Introduction

Nuclear factor-κB (NF-κB) is a dimeric transcription factor formed by hetero- or homodimerization of the five Rel family proteins which comprise RelA (p65), RelB, cRel, p52 and p50 (Foo and Nolan 1999). NF-κB is believed to play a pivotal role in immune and inflammatory responses and in the regulation of cell proliferation and apoptosis. It regulates the transcription of genes encoding proinflammatory cytokines (e.g. IL-1, IL-2, TNFα and GM-CSF), chemokines (e.g. IL-8 and RANTES), adhesion molecules (e.g. ICAM, VCAM and E-selectin), inducible enzymes (e.g. COX2 and iNOS) and also the MHC proteins important for the adaptive immune response (Baldwin 2001a). Furthermore, the gene coding for the tumor suppressor protein p53 as well as other regulators of apoptosis and cell proliferation (e.g. c-IAP-1, c-IAP-1 Bcl-XL and c-myc) are regulated by NF-κB (Baldwin 2001b).

In resting cells NF-κB is usually retained in an inactive state in the cytoplasm through binding to a member of the NF-κB inhibitor protein family, inhibitory κB (IκB), of which IκBα and IκBβ are the major inhibitors (Thompson et al. 1995). A wide variety of agonists including the proinflammatory cytokine IL-1 have been shown to activate NF-κB. NF-κB activation is mediated through activation of a specific IκB kinase (IKK) and subsequent phosphorylation of IκB molecules. Once phosphorylated, IκB is targeted for ubiquitination and degradation by proteasomes allowing translocation of the NF-κB dimer to the nucleus where it binds to specific response elements in the promoter region of the various target genes (Rothwarf and Karin 1999). A 10-bp κB response element serves as the recognition site for NF-κB and the κB response element differs from gene to gene (Yamamoto et al. 1992).

NF-κB has recently been shown to play different roles in different phases of inflammation (Lawrence et al. 2001). During the onset of inflammation NF-κB activation is associated with proinflammatory gene expression whereas NF-κB activation during the resolution of inflammation is associated with the expression of antiinflammatory genes and the induction of apoptosis.

IL-8 is a proinflammatory chemokine which activates and attracts neutrophils and T-lymphocytes from the blood into sites of infection and inflammation. IL-8 is produced by keratinocytes, fibroblasts, peripheral blood monocytes and endothelial cells by an inflammatory stimulus and is believed to play an important role in the development of inflammation (Larsen et al. 1991). Furthermore, IL-8 expression has been demonstrated to be markedly upregulated in psoriatic skin (Jiang et al. 2001). Regulation of IL-8 gene transcription occurs mainly through sequences −94 to −71 of the 5′ flanking region of the IL-8 gene. This part of the IL-8 promoter region contains binding sites for the transcription factor NF-κB (Harant et al. 1997) and the κB binding site has been demonstrated to be essential for IL-1-induced IL-8 gene expression.

p53 is a regulator of cellular apoptosis and cell proliferation but does not exhibit any proinflammatory properties (Lotem and Sachs 1998). p53 also contains a κB binding site in the promoter region and NF-κB has been shown to regulate p53 gene expression (Lawrence et al. 2001). Interestingly, the κB binding site in the p53 promoter region is different from the κB binding site in the IL-8 promoter region (Pei et al. 1999; Kim et al. 2000).

An imbalance in NF-κB signaling has previously been suggested to play an important, albeit controversial role, in the pathogenesis of psoriasis (Danning et al. 2000; McKenzie and Sabin 2003). Recently, we have shown that in involved psoriatic skin the expression of p53 is decreased compared to that in uninvolved psoriatic skin (own unpublished results) whereas it is well known that IL-8 expression is increased in involved psoriatic skin compared to that in uninvolved skin.

Vitamin D and different synthetic vitamin D analogues have been widely used in the treatment of psoriasis (Kragballe et al. 1991). The cellular actions of 1α,25-dihydroxyvitamin D3 (1α,25(OH)2D3), 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) (Haussler 1986). However, recently 1α,25(OH)2D3 has also been demonstrated to activate other signal transduction pathways resulting in regulation of transcription factors such as activator protein 1 (AP-1) (Johansen et al. 2000, 2003). Furthermore, 1α,25(OH)2D3 has been shown to inhibit IL-1α-induced IL-8 synthesis in human keratinocytes in vitro (Larsen et al. 1991) and in a melanoma cell line TNFα-induced IL-8 expression has been shown to be inhibited by 1α,25(OH)2D3 through inhibition of NF-κB DNA binding (Harant et al. 1997). Thus, the mechanism behind this modulation of NF-κB activity has not been shown.

The purpose of the present study was to determine the effect of 1α,25(OH)2D3 on NF-κB activation in cultured normal human keratinocytes and to determine whether or not 1α,25(OH)2D3 leads to a selective and differentiated regulation of NF-κB/DNA binding activity to proinflammatory (IL-8) and proapoptotic (p53) genes.

Materials and methods

Materials

Keratinocyte growth medium, human recombinant epidermal growth factor, bovine pituitary extract, gentamicin and 8–16% Tris-glycine gels were obtained from Invitrogen (Carlsberg, Calif.). 1α,25(OH)2D3 dissolved in 2-propanol was kindly provided by Leo Pharmaceuticals Products (Ballerup, Denmark). Gel shift binding 5× buffer, T4 polynucleotide kinase, T4 polynucleotide kinase 10× buffer and goat anti-mouse HRP-conjugated antibodies were obtained from Promega (Madison, Wis.). IL-8 NF-κB response element (5′-AAT CGT GGA ATT TCC TCT GAC A-3′) and p53 NF-κB response element (5′-GGG ATT GGG GTT TTC CCC TCC C-3′) were purchased from DNA Technology (Aarhus, Denmark). [γ-32P]ATP, enhanced chemiluminescence detection system (ECL), microspin G-25 columns and nick spin columns Sephadex G-50 fine DNA-grade were obtained from Amersham/Pharmacia Biotech (Uppsala, Sweden). Complete medium was purchased from Boehringer Mannheim, Germany. Trypsin, EDTA, Tris-base, Tris-HCl, DTT, O-phenylenediamine dihydrochloride tablets and boric acid were obtained from Sigma Chemical Company (St. Louis, Mo.). The IκBα, p53, p50 and p65 antibodies for Western blotting and supershifting experiments were from Santa Cruz Biotechnology (Santa Cruz, Calif.). The secondary antibody for Western blotting was purchased from DAKO, Denmark. IL-1α was obtained from RD System (Abingdon, UK). The monoclonal recombinant human IL-8 (rIL-8) antibody for the IL-8 ELISA was kindly provided by Kouji Matsushima (Dainippon Pharmaceuticals, Osaka, Japan). Bio-Rad protein assay and acrylamide were purchased from Bio-Rad (Copenhagen, Denmark).

Cell cultures

Normal adult human keratinocytes were obtained by trypsinization of skin samples from patients undergoing plastic surgery as previously described (Kragballe et al. 1985). First-passage keratinocytes were grown in serum-free low-calcium (0.09 mM) keratinocyte growth medium supplemented with 5 ng/ml human recombinant epidermal growth factor, 50 μg/ml bovine pituitary extract, and 5 μg/ml gentamicin. Cells were grown at 37°C in a humidified atmosphere containing 5% CO2. At 70–90% confluence (judged by light microscopy) keratinocyte growth medium supplemented only with 5 μg/ml gentamicin was added for 24 h. 1α,25(OH)2D3 at various concentrations (10−7–10-11 M final concentration) was added and cells were incubated from 10 min to 48 h. In separate experiments cells were further stimulated with IL-1α for 15 min. The 1α,25(OH)2D3 stock solution was diluted in 2-propanol, IL-1α was diluted in PBS, pH 7.0, and 0.15% of bovine serum albumin added. In all experiments, vehicle was added to controls. All incubations were carried out in duplicate.

Whole cell extracts

Keratinocytes were washed twice with ice-cold PBS, lysed with cold lysis buffer (50 mM Tris-HCl (pH 6.8), 10 mM DTT, 10 mM β-glycerophosphate, 10 mM NaF, 0.1 mM Na orthovanadate, 10% glycerol, 2.5% SDS, 25 mM PMSF and 50× complete), policed into tubes and boiled for 3 min. The lysates were centrifuged at 13,600 g for 3 min and the supernatants were assayed for protein concentration as previously described by Bradford (1976).

Nuclear and cytosolic extraction

Nuclear extracts were prepared as previously described (Johansen et al. 2000; Rosette and Karin 1995). Briefly a hypotonic buffer (10 mM HEPES, 10 mM KCl, 0.2 mM EDTA, 0.1 mM EGTA, 1 mM DTT, pH 7.9) containing 0.5% detergent was added to the cultured cells. Cells were policed into tubes. The suspension was passed through a 27-gauge needle six times and centrifuged at 13,600 g at 4°C for 1.5 min. The supernatant was considered the cytosolic extract. The pelleted nuclei were resuspended in a hypertonic buffer (20 mM HEPES, 0.4 mM NaCl, 1 mM EDTA and 1 mM DTT), rotated at 4°C for 30 min and centrifuged at 13,600 g at 4°C for 5 min. The supernatant was considered the nuclear extract. Protein concentration was determined as previously described by Bradford (1976) with bovine serum albumin as standard.

Labeling of consensus oligonucleotides for NK-κB

NF-κB consensus oligonucleotide (4 μl, 1.75 pmol/μl), 2 μl T4 polynucleotide kinase 10× buffer (700 mM Tris-HCl, 100 mM MgCl2 and 50 mM DTT), 4 μl [γ-32P]ATP (3000 Ci/mmol, 10 mCi/ml), 8 μl nuclease-free water and 2 μl T4 polynucleotide kinase were added to an Eppendorf tube and incubated at 37°C for 10 min. To stop the reaction, 1 μl 0.5 M EDTA was added. Finally, 200 μl TE buffer (10 mM Tris-HCl, pH 8, and 1 mM EDTA) was added. The suspension was centrifuged at 300 g for 5 min on a Nick spin column in order to remove the unlabelled oligonucleotides.

Electrophoretic mobility shift assay

Nuclear protein (3–8 μg) was preincubated for 10 min at room temperature in binding buffer (5 mM MgCl2, 2.5 mM EDTA, 2.5 mM DTT, 250 mM NaCl, 50 mM Tris-HCl (pH 7.5), 20% glycerol, 0.25 mg/ml poly(dI-dC)) and hypotonic buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.2 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 1× protease inhibitors, and 0.5 mM PMSF). 32P-labelled NF-κB probe (1 μl) was added and the suspension was incubated for an additional 20 min before loading on a 4% polyacrylamide gel in 0.5× TBE buffer (22.5 mM Tris-base, 4 mM boric acid, and 1 mM EDTA). In control experiments a specific competitor (unlabelled NF-κB oligo) or a nonspecific competitor (unlabelled SP-1 oligo) was added 10 min before addition of labeled NF-κB oligo. Supershifting experiments with p50 and p65 antibodies were also carried out in order to confirm the identity of NF-κB in the various experiments with an electrophoretic mobility shift assay (Fig. 1). The gels were dried and exposed to X-ray film at −80°C.

Fig. 1
figure 1

Control experiments using a specific competitor (unlabelled NF-κB oligo) and a nonspecific competitor (unlabelled SP-1 oligo)

Full size image

Western blotting

Cytosolic protein (20 μg), freeze-dried at −50°C was added to 20 μl loading buffer (100 μl Tris-glycine SDS sample buffer, 20 μl 0.5 M DTT, 80 μl dH2O). The suspension was boiled, cooled and centrifuged then separated by SDS-PAGE (8–16% Tris-glycine) and blotted onto nitrocellulose membranes. The gels were stained and dried and the membranes incubated overnight at 4°C in skimmed milk powder diluted in TBS (20 mM Tris-base, pH 7.6, 135 mM NaCl). Primary antibody (IκBα, 1:1000) in TBS was added and incubated for 1.5 h at room temperature. Secondary antibody (rabbit immunoglobulins/HRP, swine; 1:1000) in TBS was added and the mixture incubated for 1 h at room temperature. The membrane was washed in TBST (TBS + 0.05% Tween-20), and the reaction revealed by the ECL method.

IL-8 ELISA

A 96-well plate was coated with monoclonal IL-8 antibody in coating buffer, 1.5 μg/ml, and kept overnight at 4°C. After washing in PBS the plate was blocked for 1 h at 20°C before the samples and the standards were added and left overnight at 4°C. Then the secondary antibody was added for 1 h at 20°C, followed by the addition of peroxidase-labeled antibody diluted in assay buffer for 1 h at 20°C. After washing, the reaction was developed with O-phenylenediamine, and 30 min later the reaction was stopped by adding 1.6 N H2SO4. Optical density (OD) was measured in an ELISA reader adjusted to 490 nm.

Statistics

Results are expressed as means±SD. Statistical significance (P<0.05) was assessed by Students t-test. To test for normal distribution, a probability test was carried out.

Results

Effects of 1α,25(OH)2D3 on NF-κB DNA binding activity in cultured normal human keratinocytes

Studies were performed to determine the effect of 1α,25(OH)2D3 on NF-κB DNA binding activity to the p53 and the IL-8 κB response element in cultured normal human keratinocytes. Because the keratinocytes were not stimulated, only the basal NF-κB DNA binding activity was measured. When human keratinocytes were incubated with 1α,25(OH)2D3 (10−7 M) a time-dependent decrease in NF-κB DNA binding activity to the p53 κB response element was seen (Fig. 2a). It became significant after 6 h and reached a maximal inhibition of 69% compared to controls after 24 h (Fig. 2a). No detectable NF-κB binding activity to the IL-8 κB response element was found in cultured normal human keratinocytes incubated with either vehicle or 1α,25(OH)2D3 (10−7 M) from 10 min to 48 h (data not shown).

Fig. 2a, b
figure 2

Human keratinocytes were incubated with 1α,25(OH)2D3 (10−7 M) for the indicated times. Control represents keratinocytes cultured with vehicle only. NF-κB DNA binding activity was determined by EMSA using a synthetic p53 NF-κB target sequence. a The 32P-bands were analyzed densitometrically. NF-κB DNA binding activity is expressed as percentage of control. Each bar represents the mean±SD from four separate experiments assayed in duplicate. b IκBα expression was determined by Western blotting using a specific antibody. The intensity of the bands was analyzed densitometrically. IκBα expression is expressed as percentage of control. Each bar represents the mean±SD from at least four separate experiments assayed in duplicate. Equal loading was confirmed by incubation with an anti-TFIIB antibody (*P<0.05). Due to the number of lanes the results shown are the combined results from two different gels. A control lane was included on each gel

Full size image

The expression of IκBα was determined by Western blotting and a significant time-dependent increase in the expression of IκBα was seen when human keratinocytes were incubated with 1α,25(OH)2D3 (10−7 M) (Fig. 2b). This increase became significant after 6 h and reached a maximum after 24 h of incubation. At 6, 12 and 24 h of incubation the levels of IκBα were 178%, 189% and 211%, respectively, compared to control levels (Fig. 2b).

Human keratinocytes were also incubated with different concentrations of 1α,25(OH)2D3 (10-11–10−7 M final concentrations) for 12 h and a dose-dependent decrease in NF-κB binding activity to the p53 κB response element was seen (Fig. 3a). These findings were paralleled by a similar dose-dependent increase in cytoplasmic IκBα expression (Fig. 3b) which became significant at a concentration of 10−8 M 1α,25(OH)2D.

Fig. 3a, b
figure 3

Human keratinocytes were incubated with 1α,25(OH)2D3 at different concentrations for 12 h. Control represents keratinocytes incubated with vehicle only. NF-κB DNA binding activity was determined by EMSA using a synthetic p53 NF-κB target sequence. a The 32P-bands were analyzed densitometrically. NF-κB DNA binding activity is expressed as percentage of control. Each bar represents the mean±SD from five separate experiments assayed in duplicate. b IκBα expression was determined by Western blotting using a specific antibody. The intensity of the bands was analyzed densitometrically. IκBα expression is expressed as percentage of control. Each bar represents the mean±SD from at least five separate experiments assayed in duplicate. Equal loading was confirmed by incubation with an anti-TFIIB antibody (*P<0.05)

Full size image

1α,25(OH)2D3 modulates IL-1α-induced NF-κB DNA binding activity in cultured normal human keratinocytes

When cultured normal human keratinocytes were stimulated with IL-1α (10 ng/ml) for 15 min a significant increase in NF-κB DNA binding activity to both the p53 and the IL-8 κB response element was seen compared to stimulation with vehicle only (Fig. 4a).

Fig. 4a–c
figure 4

Human keratinocytes stimulated with or without IL-1 (10 ng/ml) for 15 min in order to demonstrate increased NF-κB DNA binding to both the p53 and the IL-8 NF-κB target sequence (a). Human keratinocytes were also incubated with 1α,25(OH)2D3 (10−7 or 10−8 M) for 12 or 24 h before stimulation with IL-1α (10 ng/ml) for 15 min. NF-κB DNA binding activity was assayed by EMSA using a synthetic p53 NF-κB target sequence (b) or a synthetic IL-8 NF-κB target sequence (c). The 32P-bands were analyzed densitometrically. NF-κB DNA binding activity is expressed as percentage of control. Control represents keratinocytes incubated with IL-1α only. Each bar represents the mean±SD from three separate experiments (*P<0.05)

Full size image

In order to determine whether 1α,25(OH)2D3 could modulate the IL-1α-induced increase in NF-κB DNA binding activity normal human keratinocytes were preincubated with 1α,25(OH)2D3 (10−8 and 10−7 M) for 12 or 24 h before stimulation with IL-1α (10 ng/ml) for 15 min. NF-κB DNA binding activity to the p53 κB response element showed a slight but significant (P<0.05) decrease in the NF-κB DNA binding activity when the keratinocytes were preincubated with 1α,25(OH)2D3 (10−7 M) for 12 and 24 h before stimulation with IL-1α (Fig. 4b). Preincubation for 12 and 24 h resulted in a 50% and a 25% decrease, respectively, compared to controls. Preincubation with 1α,25(OH)2D3 at 10−8 M resulted in only insignificant changes. When NF-κB DNA binding activity to the IL-8 promoter region was determined (Fig. 4c) a significant decrease was seen after 24 h of preincubation with 1α,25(OH)2D3 at both 10−7 and 10−8 M. The mean levels of inhibition at 10−7 and 10−8 M were 72% and 65%, respectively. Only insignificant changes were seen when keratinocytes were preincubated with 1α,25(OH)2D3 for 12 h.

In order to determine whether the 1α,25(OH)2D3-mediated downregulation in IL-1α-induced NF-κB DNA binding activity was accompanied by changes in IκBα levels, the expression of IκBα was determined in the above-mentioned IL-1α-stimulated cell samples. Preincubation for 24 h with 1α,25(OH)2D3 at 10−7 M resulted in a 50% increase in the IκBα level after IL-1α stimulation as compared to controls (P<0.05; Fig. 5). Preincubation for 12 h with 10−7 M 1α,25(OH)2D3 and preincubation with 10−8 M 1α,25(OH)2D3 for both 12 and 24 h before stimulation with IL-1α led only to insignificant increases in the IκBα levels when compared to controls. The P-values were 0.057, 0.076 and 0.116, respectively.

Fig. 5
figure 5

Keratinocytes were incubated with 1α,25(OH)2D3 (10−7 or 10−8 M) for 12 or 24 h followed by stimulation with IL-1α (10 ng/ml) for 15 min. Then the cytosolic level of IκBα was determined by Western blotting. The bands were analyzed densitometrically. The cytosolic level of IκBα is expressed as percentage of controls. Control represents keratinocytes incubated with IL-1α only. Each bar represents the mean±SD from five separate experiments assayed in duplicate. Equal loading was confirmed by incubation with an anti-TFIIB antibody (*P<0.05)

Full size image

1α,25(OH)2D3 regulates IL-1α-induced IL-8 expression in cultured normal human keratinocytes

p53 and IL-8 expression were also measured in cultured normal human keratinocytes in order to determine whether the 1α,25(OH)2D3-modulated changes in IL-1α-induced NF-κB DNA binding activity were reflected by similar changes in protein expression. p53 expression was determined by Western blotting and IL-8 expression was determined by ELISA.

Cultured keratinocytes were preincubated with 1α,25(OH)2D3 (10−7 M) for various times (6–48 h) before stimulation with IL-1α, and then 12 h later protein expression was measured. Interestingly, no significant changes in p53 expression were seen after preincubation with 1α,25(OH)2D3 (10−7 M) from 6 to 48 h before stimulation with IL-1α (Fig. 6). In contrast a significant decrease in IL-8 expression was seen after preincubation with 1α,25(OH)2D3 (10−7 M) from 6 to 48 h before stimulation (Fig. 7). After 24 and 48 h of preincubation a 47% and 52% inhibition of IL-8 expression were seen, respectively.

Fig. 6
figure 6

Whole cell extracts from human keratinocytes preincubated with 1α,25(OH)2D3 (10−7 M) or vehicle for the indicated times and stimulated with IL-1α (10 ng/ml) for 12 h were isolated and the p53 expression levels determined by Western blotting. The bands were analyzed densitometrically. Control represents keratinocytes stimulated with IL-1α only. The expression levels of p53 are expressed as a percentage of control. Each bar represents the mean±SD from five separate experiments assayed in duplicate. Equal loading was confirmed by incubation with an anti-TFIIB antibody

Full size image
Fig. 7
figure 7

Normal human keratinocytes were incubated with 1α,25(OH)2D3 (10−7 M) or vehicle for the indicated times and stimulated with IL-1α (10 ng/ml) for 12 h. ELISA was used to determine the expression levels of IL-8. IL-8 expression is expressed as percentage of control. Control represents keratinocytes stimulated with IL-1α only. Each bar represents the mean±SD from five separate experiments assayed in duplicate (*P<0.05)

Full size image

Discussion

In this study we found a difference in NF-κB binding activity to the κB sequence from the IL-8 and the p53 promoter in normal human keratinocytes in vitro. In unstimulated keratinocytes we could not detect any NF-κB binding to the IL-8 κB binding sequence whereas stimulation with IL-1α led to a significant induction of NF-κB binding. NF-κB binding to the p53 κB binding sequence was detectable even in unstimulated cells, but it was significantly increased after IL-1α stimulation. Furthermore, 1α,25(OH)2D3 stimulated the expression of IκBα. In parallel with this increase in IκBα expression, we found a time- and dose-dependent decrease in NF-κB binding to the IL-8 κB binding sequence and in IL-8 expression after preincubation of normal human keratinocytes with 1α,25(OH)2D3 and stimulation with IL-1α. A less-pronounced decrease in NF-κB binding to the p53 κB response element was seen after preincubation with 1α,25(OH)2D3 and stimulation with IL-1α, and it did not result in any changes in p53 expression.

The inhibitory effect of 1α,25(OH)2D3 on NF-κB activation has previously been demonstrated in different cell types such as keratinocytes (Komine et al. 1999), lymphocytes (Komine et al. 1999) and fibroblasts (Harant et al. 1998), but the mechanism involved has not been fully elucidated. Harant et al. (1998) have shown that 1α,25(OH)2D3 inhibits NF-κB DNA binding without affecting the translocation of NF-κB to the nucleus. Komine et al. (1999) only observed a 1α,25(OH)2D3-mediated decrease in NF-κB controlled luciferase activity in keratinocytes but did not study the mechanism. To our knowledge, our results are the first to demonstrate a 1α,25(OH)2D3-induced specific and selective regulation of NF-κB binding to the p53 and IL-8 κB response element as well as a specific regulation of the expression of IL-8 and p53.

The observed 1α,25(OH)2D3-induced selective regulation of NF-κB DNA binding activity can be explained by a specific modulation of some of the NF-κB subunits. p65 has been shown by other groups to interact with a subset of κB DNA sequences not recognized by p50 and vice versa (Kunsch et al. 1992) and interestingly the IL-8 κB response element used in their study was shown to bind p65 (Kunsch et al. 1992). This is in accordance with the findings of another study showing that p65 but not p50 is important for the NF-κB-induced IL-8 gene transcription (Kunsch and Rosen 1993) whereas the NF-κB-dependent regulation of p53 expression has been shown to be mediated through the p50/p50 homodimer (Wu and Lozano 1994; Lawrence et al.2001). IκBα is also selective in its mode of action. It predominantly binds to p65 with only weak binding to p50 (Phelps et al. 2000). It is therefore possible that the 1α,25(OH)2D3-induced increase in IκBα expression leads to a selective inhibition of p65 resulting in the observed decrease in NF-κB binding to the IL-8 κB response element and the subsequent decrease in IL-8 expression. The decrease in NF-κB binding to the p53 κB response element may be due a decrease in other NF-κB dimers than the p50/p50 homodimer, e.g. the p50/p65 heterodimer which is inhibited by the increased expression of IκBα. However, because p53 expression is regulated by the p50/p50 homodimer this decrease in NF-κB DNA binding did not result in any changes in p53 expression.

One intriguing hypothesis may therefore be that 1α,25(OH)2D3 mediates its immunomodulatory and antiinflammatory effects by selectively inhibiting NF-κB-mediated transcription of proinflammatory genes without affecting NF-κB-induced transcription of genes such as p53.

An inhibition of NF-κB activation has been suggested as part of the mechanism of action of several antiinflammatory drugs. 1α,25(OH)2D3 exerts some of its effects through binding to the VDR. The VDR belongs to the nuclear receptor superfamily which includes more than 60 different nuclear receptors including the glucocorticoid receptor (GR) (Carlberg and Polly 1999). Interestingly, inhibition of NF-κB activation has been suggested as part of the immunomodulatory effect of the glucocorticoids (Scheinman et al. 1995; Auphan et al. 1995; De Bosscher et al. 1997), and it has been shown that glucocorticoids increases IκBα transcription leading to enhanced expression of cytosolic IκBα (Scheinman et al. 1995; Auphan et al. 1995). This is similar to what was demonstrated with 1α,25(OH)2D3 in this study.

IL-1α was chosen to stimulate the keratinocytes because it is an important proinflammatory cytokine in the skin and it is expressed in keratinocytes (Wood et al. 1996). Furthermore, IL-1α has been suggested to play a key role in hyperproliferative and inflammatory skin diseases (Haase et al. 2001) and it is known to contribute to the activation of NF-κB (Wolf et al. 2001).

In the present study, we demonstrated that 1α,25(OH)2D3 inhibits NF-κB binding to the IL-8 κB binding sequence more potently than binding to the p53 κB binding sequence which leads to a significant decrease in IL-8 expression, whereas no significant changes in p53 expression were seen. We therefore propose that this selectivity at least in part may be mediated through an increased expression of IκBα. Increased expression of IκBα and selective inhibition of the expression of NF-κB-regulated proinflammatory cytokines and chemokines may be one of the mechanisms by which vitamin D analogues exert their immunomodulatory and beneficial effects in the treatment of psoriasis.