20-Hydroxycholecalciferol, Product of Vitamin D3 Hydroxylation by P450scc, Decreases NF-κB Activity by Increasing IκBα Levels in Human Keratinocytes

The side chain of vitamin D3 is hydroxylated in a sequential manner by cytochrome P450scc (CYP11A1) to form 20-hydroxycholecalciferol, which can induce growth arrest and differentiation of both primary and immortalized epidermal keratinocytes. Since nuclear factor-κB (NF-κB) plays a pivotal role in the regulation of cell proliferation, differentiation and apoptosis, we examined the capability of 20-hydroxycholecalciferol to modulate the activity of NF-κB, using 1,25-dihydroxycholecalciferol (calcitriol) as a positive control. 20-hydroxycholecalciferol inhibits the activation of NFκB DNA binding activity as well as NF-κB-driven reporter gene activity in keratinocytes. Also, 20-hydroxycholecalciferol induced significant increases in the mRNA and protein levels of the NF-κB inhibitor protein, IκBα, in a time dependent manner, while no changes in total NF-κB-p65 mRNA or protein levels were observed. Another measure of NF-κB activity, p65 translocation from the cytoplasm into the nucleus was also inhibited in extracts of 20-hydroxycholecalciferol treated keratinocytes. Increased IκBα was concomitantly observed in cytosolic extracts of 20-hydroxycholecalciferol treated keratinocytes, as determined by immunoblotting and immunofluorescent staining. In keratinocytes lacking vitamin D receptor (VDR), 20-hydroxycholecalciferol did not affect IκBα mRNA levels, indicating that it requires VDR for its action on NF-κB activity. Comparison of the effects of calcitrol, hormonally active form of vitamin D3, with 20-hydrocholecalciferol show that both agents have a similar potency in inhibiting NF-κB. Since NF-κB is a major transcription factor for the induction of inflammatory mediators, our findings indicate that 20-hydroxycholecalciferol may be an effective therapeutic agent for inflammatory and hyperproliferative skin diseases.

Inflammation and proliferation are regulated by a plethora of transcription factors, with nuclear factor-kB (NF-kB) considered to be a master regulator of these processes (reviewed in [33]). NF-kB is also important in the development, prevention and therapy of cancer [34,35,36]. NF-kB activity is stimulated by many pathways that converge on IkB kinases, including the signaling pathways activated by various cytokines, such as the proinflammatory cytokine IL-1 (reviewed in [37,38]), lipopolysaccharide (LPS) and tumor necrosis factor a (TNF-a) [39,40]. In mammals, the NF-kB family of proteins includes NF-kB1 (p105 processed to p50), NF-kB2 (p100 processed to p52), RelA (p65), RelB and cRel [41]. Phosphorylation and subsequent degradation of IkB proteins allow for translocation of cytoplasmic NF-kB into the nucleus, where NF-kB binds to specific promoter/enhancer elements to regulate the expression of specific genes [33]. NF-kB regulated genes play important roles in inflammation, immunity, cell growth and cell survival [42,43,44,45].
NF-kB activation is mediated through the activation of specific IkB kinases (IKKs) and the subsequent phosphorylation of IkB. The pathway leading to proteolysis of IkB is denoted as the canonical NF-kB activation pathway. NF-kB activation also occurs through the 'noncanonical' pathway, which does not involve IkB degradation and is activated by various agents, including interferon-a/b, lipopolysaccharide, the LMP1 protein of Epstein-Barr virus, B-cell activating factor and lymphotoxin-b [42,46].
Vitamin D and various synthetic vitamin D analogues have been widely used in the treatment of psoriasis [47] and other inflammatory/hyperproliferative skin disorders [48,49]. The cellular actions of 1,25(OH) 2 D3, the bioactive form of vitamin D, are not fully understood, but its effects have traditionally been ascribed to its binding to the vitamin D receptor (VDR) [50,51,52,53]. NF-kB plays an important role in protecting keratinocytes against apoptosis during programmed cornification [54]. In normal human keratinocytes, 1,25(OH) 2 D3 reduces NF-kB DNA binding activity by increasing IkBa protein levels, which inhibits IL-8 production [55]. A similar effect is also seen in murine macrophages [56,57]. Effects of 1,25(OH) 2 D3 on NF-kB that are not mediated by the VDR have also been reported for fibroblasts lacking the VDR [58].
In the present study we have examined the effects of 20(OH)D3 on NF-kB signaling in comparison to well defined effects of 1,25(OH) 2 D3. Since NF-kB dysregulation induces malignant transformation of HaCaT keratinocytes, but not of normal keratinocytes [59], we used in these studies both immortalized human HaCaT keratinocytes and primary epidermal keratinocytes, isolated from human neonatal foreskin (HEKn). The effects of 1,25(OH) 2 D3 on both expression of genes involved in its metabolism and the biological activity of the encoded proteins have previously been studied in these cells [26,60]. Recent data from our laboratories indicates that 20(OH)D3 can be produced by adrenal mitochondria [7], that adrenal glands ex-vivo can transform 7DHC to 5,7-diene products [61] that in the skin can be converted to biologically active vitamin D-like products [62]. Therefore, action of 20(OH)D3 on NF-kB activity would suggest a role for a novel endogenous secosteroidogenic metabolic pathway [7,8,9,10,11] in the regulation of the systemic and cutaneous immune activity.

20-Hydroxycholecalciferol inhibits NF-kB DNA binding activity in keratinocytes
In initial experiments we determined that 20(OH)D3 at 100 nM was optimal for inducing biological actions like stimulation of keratinocytes differentiation and inhibition of cell proliferation [15]. Next we examined the effect of 20(OH)D3 on NF-kB activity in keratinocytes by assaying nuclear extracts of 20(OH)D3-stimulated cells by DNA-binding assays. Primary human keratinocytes and HaCaT cells were incubated with 100 nM 20(OH)D3, nuclear extracts were prepared and incubated with an NF-kB oligonucleotide probe based on the kB binding site in the immunoglobulin light chain enhancer. As shown in figure 2A and 2C, a time dependent decrease in nuclear protein binding to the kB response element was observed in extracts from 20(OH)D3-treated cells. Inhibition of NF-kB activity was observed within 30 minutes of 20(OH)D3 addition. Maximum inhibition was reached by 4 hours, and inhibition persisted up to 24 hours. This effect was no longer observed after 48 hours. Treatment with 1,25(OH) 2 D3 also had similar inhibitory effect of NF-kB activity (data not shown). High basal NF-kB activity in HaCaT cells is probably due to serum deprivation of cells, since previously we have demonstrated that serum deprivation triggers NF-kB activation in HaCaT cells [63]. In order to determine the composition of 20(OH)D3-induced NF-kB complexes, nuclear extracts were preincubated with antibodies against the p65 and p50 NF-kB proteins and analyzed by supershift assays. As shown in figure 2B and 2D, the 20(OH)D3-induced complex contains both p50 and p65 proteins. The specificity of the binding to the kB probe was determined by incubating nuclear extract with excess unlabeled (cold) NF-kB oligonucleotide. Since excess unlabeled NF-kB oligonucleotide competed out DNA binding to the kB probe, NF-kB binding was considered specific.

20-Hydroxycholecalciferol inhibits NFkB-driven reporter gene activity in keratinocytes
In order to determine functional consequences of the decreased NF-kB DNA binding activity in the keratinocytes treated with 20(OH)D3, we performed gene reporter assays to determine NF-kB driven transcriptional activity (Fig. 3). HaCaT and normal human keratinocytes were transiently transfected with the pNFkB-Luc construct, which contained the firefly luciferase reporter gene driven by NF-kB. In the presence of 1,25(OH) 2 D3 or 20(OH)D3, basal luciferase activity decreased (Fig. 3). The inhibitory effect was more pronounced in normal human keratinocytes, with approximately a 2.5-fold decrease in the reporter activity (p,0.01) (Fig. 3A). In immortalized keratinocytes (HaCaT) the decrease in activity was less pronounced, but was also statistically significant (p,0.05) (Fig. 3A). 20(OH)D3 and 1,25(OH) 2 D3 had similar potency in inhibiting the NF-kB driven reporter in keratinocytes. Interestingly, NF-kB activity was significantly inhibited even after 24 hours of treatment with either agent (Fig. 3A).
To further characterize the inhibitory activity of 20(OH)D3 and 1,25(OH) 2 D3, NF-kB activity in HaCaT and normal human keratinocytes was stimulated by two agents known to induce NF-kB activity, LPS or IL-1a. As shown in Figure 3B, both LPS and IL-1a increased NF-kB-driven luciferase activity in HaCaT cells and normal human keratinocytes, as compared to cells treated with vehicle (,0.01% ethanol). We next examined the effects of 20(OH)D3 or 1,25(OH) 2 D3 on luciferase activity in HaCaT cells stimulated with LPS or IL-1a. Treatment with 20(OH)D3 or 1,25(OH) 2 D3 resulted in a statistically significant (p,0.05) decrease in NF-kB-driven luciferase expression in HaCaT cells stimulated by LPS or IL-1a with 20(OH)D3 and 1,25(OH) 2 D3 exhibiting similar potencies in inhibiting NF-kB activity. We than analyzed luciferase activity in cell extracts from human epidermal keratinocytes (HEKn), treated with 20(OH)D3 or 1,25(OH) 2 D3 and stimulated with LPS or IL-1a. Interestingly, the inhibition by 20(OH)D3 or 1,25(OH) 2 D3 of NF-kB activity was greater when the keratinocytes were stimulated with LPS as compared to IL-1a. 20(OH)D3 was slightly less potent in inhibiting NF-kB activity in keratinocytes when compared to 1,25(OH) 2 D3. Thus, despite the cell-type differences in the stimulation of NF-kB-dependent transcription activity by LPS versus IL-1a, 20(OH)D3 and 1,25(OH) 2 D3 inhibited NF-kB-dependent transcription.

20-Hydroxycholecalciferol inhibits translocation of the p65 NFkB protein induced by IL-1a in keratinocytes
To further characterize the inhibitory effect of 20(OH)D3 on NF-kB activity, we examined the cellular localization of the p65 NF-kB protein and the IkBa inhibitory protein in keratinocytes  were treated with 20(OH)D3 for 1, 4 or 24 hours (data not shown).

20-hydroxycholecalciferol increases IkBa protein levels in keratinocytes
Since we demonstrated by various assays (EMSA, gene reporter assays, and immunofluorescence assays) that 20(OH)D3 inhibits NF-kB activity, we next examined the underlying mechanism responsible for this activity. In the classical NF-kB pathway, NF-kB activity is sequestered it in the cytoplasm by forming a complex with inhibitory IkB proteins. Moreover, as shown in figure 4 20(OH)D3 appears to increase cellular IkB levels as determined by immunofluorescent staining. To determine whether 20(OH)D3 affects the classical NF-kB pathway, the cellular levels of IkBa and the p65 NFkB were determined at various times after 20(OH)D3 addition to cells. 20(OH)D3 induced a time-dependent increase in IkBa levels in whole cell extracts of HEKn (Fig. 5A) and HaCaT keratinocytes (Fig. 5C). IkBa was increased within 1 hour of 20(OH)D3 treatment, and by 16 hours IkBa was diminished to the levels observed in untreated cells. Similar results were obtained when cells were treated with 1,25(OH) 2 D3. In contrast, cellular levels of p65 was unaffected by 20(OH)D3 treatment of keratinocytes. As shown figure 5B and 5D, statistically significant changes were observed for IkBa levels induced by 20(OH)D3 and 1,25(OH) 2 D3 expressed relative to b-actin (p,0.05).
To further characterize the ability of 20(OH)D3 to inhibit NF-kB activity we stimulated NF-kB activity in normal human keratinocytes with IL-1a and determined IkBa levels in cytosolic extracts. We found that the concentration of IkBa levels were increased after treatment with 20(OH)D3 for 1 and 4 hours (Fig. 6A). Treatment of cells with 20(OH)D3 without IL-1a stimulation had a similar effect on IkBa levels. As shown figure 6B, statistically significant changes were observed for IkBa levels induced by 20(OH)D3 expressed relative to b-actin (p,0.05).

20-Hydroxycholecalciferol stimulates IkBa mRNA expression, but does not affect NF-kB mRNA expression in keratinocytes
To determine whether the increased IkBa protein levels in cells treated with 20(OH)D3 resulting from increased IkBa mRNA expression, we measured IkB mRNA levels by quantitative real time PCR (qPCR). As shown in figure 7 the IkBa-mRNA levels were significantly increased after 20(OH)D3 treatment of HaCaT and normal human keratinocytes. The induction by 20(OH)D3 of IkBa mRNA expression was greater in HaCaT cells than in normal keratinocytes. The effect was already detected 1 hour after treatment and returned to basal levels by 24 hours in normal keratinocytes, while the induction of IkBa mRNA persisted up to 24 hours in HaCaT cells. Moreover, the effects of 1,25(OH) 2 D3 on IkBa mRNA levels were qualitatively similar to those noted for 20(OH)D3. In contrast, mRNA levels of the p50 and p65 NF-kB subunits were unaffected by treatment with either 1,25(OH) 2 D3 or 20OHD3 (data not shown).

20-hydroxycholecalciferol requires VDR expression for its action on the NF-kB pathway in keratinocytes
We previously demonstrated that the action of 20(OH)D3 on human keratinocytes is dependent on VDR expression [15]. Therefore, we examined whether the effect of 20(OH)D3 on the NF-kB pathway was also VDR-dependent. Human keratinocytes were transiently transfected with siRNA to knock-down VDR expression, treated with 20(OH)D3 or vehicle (ethanol) and RNA isolated for gene expression analysis by qPCR In parallel experiments cell extracts were analyzed for protein expression by western blot. As shown in figure (Fig. 8B). The mRNA levels for p50 and p65 NF-kB proteins were unaffected by VDR knockdown (data not shown for p50). To further test the ability of VDR knockdown on NFkB translocation, we transfected cells with scrambled or VDR siRNA and treated them with 20(OH)D3, and than examined the cellular localization of the p65 NF-kB protein and the IkBa inhibitory protein in keratinocytes by fluorescent microscopy. In summary, significantly less IkBa protein was localized in the cytoplasm after 20(OH)D3 treatment of VDR knockdown cells as compared with scrambled siRNA-treansfected cells (Fig. 8D). 20(OH)D3 treatment of cells nearly completely blocked the nuclear translocation of p65. In contrast, 20(OH)D3 treatment  of VDR knockdown cells did not block the nuclear translocation of p65 ( Figure 8E).

Discussion
We have previously shown that 20(OH)D3 is a product of vitamin D3 metabolism by cytochrome P450scc (see Fig. 1) [7,10]. Moreover, 20(OH)D3 has significant biological activity in human keratinocytes, as it inhibits their proliferation and stimulates their differentiation [15]. In the present study we demonstrate that 20(OH)D3 is a potent inhibitor of NF-kB activity. Moreover, 20(OH)D3 treatment also increases IkBa protein levels through induction of IkBa mRNA expression. IkBa induction by 20(OH)D3 requires VDR expression, indicating that 20(OH)D3 acts through the classical vitamin D and NFkB regulatory pathways.
The inhibitory effect of 20(OH)D3 on NF-kB activity in keratinocytes was shown by several complementary approaches including NF-kB dependent DNA binding assays, NFkB-driven reporter gene activity assays, as well as western blotting and immunofluorescence analysis of the translocation of p65 NF-kB subunit from cytoplasm into the nucleus. The inhibitory effect of 20(OH)D3 was rapid (within 30 minutes), reached a maximum by 4 hr after addition, and remained detectable as long as 24 hr after addition. The inhibitory effect of 20(OH)D3 on NF-kB-dependent transcriptional activity by luciferase reporter-gene analysis (Fig. 3) paralleled the time course of 20(OH)D3 inhibition of NF-kBdependent DNA binding activity by EMSA (Fig. 2). The inhibitory effect of 20(OH)D3 on NF-kB was greater in normal human keratinocytes as compared to the effect in HaCaT keratinocytes. This discrepancy might be secondary to the immortalization of HaCaT cells, which might render them less sensitive to 20(OH)D3 treatment. Moreover, 20(OH)D3 had similar potency to that of the well characterized 1,25(OH) 2 D3 in inhibiting NF-kB activity in keratinocytes (no statistically significant difference). The vitamin D analogs, 1,25(OH) 2 D3 and 1,24(OH) 2 D3, have been previously reported to inhibit NF-kB activity [57]. Also, 1,25(OH) 2 D3 has been previously shown to regulate NF-kB DNA binding activity in human keratinocytes through an increase in IkBa expression [55]. In this study, 1,25(OH) 2 D3 inhibited NF-kB binding to the IL-8 kB binding sequence more potently than binding to the p53 kB binding sequence. This selectivity may be mediated through an increased IkBa expression, indicating that vitamin D analogues may exert their immunomodulatory effects through NF-kB regulated proinflammatory cytokines and chemokines. In our study we tested the effect of the novel analog of vitamin D3, 20(OH)D3, not only on NF-kB activity, but also on protein and mRNA levels, as well as the role of VDR in the effect of 20(OH)D3 on NF-kB. We clearly demonstrate that 20(OH)D3 inhibits NF-kB activity with potency similar to that of calcitriol (1,25(OH) 2 D3, the endogenous active form of vitamin D3). The mechanism of action or 20(OH)D3 appears to be very similar to that of 1,25(OH) 2 D3. The hydroxyl group of 20(OH)D3 is attached at the C20 position [7]; which is interesting since the attachment at C1 is considered to be required for full biological activity and calcemic effects [2,5].
To further characterize the action of 20(OH)D3 on inhibiting NF-kB activity we used known stimulators of the NF-kB pathway, LPS and IL-1a [64]. Keratinocytes can produce a plethora of cytokines including interleukin (IL)-1 and tumor necrosis factor a (TNF) (reviewed in [37]). IL-1 activates keratinocytes and promotes their proliferation and migration. Also, LPS is considered as a potent NF-kB stimulator [65]. In the present study, 20(OH)D3 was found to attenuate NF-kB transcriptional activity induced by both LPS or IL-1 in HaCaT cells and primary keratinocytes, and exhibited a similar potency to 1,25(OH) 2 D3. Interestingly, in normal keratinocytes the inhibition of NF-kB activity stimulated with IL-1a by vitamin D3 hydroxyderivatives was less pronounced then the inhibition of activity stimulated with LPS. This indicates that immortalization changes the responsiveness of keratinocytes to various stimuli, as it has been demonstrated in case of neuropeptides [63,66]. Nonetheless, both 20(OH)D3 and 1,25(OH) 2 D3 inhibit NF-kB activity in keratinocytes induced by recognized proinflammatory stimuli. Stimulation of NF-kB activity by LPS and IL-1a does not alter the action of 20(OH)D3 on NF-kB activity, since 20(OH)D3 treatment of cells with or without exogenous stimulation had similar effects on p65 localization and IkBa levels. Although NF-kB can be activated through both classical and alternative signaling pathways, previous studies have indicated that IL-1 and LPS activate NF-kB through the classical signaling pathway [46]. In this general pathway, p50:p65 dimers are sequestered in the cytoplasm by IkB proteins. LPS and IL-1 stimulate IkB kinase activity, resulting in the subsequent IkB phosphorylation and ubiquitinylation. Then IkB is targeted for proteosomal degradation, which allows p50:p65 dimers to translocate to the nucleus, bind to DNA and activate the transcription of NF-kB-dependent genes. Consistent with this general pathway we show that both LPS and IL-1 stimulate NF-kB transcriptional activity as well as result in IkB degradation. Most importantly, we show that 20(OH)D3 acts as an immunosuppressive agent in human keratinocytes by blocking the activation of this signaling pathway by both IL-1 and LPS. 20(OH)D3 not only inhibits the translocation of the p65 NF-kB protein from cytoplasm to nucleus in keratinocytes, but also increases the cellular levels of the inhibitory NF-kB protein, IkB, thus sequestering the NF-kB in the cytoplasm as transcriptionally inactive NF-kB/IkB complexes. Since recent studies demonstrate that activation of the alternative NF-kB pathway can also lead to the translocation of p65containing dimers into the nucleus [42], our data cannot exclude the possibility that 20(OH)D3 also blocks this signaling pathway as well. Detailed analysis of the alternative signaling pathway will be the subject of future studies.
In previous studies we showed that the action of 20(OH)D3 on proliferation and differentiation in keratinocytes requires VDR expression [15]. In the present study we find that silencing VDR expression in keratinocytes blocks the inhibitory actions of 20(OH)D3 on NF-kB activity (Fig. 8). Therefore, our data indicates that both 20(OH)D3 (novel ligand) and 1,25(OH) 2 D3 (classical ligand) suppress NF-kB activity through a VDRmediated signaling pathway. Although the mechanism of this pathway in inhibiting NF-kB activity requires more in-depth analysis, our studies demonstrate that 20(OH)D3 can induce antiinflammatory actions similar to those mediated by calcitriol (1,25(OH) 2 D3) via the VDR-mediated inhibition of NF-kB activity. Interestingly vitamin D analogues are now widely used drugs for the treatment of psoriasis, an inflammatory and hyperproliferative dermatoses (reviewed in [49]). Therefore, we believe that 20(OH)D3 holds promise as a novel therapeutic agent in the prevention and therapy of inflammatory, auto-immune and hyperproliferative skin diseases.
Recently, new and important immunomodulatory effects of vitamin D analogs have been characterized, especially those for 1,25(OH) 2 D3 [1,25,67]. Inhibitors targeting the NF-kB signaling pathway effectively suppress NF-kB activity, protect and relieve inflammatory symptoms, and induce apoptosis of tumor cells. NF-kB represents an attractive drug target for therapy of inflammatory and autoimmune disorders, as well as for cancer. Thus, 20(OH)D3 is a new powerful analog of vitamin D3 that is produced by enzymatic activity of CYP11A1 [7,10], and have pleiotropic activities through its ability to modulate the NF-kB signaling pathway as illustrated in figure 9. Increased expression of IkBa and inhibition of NF-kB activity in keratinocytes induced by 20(OH)D3 may be one mechanism by which this (potentially endogenous) vitamin D analog could exert beneficial effects in inflammatory and auto-immune disorders.

Cell culture
Immortalized human keratinocytes (HaCaT), which are frequently used for studies on biological effect of 1,25(OH) 2 D3 [30,68], were cultured in Dulbecco's Modified Eagle Medium supplemented with glucose, L-glutamine, pyridoxine hydrochloride (Cell Grow), 5% fetal bovine serum (Sigma) and 1% penicillin/streptomycin/amphotericin antibiotic solution (Sigma) [6]. In order to eliminate potential interference by sterols present in the serum [69], 5% charcoal/dextran-treated bovine serum (HyClone) was used to test the effects of active forms of vitamin D. In addition, cells were serum-deprived for 24 h before treatment. Normal human epidermal keratinocytes were isolated from neonatal foreskin (HEKn) and grown in KGM medium supplemented with KGF (Lonza) on collagen-coated plates [15]. For experiments cells in their third passage were used.

Immunofluorescent staining
HEKn cells were seeded onto cover glasses in 6-well plate and treated with 100 nM of 20(OH)D3 for 24 h. Control cells were treated with solvent (,0.1% ethanol). After treatment cells were washed in PBS and fixed in 4% paraformaldehyde. Cells were than incubated in permeabilizing solution (0.2% Triton-X 100 in PBS) for 5 min, washed with PBS and blocked in 2% BSA for 30 min. Primary antibody, either goat anti-rabbit-p65 (1:100) or goat anti-rabbit-IkB (1:100) in 1% BSA, was added to the cells and incubated overnight at 4uC. After extensive washing in PBS, cells were incubated in the secondary antibody solution comprising goat-anti-rabbit-Alexa Fluor 488 (Invitrogen, 1:500 in PBS) and incubated for 1 h at room temperature in the dark. Cells on cover glass were washed and mounted with mounting medium containing propidium iodine (Vectashield). Stained cells were analyzed using a fluorescent microscope at 406 magnification.

Transfection and Reporter assay
The constructs pLuc, pHRLTK and NFkB-Luc have been described previously [70]. According to our protocols [63,66], HaCaT and normal epidermal keratinocytes were transfected using Lipofectamine Plus (Invitrogen, Carlsbad, CA) in DMEM or KGM medium, with firefly luciferase reporter gene plasmid and with phRL-TK (expresses Renilla luciferase and serves as normalization control; Promega, Madison, WI). At 24 hr after transfection, the medium was changed and the cells were treated with the vitamin D3 derivatives or vehicle (,0.1% ethanol) for 0.5, 1, 4, 16 and 24 h. Following this protocol, cells were also treated with 10 ng/ml interleukin 1a (IL-1a) (Sigma) or 1 mg/ml LPS (Sigma) for 30 minutes. The Firefly luciferase and Renilla luciferase signals were recorded with a TD-20/20 luminometer (Turner Designs, Sunnyvale, CA); background luminescence was subtracted and the resulting promoter specific firefly signal was divided by the Renilla signal (proportional to the number of transfected cells). The values obtained were calculated relative to control (untreated) cells, and expressed as relative fold change.

siRNA transfection
Keratinocytes were transfected with 2 nM VDR or scrambled siRNA (Dharmacon), on-Target plus smart pool human VDR or on-Target plus siControl non-targeting pool, using lipofectamine plus (Invitrogen) in DMEM medium. Twenty four hours after transfection, cells were treated for an additional 4 h with 100 nM 20(OH)D3 or vehicle (ethanol), mRNA was isolated and used for gene expression analysis or cells were stained for IkBa or NFkB p65 and examined for protein localization by fluorescent microscopy.

Preparation of cell lysates
Cells were treated with 20(OH)D3 or 1,25(OH) 2 D3, and whole cell lysates were prepared as described previously [71,72]. Cells were resuspended in RIPA buffer containing protease inhibitor cocktail (Sigma) and PMSF. Nuclear extracts were prepared as described previously [73]. In brief, HaCaT or normal human keratinocytes were treated with 100 nM 20(OH)D3 for 0, 0.5, 1, 4, 16 or 24 h, and then stimulated with or without interleukin 1a (10 ng/ml) for an additional 30 min. The cells were harvested, pelleted and resuspended in STM buffer (20 mM Tris-HCl, 250 mM sucrose and 1.1 mM MgCl 2 ). The nuclear pellet was resuspended in 30 ml nuclear extraction buffer containing 0.4 M KCl, 5 mM 2-mercaptoethanol and protease inhibitor cocktail (1:100 dilution, Sigma) in STM buffer, and incubated on ice for 30 min with intermittent shaking and then centrifuged at 14,0006g for 20 min at 4uC. The protein content in the supernatant was quantified using the Bradford protein assay kit [74]. Cytosolic extracts were prepared as described previously with minor modifications [55]. Cell pellets were resuspended in hypotonic buffer (10 mM HEPES pH 7.9, 10 mM KCl, 0.2 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5% NP40, protease inhibitor cocktail (Sigma) and 20 mM PMSF), and after a 20 min of incubation on ice the suspension was centrifuged at 4uC for 10 minutes at 5,0006g. The centrifuged supernatant was considered the cytosolic extract.
The levels of IkBa, the p65 NF-kB protein, and b-actin were assessed by immunoblotting. Primary antibodies used were the rabbit polyclonal antibodies directed against IkB-a (Santa Cruz, 1:500 dilution); p65 (Santa Cruz, 1:500 dilution) and b-actinperoxidase (Sigma, 1:5000 dilution). The secondary antibody used was anti rabbit IgG (Santa Cruz, 1:7,000 dilution) and anti-mouse IgG (Santa Cruz, 1:5,000) conjugated to horseradish peroxidase. Signals were detected using ECL kit Supersignal West Pico Chemiluminescent Substrate (Pierce). The intensity of bands was measured using ImageJ Software. Results for whole cell and cytosolic protein levels were expressed relative to b-actin levels [74]. Levels of VDR and b-actin 24 h after VDR siRNA transfection were assessed in western blots with (VDR(D-6)) antibody (Santa Cruz, 1:400).

Electrophoretic mobility shift assay (EMSA)
DNA binding activity was determined by EMSA using a consensus NF-kB IRDye-labeled oligonucleotide probe (LI-COR). The DNA binding reaction consisted of 2.5 to 5 mg of the nuclear extract, the NF-kB probe and gel shift binding buffer containing of 2.5 mM DTT, 0.25% Tween-20 and 0.25 mg/ml poly(dI) :poly(dC). The reaction was carried out at room temperature in the dark for 30 min. For supershift assays 1 mg of p65 or p50 antibody (gift of the NCI Preclinical Repository) was added to the nuclear extract prior to DNA binding and incubated for 30 min at 4uC. Orange loading dye was added to samples which were loaded on pre-run 5% TBE gels and run at 70 V for 2 h. The gel was scanned using an Odyssey Infrared Imaging System (LI-COR, Inc,. Lincoln, NE).

Real-time RT PCR
The RNA from HaCaT and normal keratinocytes treated with 20(OH)D3 or 1,25(OH) 2 D3 was isolated using Absolutely RNA Miniprep Kit (Stratagen). Reverse transcription (100 ng/reaction) was performed with Transcriptor First Strand cDNA Synthesis Kit (Roche). Real-time PCR was performed using cDNA diluted 10fold in sterile water and a TaqMan PCR Master Mix (n = 3). Reactions were performed at 50 uC for 2 min, 95 uC for 10 min and than 50 cycles of 95 uC for 15 s, 60 uC for 1 min). The primers and probes were designed with universal probe library (Roche): IkBa primers (left: GTCAAGGAGCTGCAGGAGAT and right: GATGGCCAAGTGCAGGAA), probe #38; NFkB1 primers (left: ACCCTGACCTTGCCTATTTG and right: AGCTCTTTTTCCCGATCTCC), probe #39; and RelA primers (left: CGGGATGGCTTCTATGAGG and right: CTCCAGGTCCCGCTTCTT), probe #47; VDR primers (left: CTTACCTGCCCCCTGCTC and right AGGGTCAGG-CAGGGAAGT), probe #58. The data was collected on a Roche Light Cycler 480. The amounts of product were compared to Cyclophilin B using a comparative C T method.

Statistical analysis
Data are presented as means6STDEV and were analyzed with Student's t-test (for 2 groups) and one-way Anova with appropriate post-hoc test (for more than 2 groups) using Excel (Microsoft) and Prism 4.00 (GraphPad Software, San Diego), respectively. Statistically significant differences are denoted with asterisks: *P,0.05, **P,0.001.