Direct transcriptional regulation of RelB by 1alpha,25-dihydroxyvitamin D3 and its analogs: physiologic and therapeutic implications for dendritic cell function.

The nuclear factor-kappaB (NF-kappaB) protein RelB plays a unique role in dendritic cell (DC) function and, as such, is an important regulator of antigen presentation and immune regulation. In this study, inhibition of RelB expression in DCs exposed to an analog of the active form of vitamin D3 (1alpha,25-dihydroxyvitamin D3 (1alpha,25-(OH)2D3)) was observed and shown to be mediated by the vitamin D receptor (VDR). Potential vitamin D response elements were identified within promoter regions of human and mouse relB genes. In gel shift experiments, these motifs specifically bound VDR.retinoid X receptor-alpha complexes. Reporter assays confirmed that transcriptional activity of human and mouse relB promoters was inhibited by 1alpha,25-(OH)2D3 agonists in a DC-derived cell line. The inhibition was abolished by mutagenesis of the putative vitamin D response elements and was enhanced by overexpression of VDR. Mutagenesis of NF-kappaB response elements within the relB promoter did not affect the magnitude of 1alpha,25-(OH)2D3 analog-mediated inhibition, ruling out an indirect effect on NF-kappaB signaling. Glucocorticoid caused additional inhibition of relB promoter activity when combined with the 1alpha,25-(OH)2D3 analog. This effect was dependent on the integrity of the NF-kappaB response elements, suggesting separate regulatory mechanisms for the two steroid pathways on this promoter. We conclude that relB is a direct target for 1alpha,25-(OH)2D3-mediated negative transcriptional regulation via binding of VDR.retinoid X receptor-alpha to discrete DNA motifs. This mechanism has important implications for the inhibitory effect of 1alpha,25-(OH)2D3 on DC maturation and for the potential immunotherapeutic use of 1alpha,25-(OH)2D3 analogs alone or combined with other agents.

Dendritic cells (DCs) 1 occupy a unique role in initiating immune responses as a result of their ability to mingle with and potently activate naïve T-cells (1,2). A burgeoning literature also demonstrates an important function for DCs in maintaining peripheral immune tolerance (3). The degree to which DC function can be polarized to induce immune sensitization or tolerance is highlighted by advances toward the therapeutic use of DCs to both boost (for neoplasia and vaccination) and inhibit (for transplantation and autoimmunity) antigen-specific cellular immunity (1,2,4,5). This functional plasticity is linked with a collection of phenotypic changes (termed maturation) that convert the DC from a cell with modest antigenpresenting capacity to one with high surface levels of peptide⅐major histocompatibility complex complexes and costimulatory ligands (1,2). Triggering of the DC maturation program is induced by engagement of surface receptors for microbial products, pro-inflammatory cytokines, and coreceptors expressed by activated T-cells (1,2). Maturational stimuli are channeled through intracellular signaling cascades, the targeting of which has been identified as a key strategy in modulating DC phenotype for the purpose of immunotherapy (6).
Prominent among the signals that regulate DC maturation is the nuclear factor-B (NF-B) pathway (1,2,6,7). Rel/NF-B proteins are a family of transcription factors that serve as pivotal regulators of immune, inflammatory, and acute-phase responses (8 -10). There are five known mammalian Rel/NF-B proteins, Rel (c-Rel), p65 (RelA), RelB, p50 (NF-B1), and p52 (NF-B2), that function as dimers held latently in the cytoplasm by inhibitor proteins (IB). Cellular activation leads to IB phosphorylation and translocation of NF-B dimers to the nucleus, where they act directly upon regulatory elements within the promoter regions of many genes (8 -10). Individual NF-B proteins vary in their cellular distribution, binding partners, mechanisms and kinetics of activation, and target genes (9,10). Several lines of evidence implicate RelB as a critical regulator of the differentiation and maturation of DCs. RelB-deficient mice lack mature myeloid DCs (11,12), and DCs in which RelB expression is inhibited retain an immature phenotype and are associated with induction of immune tolerance in vivo (13). Inhibition of RelB nuclear translocation in DCs has also been observed following the use of tolerogenic immunosuppressive regimens in experimental models of allotransplantation (14).
We have recently reported that the active form of the steroid * This work was supported in part by National Institutes of Health Grant DK59505 (to M. D. G.) and Grants DK25409 and DK58546 (to R. K.) and by the Mayo Foundation CR75 Program (to M. D. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  hormone 1␣,25-dihydroxyvitamin D 3 (1␣,25-(OH) 2 D 3 ) and its analogs, which are known to potently inhibit DC maturation (15), selectively inhibit mRNA and protein expression of RelB in bone marrow-derived DCs (15,16). The inhibition of RelB in DCs is further attenuated by addition of glucocorticoid, and DCs generated in the combined presence of 1␣,25-(OH) 2 D 3 and glucocorticoid agonists exhibit a highly immature phenotype (16). The functional effects of 1␣,25-(OH) 2 D 3 and its analogs are predominantly mediated by the vitamin D receptor (VDR), which then acts as a transcriptional regulator by binding to vitamin D response elements (VDREs) within the promoters of responsive genes, most commonly as a heterodimer with retinoid X receptor-␣ (RXR␣) (17,18). Negative regulation by 1␣,25-(OH) 2 D 3 of immune-related gene products such as interleukin (IL)-2, interferon-␥, and IL-12 p40 has been documented, but it has not been possible to clearly identify VDREs in the promoters of these genes (19 -21). In this report, we present evidence that 1␣,25(OH) 2 D 3 -mediated inhibition of RelB in DCs is a VDR-dependent process that operates through bona fide VDREs within the promoter regions of both human and mouse relB genes and that may be augmented by concurrent interference with separate NF-B response elements (NF-B-REs) in the relB promoter.
Cell Culture and Transient Transfection-Murine bone marrow-derived DCs (BMDCs) were prepared as described previously (23). D 3 analog and dexamethasone (Sigma) were added on days 2, 4, and 6 of culture to final concentrations of 10 Ϫ10 and 10 Ϫ7 M, respectively. Mouse D2SC1 cells (provided by Dr. Sang-Mo Kang, University of California, San Francisco, CA) (24) were cultured in Iscove's modified Dulbecco's medium containing L-glutamine, penicillin/streptomycin, and 5% fetal bovine serum. Cells were transiently transfected with luciferase reporter plasmids, the pRL-TK reference Renilla luciferase plasmid (Promega, Madison, WI), and expression plasmids using FuGENE 6 reagent (Roche Applied Science) in accordance with the manufacturer's instructions.
Indirect Immunofluorescence-Day 7 BMDCs from wild-type VDR and VDR knockout mice were seeded on 10-well microscope slides (Erie Scientific Co., Portsmouth, NH), fixed in 3% paraformaldehyde for 15 min on ice, washed three times with phosphate-buffered saline (PBS), permeabilized in 0.2% Triton X-100 in PBS for 10 min, and washed with PBS. After blocking for 1 h in PBS and 5% nonfat dry milk, cells were incubated with anti-mouse RelB polyclonal antibody (1:150 dilution) for 1 h at room temperature, followed by three washes with PBS and 5% nonfat dry milk. Finally, cells were incubated with secondary antibody (Alexa Fluor 488-conjugated goat anti-rabbit IgG, 4 g/ml) for 45 min in PBS and 5% nonfat dry milk, followed by three washes. Slides were mounted with Vectashield® mounting medium (Vector Laboratories, Inc., Burlingame, CA) and examined by confocal laser-scanning microscopy (LSM510, Carl Zeiss, Inc., Göttingen, Germany).
Expression Constructs and Reporter Plasmids-A polynucleotide fragment containing the entire coding region of the mouse VDR transcript (GenBank TM /EBI accession number D31969) was amplified by PCR using sequence-specific oligonucleotide primers (sense primer, 5Ј-CTGTGAGTCTTCCAGGAGAGCACC-3Ј; and antisense primer, 5Ј-TCAGGAGATCTCATTGCCAAACACC-3Ј) and cDNA prepared from activated murine T-cells and then ligated into the mammalian expression vector pcDNA3.1(ϩ) by the restriction sites HindIII and XbaI. The cloning of the human relB promoter (containing 1.1 kb of sequence 5Ј to the translational start site) and its corresponding NF-BI and NF-BII mutants into the pGL3-Basic vector (Promega) has been described previously (25). A fragment of genomic DNA containing 0.8 kb of sequence 5Ј to the start site of the mouse relB gene was isolated by screening a genomic library prepared from D3 embryonic stem cell DNA (mouse strain 129/Sv) with a mouse full-length cDNA probe. One phage encoding the relB promoter region was isolated and digested with XbaI and XhoI. The resulting 1.48-kb fragment was ligated into a modified pBluescript vector and then transferred to pGL3 upstream of the firefly luciferase reporter gene. Mutagenesis of plasmid constructs was performed using the QuikChange TM site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The following mutated plasmid constructs were generated: the human relB promoter with mutated VDRE motif A, the human relB promoter with mutated VDRE motif B, the human relB promoter with mutated VDRE motifs A and B, the mouse relB promoter with mutated VDRE, the human relB promoter with mutated NF-BI and NF-BII, and the mouse relB promoter with mutated NF-BI and NF-BII (see Fig. 2B for mutated sequences). The sequences of all wild-type and mutant constructs were confirmed by direct sequencing.
Gel Shift Assays and Immunoblotting-30 -33-base oligonucleotide probes were prepared that straddled the putative VDRE motifs present in the human and mouse relB promoter regions. Complementary strands were synthesized and annealed to sense strands at a molar ration of 1:1 in 100 mM Tris and 50 mM NaCl (pH 7.5) by heating to 100°C for 10 min and then cooling down to room temperature slowly. The annealed double-stranded oligonucleotides were labeled with [␥-32 P]ATP and T4 polynucleotide kinase (Roche Applied Science). Unincorporated oligonucleotides were removed using a NucTrap® probe purification column (Stratagene). Non-radiolabeled competitor oligonucleotides containing sequence for the mouse osteopontin VDRE and the human AP-1 (c-Jun)-binding sequence were used as positive and negative controls, respectively. Recombinant human full-length VDR and RXR␣ were prepared as glutathione Stransferase fusion proteins (26). Labeled oligonucleotides (2 pmol), with or without varying ratios of competitor oligonucleotides (10:1, 50:1, and 100:1), were mixed with VDR and RXR␣ (1 g each) in the binding buffer. The reaction mixtures were incubated at room temperature for 30 min and then separated by electrophoresis on 4% polyacrylamide gels in 0.25ϫ Tris borate/EDTA buffer. Dried gels were exposed to x-ray film.
Luciferase Reporter Assays-Mouse D2SC1 cells were seeded in 6-well plates at 5 ϫ 10 5 cells/well. Twenty-four hours later, the cells were transfected with 1 g of plasmid-encoded promoter construct and 10 ng of pRL-TK plasmid (encoding Renilla luciferase under the control of the thymidine kinase promoter) as an internal control. In some experiments, the cells were cotransfected with 0.5 g of mouse VDR expression construct in pcDNA3.1. Ten hours later, the medium was removed and replaced with control medium or with medium containing D 3 analog at final concentrations of between 10 Ϫ12 and 10 Ϫ8 M with or without 10 Ϫ7 M dexamethasone. After an additional 24 h, the cells were harvested and assayed for reporter gene activity and Renilla luciferase activity using the dual-luciferase assay kit (Promega) according to the manufacturer's instructions. Final results for each sample were recorded as Renilla adjusted relative light units.
Data Analysis-All experiments were carried out a minimum of three times with consistent results. For all reporter assays, duplicate or triplicate samples for each condition were prepared, and final results are expressed as means Ϯ S.D. Statistical differences between individual experimental conditions were determined using two-tailed, unpaired Student's t test with significance assigned to p Ͻ 0.05.

Inhibited DC Expression of RelB by 1␣,25-(OH) 2 D 3 Agonist Is a VDR-dependent Process-
To determine whether 1␣,25-(OH) 2 D 3 -mediated inhibition of RelB in DCs is dependent on the physiologic receptor (VDR), BMDCs were generated from wild-type VDR and VDR-deficient mice in the absence or presence of an optimized concentration (23) of D 3 analog and were immunofluorescently stained for RelB (Fig. 1). Cultures treated with the glucocorticoid dexamethasone were also examined. Untreated day 7 BMDCs stained strongly for RelB, with many cells having intranuclear as well as cytoplasmic staining. A clear reduction in RelB immunofluorescence was induced by both D 3 analog and dexamethasone in wild-type VDR BMDCs, but only by dexamethasone in VDR-deficient BMDCs. The results are consistent with a VDR-mediated inhibitory action of D 3 analog on RelB expression.
Human and Mouse relB Promoter Regions Contain Putative VDREs That Bind to the VDR⅐RXR␣ Complex-Genomic DNA sequences 5Ј to the start codons of the human and mouse relB genes were examined for potential VDREs. With the canonical DR3 VDRE hexamer ((A/G)G(T/G)TCA) as a benchmark (18), motifs consisting of two hexameric repeats conforming to an N(G/C)N(T/A)(G/C)(T/A) sequence and separated by three nucleotides were sought. Two such sequences were identified in the human relB promoter region (designated as human relB motif A (Ϫ799 to Ϫ785) and human relB motif B (Ϫ443 to Ϫ429)), and one was identified in the mouse promoter region (designated as the mouse relB motif (Ϫ602 to Ϫ588)). The positions and sequences of these motifs are illustrated in Fig.  2A. The sequences of the mutated motifs that were generated for use as controls in subsequent experiments are shown in Fig.  2B. The abilities of these three putative VDREs to complex with VDR and RXR␣ were tested in gel shift experiments (Fig.  3). Radiolabeled oligonucleotides containing human motifs A and B were found to complex with VDR and RXR␣ together, but not with either protein alone (Fig. 3A). The binding was competed in a dose-dependent fashion by non-radiolabeled oligonucleotides containing the same sequence or the sequence of a positive regulatory VDRE from the mouse osteopontin pro-moter (26), but not by oligonucleotides containing an AP-1binding site. Radiolabeled oligonucleotides in which the putative VDREs were mutated demonstrated an absent or markedly reduced ability to complex with VDR and RXR␣. Comparable results were obtained with oligonucleotides incorporating the mouse wild-type and mutant relB motifs (Fig. 3B). The relative affinity of the mouse motif for VDR and RXR␣ was compared with that of the human motifs and with the osteopontin VDRE in a competitive gel shift assay (Fig. 3C). The ability of human motifs A and B to compete with the radiolabeled mouse motif was less than that of the mouse motif itself, whereas the osteopontin VDRE competed more potently than any of the relB sequences. At a 10-fold excess of non-radiolabeled oligonucleotide, the mouse relB motif was associated with a 50% reduction in the density of the shifted band compared with 16, 30, and 93% for human motifs A and B and the osteopontin motif, respectively (Fig. 3C). We concluded that the identified sequences from the human and mouse relB promoter regions represent bona fide binding motifs for the VDR⅐RXR␣ complex, with the single mouse sequence having greater affinity than either of the two human sequences.

Human and Mouse relB Promoter Activities Are Negatively Regulated by 1␣,25-(OH) 2 D 3 Agonist in a Manner That Is Dependent on the Putative VDREs and That Is Enhanced by
Increased VDR Expression-Promoter region sequences from the human and mouse relB genes, including the putative VDREs, were ligated into a luciferase-encoding plasmid and employed in reporter assays using the murine DC-derived cell line D2SC1 (24). Detectable low level expression of VDR by this cell line was confirmed at the mRNA and protein levels (data not shown). Constructs were also generated in which the putative VDRE motifs were mutated to sequences shown by gel shift to have little affinity for VDR⅐RXR␣ (Figs. 2B and 3). For the human construct, motifs A and B were mutated singly and together. Promoter activities for this panel of reporter constructs were measured in D2SC1 cells in the absence and presence of D 3 analog and are expressed as the percent reduction associated with D 3 analog treatment (Fig. 4, two similar experiments shown). Both human and mouse wild-type promoter activities were significantly inhibited by 1␣,25-(OH) 2 D 3 agonist. In Ͼ10 separate experiments, D 3 analog treatment was associated with a consistent significant reduction in both human and mouse wild-type promoter activities that varied between 30 and 80%. In contrast, the human double mutant VDRE and mouse mutant VDRE promoters were minimally inhibited (0 -10% in multiple experiments). The human single mutants were inhibited by D 3 analog to a lesser degree than the human wild-type promoter, although the difference did not consistently reach statistical significance.
The influence of VDR expression levels on inhibition of human and mouse relB promoter activities by D 3 analog was next examined using the same reporter assay protocol with cotransfection of a plasmid encoding mouse VDR or an empty expression vector (Fig. 5). The concentration of D 3 analog was titrated from 10 Ϫ12 to 10 Ϫ8 M, and results were compared with untreated D2SC1 cells in the absence or presence of VDR overexpression. For both human and mouse relB promoters, the absolute promoter activity was significantly lower, and the percent reduction compared with that in untreated cells was greater at all concentrations of D 3 analog for the VDR-overexpressing cells. For example, at the suboptimal concentration of 10 Ϫ12 M D 3 analog, the percent reduction in promoter activity for cells overexpressing VDR was 40% for the human relB promoter and 45% for the mouse relB promoter compared with 25 and 9%, respectively, for cells not overexpressing VDR. At the optimal concentration of 10 Ϫ10 M, the equivalent results

FIG. 1. Inhibition of RelB expression in BMDCs by D 3 analog requires VDR expression.
Shown is the immunofluorescent detection of RelB in DCs derived from the bone marrow of wild-type VDR (upper panels) and VDR knockout (lower panels) mice. Bone marrow cultures were carried out with no addition (CONTROL) or in the presence of D 3 analog or of dexamethasone and were stained for RelB. Cytoplasmic and nuclear staining of DCs was clearly present in control cultures from both animals (left panels). Reduced staining for RelB was evident following D 3 analog treatment of cultures from wild-type VDR bone marrow, but not from VDR knockout bone marrow (middle panels). Exposure of either wild-type VDR or VDR knockout bone marrow cultures to dexamethasone resulted in reduced RelB immunofluorescence (right panels).
The results clearly support the contention that the VDR⅐RXR␣-binding motifs identified in the human and mouse relB promoters represent negative regulatory VDREs and are necessary for 1␣,25-(OH) 2 D 3 -mediated inhibition of RelB expression in DCs. Furthermore, the magnitude of 1␣,25-(OH) 2 D 3 -mediated inhibition of relB gene transcription in DCs is influenced by the expression level of VDR.

Inhibition of the relB Promoter by 1␣,25-(OH) 2 D 3 Agonist Is Independent of NF-B-REs, but the Additive Effects of Glucocorticoid Are Mediated through NF-B-REs-Transcriptional
expression of the human relB gene is positively regulated by two NF-B-REs (25). As 1␣,25-(OH) 2 D 3 has been reported to interfere with NF-B signaling (21,27), the effect of eliminating the NF-B-REs on relB promoter activity in D 3 analogtreated D2SC1 cells was determined. Cells were cotransfected with VDR along with the human wild-type relB reporter construct or with a construct in which the two NF-B-REs were inactivated by mutagenesis (see Fig. 2B for the sequences of wild-type and mutant NF-B-REs) and were exposed to graded concentrations of D 3 analog. As shown in Fig. 6, reporter activity from the human mutant NF-B-RE promoter was consistently lower than that from the wild-type promoter, but the degree of inhibition by each concentration compared with that in untreated cells was very similar for both constructs. At D 3 analog concentrations of 10 Ϫ12 , 10 Ϫ10 , and 10 Ϫ8 M, the wildtype promoter activity was inhibited by 40, 86, and 89%, respectively, whereas equivalent degrees of inhibition for the mutant NF-B-RE promoter were 58, 92, and 89%. The ability of a glucocorticoid agonist (dexamethasone) to additively inhibit relB promoter activity in combination with D 3 analog was then tested using the human wild-type and mutant NF-B-RE constructs (Fig. 7A). In contrast, no additional dexamethasoneassociated inhibition of the mutant NF-B-RE promoter occurred. The equivalent NF-B-REs from the mouse relB promoter were also identified and mutated (see Fig. 2B). As shown in Fig. 7B, the effect of dexamethasone in combination with D 3 analog on mouse wild-type and mutant NF-B-RE relB promoters was closely comparable to the results obtained with the human construct. In the experiments shown, the addition of 10 Ϫ7 M dexamethasone to 10 Ϫ10 M D 3 analog resulted in an increase in the degree of inhibition of promoter activity from 28 to 55% for the human wild-type relB promoter and from 31 to 52% for the mouse relB promoter. For the human and mouse mutant NF-B promoters, the degree of inhibition for D 3 analog alone was 33 and 26%, respectively, compared with 25 and 25% for D 3 analog and dexamethasone. Comparable results were obtained in multiple repeat experiments. We concluded that 1␣,25-(OH) 2 D 3 -mediated inhibition of relB promoter activity in DCs operates independently of NF-B-REs, but is capable of additively inhibiting relB gene transcription when combined with an antagonist of NF-B signaling such as glucocorticoid. DISCUSSION The results of this study demonstrate a direct negative regulatory effect of 1␣,25-(OH) 2 D 3 on the promoter region of the gene encoding RelB, a pivotal NF-B component in the regulation of DC differentiation and maturation (11)(12)(13)(14). The potential binding motifs for VDR⅐RXR␣ that were identified in both mouse and human promoters proved to have specific affinity for recombinant VDR⅐RXR␣ in gel shift experiments. Further- FIG. 2. A, shown are the putative VDREs in the promoter regions of human and mouse relB genes. The sequences and positions of two potential VDREs within the human relB gene 5Ј to the start codon (ATG) and of one potential VDRE within the same region of the mouse relB gene are shown. The human sequences were designated motifs A and B, respectively. The hexameric sequences corresponding to the putative VDR⅐RXR␣-binding sites are shown in boldface. B, the positions and sequences of nucleotide motifs from human and mouse relB promoter regions that were identified as VDREs and NF-B-REs are listed along with the sequences to which these motifs were mutated for experimental controls. Mutated nucleotides are shown in italics. more, using a panel of luciferase reporter constructs, it was possible to demonstrate that 1␣,25-(OH) 2 D 3 -and D 3 analogmediated negative regulation of relB promoter activity occurs in a DC-derived cell line and is dependent upon the presence of the VDR⅐RXR␣-binding motifs. Although two VDREs were identified in the human promoter and only one in the mouse promoter, the inhibitory effects of 1␣,25-(OH) 2 D 3 agonists on the two promoters were closely comparable, an observation that may be explained by the relatively higher affinity of the mouse VDRE for VDR⅐RXR␣ in competitive gel shift assays. The physiologic relevance of the mechanism is supported by the demonstration of VDR-dependent attenuation of RelB expression in DCs derived from murine bone marrow.
Manipulation of the NF-B signaling pathway has garnered substantial attention as a promising therapeutic intervention for inflammatory and immune-mediated diseases (6 -10). The primary impetus for applying NF-B inhibition to autoimmunity and transplantation stems from the recognition that a diverse array of triggering stimuli for cognate immunity are channeled through this intracellular pathway (1,5,(7)(8)(9). The central role for DCs in orchestrating antigen-specific T-cell and B-cell responses (1,2) and the essential function of RelB in DC differentiation and immunostimulatory capacity (11)(12)(13) provide an excellent example of how discrete manipulation of NF-B activity might be applied to the prevention or treatment of inappropriate immune activation. This contention is supported by the recent demonstration by Martin et al. (13) that direct inhibition of RelB expression in bone marrow cultures results in the generation of immature DCs that are associated with antigen-specific suppression of secondary T-cell responses when administered to sensitized animals.
The immunomodulatory effects of the vitamin D endocrine system have been studied for Ͼ20 years, and in vitro and in vivo studies have identified the DC as a primary target of 1␣,25-(OH) 2 D 3 -mediated inhibitory effects (15, 23, 28 -31). The expression of multiple maturation-induced proteins is inhibited in DCs exposed to 1␣,25-(OH) 2 D 3 agonists (23, 28 -30). Functionally, the phenotype of 1␣,25-(OH) 2 D 3 -conditioned DCs is an immature one with relatively poor capacity to induce antigenspecific T-cell proliferation and a tendency to promote tolerance to minor histocompatibility alloantigens in vivo (15, 23, 28 -30). Furthermore, in studies by Gregori et al. (32,33), the administration of 1␣,25-(OH) 2 D 3 or related analogs, with or without additional immunosuppressive agents, was associated with protection against autoimmunity and allograft rejection and with expansion of CD4 ϩve /CD25 ϩve regulatory T-cell populations. These observations were suggested to result from in vivo modulation of DC/T-cell interactions to favor the generation of antigen-specific regulatory T-cell populations, a mechanism that has been evoked by others to explain the tolerance induced by inoculation with or targeting of antigen to immature DCs (3,34). Although it is clear that additional individual genes may be regulated in cells of the immune system by 1␣,25-(OH) 2 D 3 agonists (15), our finding of direct transcriptional suppression of a key signaling protein (RelB) represents a discrete VDRmediated mechanism whereby such agents may promote  5-13). Competing oligonucleotides were the individual relB motifs themselves (Self; lanes 5-7), a canonical VDRE from the mouse osteopontin gene promoter (lanes 8 -10), and an AP-1 response element (AP-1 RE; lanes [11][12][13]. Oligonucleotides containing mutated sequence at the putative VDREs (Mutant) were also tested (lanes 14 -17). In C, the radiolabeled mouse relB motif was incubated with VDR and RXR␣ in the absence of a competing non-radiolabeled oligonucleotide (lane 1) or in the presence of graded amounts of competing oligonucleotides containing the same sequence (lanes 2-4), the putative human relB VDREs (motif A (Human A; lanes 5-7) and motif B (Human B; lanes 8 -10)), and the mouse osteopontin VDRE (lanes 11-13). The proportionate reduction in the density of the major shifted band compared with the control reaction (lane 1) is shown at the bottom of the lanes for each competitive reaction (% RED.).
"tolerogenic" antigen presentation. Furthermore, the separate effects of D 3 analog and glucocorticoid on the relB promoter provide a mechanistic basis for the additive or synergistic effects of 1␣,25-(OH) 2 D 3 agonists on immune-mediated disease (35).
The demonstration that the magnitude of transcriptional repression of relB by 1␣,25-(OH) 2 D 3 agonists is influenced by the level of VDR expression has important implications for in vivo potency of immunomodulatory D 3 analogs. Human tonsillar DCs (generally a site of ongoing active immune responses) constitutively express VDR (36), whereas lymphocyte-depleted mouse splenocytes (a mixture of macrophage/monocytes and DCs) demonstrate induction of VDR following a retroviral infection (37). Hewison et al. (38) have also demonstrated that VDR expression undergoes regulation during DC differentiation from monocytes. The fact that VDR is an inducible protein within immune cell populations suggests that immunotherapy using D 3 analogs is likely to target the DCs involved in an emerging or established immune injury. With regard to RelB repression, this would imply that newly recruited DCs and DCs undergoing maturation-inducing stimulation may be specifi-cally modified by 1␣,25-(OH) 2 D 3 and related analogs to retain an immature phenotype. Whether VDR is regulated in DCs by additional endogenous or exogenous factors remains to be determined. It is interesting, however, that Cantorna et al. (39) have identified an interplay between dietary calcium and protection against autoimmunity in 1␣,25-(OH) 2 D 3 -treated animals. Polymorphisms of the VDR gene have also been linked with predisposition to immune-mediated disease (40). Although the mechanisms for these observations are not known at present, it is likely that environmental and genetic factors that influence base-line and inducible VDR expression also affect susceptibility to 1␣,25-(OH) 2 D 3 -mediated immunosuppression.
The possible mechanisms whereby 1␣,25-(OH) 2 D 3 bound to VDR⅐RXR␣ negatively regulates transcription of certain genes FIG. 4. Human and mouse relB promoter activities are negatively regulated by D 3 analog via a VDRE-dependent mechanism. The DC-derived cell line D2SC1 was transfected with a panel of luciferase reporter constructs containing human or mouse wild-type (WT) or mutant (Mut) relB promoter sequence in the presence or absence of 10 Ϫ10 M D 3 analog. For each construct, the promoter activity (measured as relative light units (RLU)) was measured for multiple replicates of untreated and D 3 analog-treated cells, and the result are expressed as the mean Ϯ S.D. of the proportionate reduction associated with D 3 analog treatment (% reduction in relative light units). The results of two similar experiments are shown. †, p Ͻ 0.05 compared with the results for the human wild-type promoter; ‡, p Ͻ 0.05 compared with the results for the mouse wild-type promoter.
FIG. 5. Overexpression of VDR results in an increase in the sensitivity of the human and mouse relB promoters to inhibition by D 3 analog. D2SC1 cells were transiently transfected with luciferase reporter constructs containing human wild-type (upper panel) and mouse (lower panel) relB promoter sequences and were cotransfected with empty plasmid (VDR Ϫ) or a VDR expression plasmid (VDR ϩ). An immunoblot for VDR is shown to illustrate VDR overexpression in VDR ϩ cells. Following transfection, the cell populations were exposed to graded concentrations of D 3 analog (from 0 to 10 Ϫ8 M), and promoter activity was measured (in relative light units) and is expressed as means Ϯ S.D. The proportionate reduction in promoter activity compared with untreated cells is also shown for individual conditions as a percentage above each bar. †, p Ͻ 0.05 compared with VDR Ϫ cells.
include competitive displacement of positive regulatory transcription factor complexes, recruitment of corepressor proteins, and direct interference with assembly of the transcriptional machinery. Regarding 1␣,25-(OH) 2 D 3 -mediated inhibition of immune-related genes, Cippitelli and Santoni (20) demonstrated that two VDR⅐RXR␣-binding regions in the interferon-␥ promoter are responsible for negative regulation of this gene, with the potential to interfere with both AP-1 recruitment and transcriptional complex assembly. D'Ambrosio et al. (21) characterized the inhibition by 1␣,25-(OH) 2 D 3 of IL-12 p40 promoter activity in DCs as being mediated through interference with NF-B transcriptional activation, but did not detect direct binding of VDR to this promoter region. Alroy et al. (19) and Takeuchi et al. (41) identified a region within the human IL-2 promoter in which a VDR⅐RXR␣-binding domain and an NFATp (nuclear factor of activated T-cells p)/AP-1 domain overlap. Interestingly, the DNA region to which VDR⅐RXR␣ bound did not closely conform to any reported VDREs, and the corresponding mouse sequence failed to bind VDR⅐RXR. DNAbound VDR⅐RXR␣ was shown to complex with NFATp and to destabilize its association with AP-1 components. Towers and Freedman (42) characterized a variant VDRE half-site in the promoter of the granulocyte/macrophage colony-stimulating factor gene that overlaps with an NFATp/AP-1 site and that mediates transcriptional repression upon binding VDR alone. The VDREs we have identified in the mouse and human relB promoters conform more closely to canonical DR3 VDREs than those described for the IL-2 and granulocyte/macrophage colony-stimulating factor promoters and do not detectably bind VDR alone. We have not, to date, identified a potential overlapping binding site for positive regulatory complexes associated with the relB promoter VDREs, and our results with mutant NF-B-RE promoter constructs rule out the possibility that binding of VDR⅐RXR␣ to the VDREs acts by interfering with the function of these NF-B-REs. The characterization of nuclear proteins associated with DNA-bound VDR⅐RXR␣ complexes in DCs and of the other signaling pathways involved in relB transcription may provide additional insights.
In conclusion, we have shown that the promoter region of the gene encoding the NF-B family member RelB is a direct target of the vitamin D system in mouse and human via one or more non-classical hexameric repeats that directly bind VDR⅐RXR␣ and that mediate negative transcriptional regulation. The unique influence of RelB expression on DC function identifies this novel mechanism as a key element in the immunotherapeutic properties of 1␣,25-(OH) 2 D 3 and its analogs. D2SC1 cells were transfected with luciferase reporter constructs containing human (A) or mouse (B) wild-type or mutant NF-B (NF-kappa B Mut) relB promoter sequences and were untreated (No Addition) or exposed to 10 Ϫ10 M D 3 analog in the absence (D3 Analog) or presence (D3 Analog ϩ Dex) of 10 Ϫ7 M dexamethasone. Results are expressed as means Ϯ S.D. of the relative light units. The proportionate reduction in promoter activity compared with untreated cells is also shown for individual conditions as a percentage above each bar. †, p Ͻ 0.05 compared with untreated cells; ‡, p Ͻ 0.05 compared with D 3 analog treatment alone.