrated that 1α,25(OH)2D3 was able to reduce VCAM-1 level previously enhanced by TNF-α, though in their study this effect was not seen on E-selectin expression, which might be explained by the experimental conditions. In fact, in their experiments, HUVECs were incubated with the association of TNF-α and 1α,25(OH)2D3, adding the two compounds at the same time.
Another question worth to mention is the intervention concentration of 1α,25(OH)2D3. In this study we pretreated HUVECs with various concentrations of 1α,25-(OH)2D3 (10 -9 to 10 -7 M) and found that 1α,25(OH)2D3 at 10 -8 M had the strongest inhibitory effect on TNF-α induced adhesion molecules expression. Likewise, Kudo et al [20]. showed in their study that, in human coronary arterial endothelial cells (HCAECs), 1α,25(OH)2D3 at a concentration of 10 -8 M also had a slightly stronger suppressive effect on VCAM-1 expression after stimulation with TNF-α in contrast to 1α,25(OH)2D3 concentrations of 10 −7 M and 10 −9 M. Martinesi et al [19]. reported that in the absence of growth factors, 1α,25(OH)2D3 at 10 -8 M had a mildly stronger inhibitory effect on the expression of adhesion molecules in HUVECs when compare to 1α,25(OH)2D3 at 10 -7 M. Since the hormone alone cannot alter the expression of adhesion molecules [19], it's possible that a higher concentration (10 -7 M) may have severer inhibitory effects on HUVECs proliferation (as reported by Zehnder et al. [21]), and thus leads to a decreased expression of these molecules. Of note, there is also a study showing that 1α,25(OH)2D3 at 10 -7 M had more obvious inhibitory effect on E-selectin expression as well as NF-κB activation when compared to consistent with levels obtained in healthy human plasma after administration of a normal dose, that is 10 −9 to 10 −7 M as reported by literatures [22,23].
Inhibition of NF-κB signaling with NF-κB SN50 reduced TNF-α stimulated VCAM-1 and E-selectin expression in HUVECs, indicating that an intact NF-κB signaling is required for in ammatory cytokine induced endothelial activation. Previous investigations have reported that expression of NF-κB was greater in vascular endothelial cells of subjects with vitamin D de ciency [24]. Hence, it's possible that vitamin D exerts its inhibitory effect on endothelial activation by inhibiting NF-κB signaling pathway. It is well known that the activity of NF-κB includes IκBα and p65 NF-κB phosphorylation, p65 NF-κB nuclear translocation, and p65 NF-κB binding to target gene promoters. Thus, the activity of NF-κB can be regulated at multiple sites. The present study demonstrated that 1α,25-(OH)2D3 can affect p65 NF-κB binding to VCAM-1 and E-selectin promoters. Although the exact mechanism by which vitamin D disrupts this interaction remains unclear, our data suggest that, in HUVECs, at least part of the mechanism stems from an increase or stabilization of IkBα, thus abolishing p65 nuclear translocation. In agreement with our ndings, several studies reported that in mesangial cells, pancreatic islet cells and mouse embryonic broblasts, vitamin D can also inhibit NF-κB signaling through increasing IkBα and reducing p65 NF-κB nuclear translocation, and eventually abolishing its binding to gene promoters [25][26][27]. Based on this point, more studies are needed to elucidate how vitamin D regulates IkBα. However, it should be noted that in human proximal tubular cells, paricalcitol, a vitamin D analogue, in uences neither IkBα nor p65 nuclear translocation, but only abolishes the binding of p65 to target gene promoter [28]. Although the exact reason behind this discrepancy is still unknown, it could be related to different cell types.
As expected, in HUVECs transfected with a speci c VDR-siRNA, the inhibitory effect of 1α,25(OH)2D3 on VCAM-1 and E-selectin expression as well as NF-κB signaling pathway was abolished, indicating that VDR is required to mediate the repressive action of 1α,25(OH)2D3. Meanwhile, 1α,25-(OH)2D3 can upregulate the expression of VDR in HUVECs, which physically interacts with p65 and potentially blocks p65 binding to DNA. This increased VDR expression and VDR-p65 physical association may serve as another mechanism by which vitamin D destroys the binding of p65 to gene promoters. In fact, it was also reported in broblasts, human proximal tubular cells and osteoblasts [26, 28, 29], but not in mesangial cells [27], which might be explained by the cell-type speci city.
Though our data strongly indicate that 1α,25(OH)2D3 inhibits endothelial adhesion molecules expression primarily by triggering a VDR-mediated sequestration of NF-κB pathway, it remains to be elucidated whether this mechanism is applicable in vivo. Besides, 1α,25(OH)2D3 may regulate endothelial activation by other routes as well, given that vitamin D has pleiotropic effects.

Conclusions
In conclusion, we have shown in this study that 1α,25-(OH)2D3 can inhibit proin ammatory cytokine induced endothelial activation, and this potentially endothelial protective role seems to be mediated by its ability to induce the VDR-mediated sequestration of NF-κB signaling. In view of the importance of endothelial cell activation in the occurrence and development of CV events, the endothelial protective effect of 1α,25-(OH)2D3 may become a novel target for the prevention and treatment of such diseases.

Cell culture
Normal cryopreserved HUVECs (obtained from KeyGEN BioTECH, Nanjing, China) were thawed rapidly in 37-40 degrees water bath and grown in F12K medium (KeyGEN BioTECH, Nanjing, China) at 37 °C in a 5% CO2 humid incubator. When the cells reached 80% con uence, they were passaged using 0.25% Trypsin (KeyGEN BioTECH, Nanjing, China). Cells at passages 3 to 4 were used for experiments.
Transfection with VDR-siRNA

Western Blot Analysis
The preparation of whole-cell lysates and Western blot analysis of protein expression were carried out using routine procedures. Brie y, HUVECs after various treatments as indicated were washed with ice-cold PBS and lysed in the presence of Protease Inhibitor Cocktail (Sigma-Aldrich, USA) for 15 min in ice-cold lysis buffer. Cell extracts were centrifuged at 14,000 x g for 15 min in a 4°C pre-cooled centrifuge and the supernatant was stored at -70°C. For Western blot analysis, protein lysates were separated on 10% SDS-PAGE gels, transferred to nitrocellulose membrane and immunoblotted using standard protocols. Proteins were visualized using horseradish peroxidase detection reagents according to manufacturer's instructions (New-SUPER ECL, KeyGEN BioTECH, Nanjing, China), and exposed to autoradiographic lm (G:BOX chemiXR5). Gel-Pro32 software was used to grayscale the results. The primary antibodies: Anti-VCAM-1 (ab134047), anti-E-selectin (ab18981), anti-p65 NF-κB (ab32536), anti-phospho-p65 NF-κB (phospho S529) (ab109458), anti-IκBα (ab32518), anti-phospho-IκBα (phospho S36) (ab133462) and anti-VDR (ab109234), were all obtained from Abcam (Cambridge, UK).
ChIP assays NF-kB binding to VCAM-1 and E-selectin gene promoters in HUVECs was determined by ChIP assays using a commercially available ChIP assay kit (Lake Placid, NY, USA). Brie y, HUVECs after various treatments as indicated were treated with 1% formaldehyde to cross-link histones to DNA, and then treated with stop buffer for 5 minutes to stop the cross-linking. The chromatin was extracted and fragmented by sonication. The sonicated chromatins were incubated with anti-p65 antibody overnight at 4°C, followed by incubation with protein A-agarose for 2 h. The precipitates were washed, and chromatin complexes were eluted. After reversal of the cross-linking, the DNA was puri ed, and used as templates for PCR using the primers anking the NF-kB binding sites in VCAM-1 and E-selectin gene promoters. The sequences of primers used for ChIP assay were as follows: VCAM-1, 5'-GAGGAGCAGGTAGGACTT -3' (forward), and 5'-CTGAGGTCTGGAATCTATAACT -3' (reverse); and E-selectin, 5'-GCCTCTCACCTCAGCCTTGTAG-3' (forward), and 5'-ACATTGTGCCAACATCAGTATCCT-3' (reverse). The PCR products were run on 1.5% agarose gel and stained with ethidium bromide.
Immunostaining HUVECs were pretreated with or without 10 -8 M 1α,25(OH)2D3 for 24 h and then exposed to TNF-α 40ng/ml for 1h. The cells were xed with 4% paraformaldehyde for 30 min, and incubate with normal goat serum for 20 min to block unspeci c binding of the antibodies. After that, cells were stained with anti-p65 antibody for 2h in a humidi ed chamber, followed by staining with uorescein isothiocyanate (FITC)-conjugated secondary antibody (Abcam, Cambridge, UK) for 1h in the dark. Each slide was stained with 4-diamidino-2-phenylindole (DAPI) for 5min to visualize the nucleus. Slides were viewed with an Olympus IX51 microscope equipped with a digital camera (Japan).