The functions of dendritic cells (DCs) are tightly regulated such that protective immune responses are elicited and unwanted immune responses are prevented. 1α25-dihydroxyvitamin D3 (1α25(OH)2D3) has been identified as a major factor that inhibits the differentiation and maturation of DCs, an effect dependent upon its binding to the nuclear vitamin D receptor (VDR). Physiological control of 1α25(OH)2D3 levels is critically dependent upon 25-hydroxyvitamin D3-1α-hydroxylase (1αOHase), a mitochondrial cytochrome P450 enzyme that catalyzes the conversion of inactive precursor 25-hydroxyvitamin D3 (25(OH)D3) to the active metabolite 1α25(OH)2D3. Using a human monocyte-derived DC (moDC) model, we have examined the relationship between DC VDR expression and the impact of exposure to its ligand, 1α25(OH)2D3. We show for the first time that moDCs are able to synthesize 1α25(OH)2D3 in vitro as a consequence of increased 1αOHase expression. Following terminal differentiation induced by a diverse set of maturation stimuli, there is marked transcriptional up-regulation of 1αOHase leading to increased 1αOHase enzyme activity. Consistent with this finding is the observation that the development and function of moDCs is inhibited at physiological concentrations of the inactive metabolite 25(OH)D3. In contrast to 1αOHase, VDR expression is down-regulated as monocytes differentiate into immature DCs. Addition of 1α25(OH)2D3 to moDC cultures at different time points indicates that its inhibitory effects are greater in monocyte precursors than in immature DCs. In conclusion, differential regulation of endogenous 1α25(OH)2D3 ligand and its nuclear receptor appear to be important regulators of DC biology and represent potential targets for the manipulation of DC function.
Changes in dendritic cell (DC)3 function represent the net cellular response to a large repertoire of microenvironmental positive and negative stimuli (1). An understanding of how developing or terminally differentiated DCs integrate these signals is of crucial importance in terms of defining regulatory pathways that contribute to the homeostatic control of their functions.
One important regulatory mechanism involves the actions of the secosteroid hormone, 1α25-dihydroxyvitamin D3 (1α25(OH)2D3), the biologically active metabolite of vitamin D3, which influences the immune response at multiple levels (2), including at the level of the APC. For example, 1α25(OH)2D3 inhibits the release of inflammatory cytokines such as IL-12 (3) or TNF-α by macrophages (4). Furthermore, in vitro 1α25(OH)2D3 inhibits the differentiation and maturation of human monocyte-derived DCs (moDCs), leading to reduced cell surface expression of CD40, CD80, and CD86 costimulatory molecules, reduced allostimulatory function, and decreased survival (5, 6). Similar findings are observed in murine bone marrow-derived DCs which have been propagated in the presence of 1α25(OH)2D3 analogs (7). Indeed, infusion of small numbers of these altered DCs is sufficient to induce tolerance to male Ags in an HY mismatch model (7).
The major 1α25(OH)2D3-mediated effects upon DCs are mediated through a nuclear vitamin D receptor (VDR) (8). The ligand-VDR complex heterodimerizes with the 9-cis-retinoic acid liganded retinoid X receptor, before binding as a transregulatory complex to vitamin D3-responsive cis elements in target gene promoters (9). These genes include IL-12p35 and p40 (3), that are expressed upon CD40-mediated licensing of DCs and consequently the repressive effect of 1α25(OH)2D3 upon IL-12 secretion is lost in DCs derived from VDR knockout mice (7). Indeed, the lymph nodes in VDR knockout mice are increased in size as compared with wild type and the proportion of mature DCs within them is greater, suggesting that 1α25(OH)2D3 exerts an essential negative regulatory function in vivo that limits the degree to which DCs undergo spontaneous maturation (7).
The observed effects of 1α25(OH)2D3 upon APCs in vitro are at concentrations greater than that readily measured in vivo, suggesting that its influence upon DC function occurs across microenvironmental gradients where 1α25(OH)2D3 is generated locally. Thus, the impact of 1α25(OH)2D3 upon DC function is likely to be dependent not only upon cellular expression of VDR but also its local synthesis and or breakdown. 1α25(OH)2D3 is generated upon the 1-α hydroxylation of an inactive precursor metabolite, 25-hydroxyvitamin D3 (25(OH)D3), which exists at a 2–3 log higher concentration than the daughter molecule (10). The level of 1α25(OH)2D3 is, thus, critically dependent upon cellular expression of the enzyme, 25-hydroxyvitamin D3 (25(OH)D3) 1-α-hydroxylase (1αOHase) which catalyzes this reaction. 1αOHase is expressed predominantly in proximal renal tubular cells (11) where it is subject to tight regulation, but is also expressed in cells from extrarenal sites including skin, endothelium, lymphoid organs, and decidua (12, 13, 14). Macrophages constitute one of the primary sources of extrarenal 1α25(OH)2D3 production by virtue of their ability to up-regulate 1αOHase expression (15). Indeed, in the context of aberrant immune responses such as those associated with systemic sarcoidosis, unregulated 1α25(OH)2D3 synthesis can lead to defective calcium homeostasis (16).
Myeloid DCs can be differentiated from monocyte/macrophage precursors in vitro and share much of the cellular apparatus required to react to microbial pathogens and other proinflammatory signals (17). This suggests the possibility that the capacity to synthesize 1α25(OH)2D3 is conserved within DCs. Therefore, the aim of this study was to determine whether moDCs also generate 1α25(OH)2D3 and if so, under what conditions. Further, we aimed to establish the relationship between the synthesis of 1α25(OH)2D3 and the expression of the VDR in DCs. We confirm that differentiation and maturation of human moDCs is associated with significant 1α25(OH)2D3 synthesis, as a result of transcriptional up-regulation of 1αOHase. Furthermore, we demonstrate that as moDCs differentiate, there is a significant down-regulation of VDR expression. This differential regulation of inhibitory ligand and its nuclear receptor are indicative of potential autocrine/paracrine regulation of DC differentiation and function, and suggest possible targets for manipulation of DC function.
Materials and Methods
Generation of moDCs
PBMCs were derived from buffy coats of healthy donors by density centrifugation over Lymphoprep (Robbins Scientific, Sunnyvale, CA), resuspended in RPMI 1640/2% human AB serum, and allowed to adhere to plastic at 37°C for 2 h. Nonadherent cells were then removed and monocytes were cultured in IMDM (Life Technologies, Rockville, MD) plus 5% human AB serum (HD Supplies, Buckinghamshire, U.K.) plus 2 mM l2 flasks (Iwaki, Tokyo, Japan). In parallel experiments, moDCs were propagated under serum-free conditions in macrophage serum-free media (Life Technologies). 1α25(OH)2D3 (10−8 M), 25(OH)D3 (10−8–10−7 M), or vehicle (Sigma-Aldrich, St. Louis, MO) were added in some experiments either at the beginning of culture or at time points indicated in the text. Media were replenished at 3-day intervals. Immature DC were resuspended in the above media ± vitamin D3 metabolites (on day 7) and incubated in 12-well plates at a concentration of 5 × 105/ml and matured for 48 h in the presence of one of the following maturation stimuli: 5 μg/ml G28.5 CD40 Ab (kindly provided by Prof. J. Gordon, Department of Immunology, University of Birmingham, Birmingham, U.K.) plus 5 μg/ml goat anti-mouse Ig Fc cross-linker (Pierce, Rockford, IL), 50 ng/ml TNF-α (Peprotech, London, U.K.), 1000 U/ml IFN-γ, 1 μg/ml LPS Escherichia coli, or 20 μg/ml poly I:C (all obtained from Sigma-Aldrich). In control experiments, 5 μg/ml nonspecific mouse IgG1 MOPC21 (Sigma-Aldrich) plus 5 μg/ml goat anti-mouse Ig Fc cross-linker (Sigma-Aldrich) or PBS were substituted for the various maturation stimuli. In experiments to test the regulation of 1αOHase, 20 μM p38 mitogen-activated protein kinase (MAPK) inhibitor SB203580 (Sigma-Aldrich) or 50 μM of the NF-κB inhibitor N-α-tosyl-l-phenylalanine chloromethyl ketone (TLCK; Sigma-Aldrich) were added 60 min before the maturation stimulus.
Flow cytometric analysis
The cell surface phenotype of moDCs or DCs propagated in the presence of vitamin D3 metabolites was analyzed using the following mAbs: anti-CD1a-fluorescecin isothiocyanate (HI149;IgG1), anti-CD14-PE (M5E2. IgG2a), anti-CD83-PE (HB15E, IgG1), anti-CD86-PE (IT2.2, IgG2b), anti-HLA-DR-Cychrome (G46.6, IgG2a; San Diego, CA), and the isotype controls labeled with the appropriate fluorochrome (MOPC21, IgG1; G155–178, IgG2a; and 27–35, IgG2b) were obtained from BD PharMingen (San Diego, CA). The staining of the samples was measured on a FACS Coulter Epics XL (Beckman Coulter, Fullerton, CA), using System II software (Beckman Coulter), and was analyzed with WinMDI software (from Prof. J. Trotter, The Scripps Research Institute, San Diego, CA).
In experiments to test the effect of vitamin D3 metabolites upon IL-12 production by moDCs, day 7 DCs were incubated for 48 h in 48-well plates at a concentration of 5 × 105/ml with CD40 ligand (CD40L)-transfected or control nontransfected L cells at a ratio of 2:1. IL-12p70 and IL-12p40 levels were evaluated by ELISA kits according to the manufacturer’s instructions (BD PharMingen). The IL-12p40 kit detects the IL-12p40 homodimer or monomer but not the bioactive IL-12p70 heterodimer.
Cytospins were prepared from monocytes, immature moDC, and mature moDC, washed with PBS and stained for 1 h at room temperature with sheep anti-1αOHase Ab (12), which had been diluted 1/200. After washing, the slides were incubated with secondary biotinylated donkey anti-sheep Ab (Binding Site, Birmingham, U.K.) for 30 min, followed by a second 30-min incubation with a streptavidin-biotin complex (strepABComplex; DAKO, Cambridgeshire, U.K.). The staining was visualized using 3′-diaminobenzidine tetrahydrochloride (Sigma-Aldrich). Finally, the cells were counterstained with Mayer’s hematoxylin. The percentage of stained cells and intensity of staining were measured by counting a minimum of 200 cells per slide. All analyses were performed single blind.
These were performed as previously described (18). For the detection of NF-κB binding activity, an oligonucleotide containing the HIV-long terminal repeat NF-κB consensus binding sequence (5′-GAT CAG GGA CTT TCC GCT GGG GAC TTT CC-3′) was used.
Phospho-p38 immunoblots were performed as described previously (18) using a phosphospecific p38 MAPK (Thr180/Tyr182) Ab (New England Biolabs, Beverly, MA) and were compared with total p38 MAPK.
Quantitative RT-PCR analysis of 1α-OHase, VDR, and 25(OH)D3 24-hydroxylase (24(OH)ase) mRNA expression
1αOHase, VDR, and 24(OH)ase mRNA levels were analyzed using an ABI 7700 sequence detection system (PE Biosystems, Warring, U.K.) using previously reported protocols (12). RNA was extracted from cell pellets using the StrataPrep total RNA miniprep kit (Stratagene, Amsterdam, The Netherlands) and was reverse-transcribed using AMV reverse transcriptase (Promega, Madison, WI) and random hexamers in 40-μl reaction volumes according to manufacturer’s instructions. Amplification of cDNAs was performed in 25-μl volumes on 96-well plates, in a reaction buffer containing TaqMan Universal PCR Master Mix, 3 mM Mn(Oac)2, 200 μM dNTPs, 1.25 U AmpliTaq Gold polymerase, 1.25 U AmpErase UNG, 150 nmol TaqMan probe, 900 nmol primers, and 50 ng of cDNA. All reactions were multiplexed with the housekeeping gene 18S rRNA, provided as a preoptimized control probe (PE Biosystems) enabling data to be expressed in relation to an internal reference to allow for differences in sampling and RT efficiency. Data were obtained as cycle threshold (Ct) values (the cycle number at which logarithmic PCR plots cross a calculated threshold line) according to the manufacturer’s guidelines and were used to determine ΔCt values (ΔCt = Ct of the target gene − Ct of the housekeeping gene). Measurements were conducted in triplicate for each individual experiment. All PCRs were conducted with the primers and probes outlined below using the following reaction conditions: 50°C for 2 min, 95°C for 10 min; 44 cycles of 95°C for 15 s; 60°C for 1 min.
Primers and probes for PCRs
PCR primer and probe sequences for 1αOHase were as follows: forward primer, 5′-CACCCGACACGGAGACCTT-3′; reverse primer, 5′-TCAACAGCGTGGACACAAACA-3′ and TaqMan probe, 5′-TCCGCGCTGTGGGCTCGG-3′. For 24(OH)ase: forward primer, 5′-CAAACCGTGGAAGGCCTATC-3′; reverse primer, 5′-AGTCTTCCCCTTCCAGGATCA-3′ and Taqman probe, 5′-ACTACCGCAAAGAAGGCTACGGGCT-3′. For VDR: forward primer, 5′-CTTCAGGCGAAGCATGAAGC-3′; reverse primer, 5′-CCTTCATCATGCCGATGTCC-3′ and TaqMan probe, 5′-AAGGCACTATTCACCTGCCCCTTCAA-3′. In each case, the housekeeping gene used to derive ΔCt values was 18S rRNA which was analyzed using primers and probes provided by the manufacturer (PE Biosystems).
Cells were washed in serum-free RPMI 1640 and were then resuspended at 1 × 106 cells/ml. Aliquots of this suspension (400 μl) were then incubated with 10 nM [3H]25(OH)D3 (specific activity 180 Ci/mmol; Amersham, London, U.K.) for 4 h at 37°C. The reaction was terminated by freezing at −20°C. Cell extracts and medium were combined and vitamin D metabolites were extracted in 2.5 ml of chloroform:methanol (4:1 vol:vol). Following evaporation of the organic phase, steroids were resuspended in 50 μl of dichloromethane:isopropanol (9:1 vol:vol). Standard lanes were included which contained only [3H]25(OH)D3 or [3H]1α25(OH)2D3. Production of [3H]1α25(OH)2D3 was measured on a Bioscan System 200 imaging TLC plate scanner (Bioscan, Edmonds, WA). Data were reported as fmol of 1α25(OH)2D3 produced per hour per 106 cells.
Analysis of nuclear binding of 1α25(OH)2D3
Numbers of intracellular receptors for 1α25(OH)2D3 were assessed by whole cell nuclear association assays. Briefly, cells were resuspended in serum-free RPMI 1640 to give 0.5 × 107 cells/ml. Aliquots (100 μl) of these suspensions were then added to glass tubes containing a single saturating dose (5 nM) of [3H]1α25(OH)2D3 (specific activity 180 Ci/mmol; Amersham) in the presence or absence of a 200-fold excess of unlabeled 1α25(OH)2D3 (to determine nonspecific binding). Cells were then incubated for 1 h to allow association of ligand-VDR complexes with nuclei and then washed twice with 1 ml of PBS at 4°C to remove unincorporated [3H]1α25(OH)2D3. The final cell pellets were incubated with 0.5 ml of lysis buffer (0.25 M sucrose, 20 mM HCl, 1.1 mM magnesium chloride, and 0.5% Triton X-100, pH 7.4) and centrifuged at 1000 × g for 5 min. This step was repeated and the resulting crude nuclear pellet was resuspended in 100 μl of PBS and 100 μl of ethanol, and transferred to 4 ml of scintillation fluid before scintillation counting. Maximal binding was determined as fmol 1α25(OH)2D3 bound per 106 cells and then converted to VDR per cell using Avogadro’s constant.
Statistical analysis was performed using one-way or repeated measure ANOVA, with posthoc testing using the Student Newmans Keuls multiple comparison posttest (Instat version 2.04a; GraphPad Software, San Diego, CA).
Human moDCs synthesize 1α25(OH)2D3
To determine whether human DCs are able to synthesize 1α25(OH)2D3, monocytes were incubated in the presence of IL-4 and GM-CSF for 7 days and were then matured by CD40 ligation for 48 h. At various time points, RNA was prepared and real time RT-PCR was performed to determine 1αOHase mRNA expression. Differentiation and maturation of moDCs was accompanied by a marked increase in 1αOHase mRNA expression as compared to the starting population, increasing >10-fold in day 7 immature moDCs and >60-fold in moDC matured by CD40 ligation (Fig. 1⇓A). The amplified product corresponded to the expected size and sequence indicating that, in common with macrophages, the transcript is identical to that detected in proximal renal tubular cells (data not shown). We also examined the effect of other maturation stimuli representative of both T cell and microbial origin for their ability to induce the transcriptional up-regulation of 1αOHase. As shown in Fig. 1⇓B, under the conditions tested, CD40 ligation, TNF-α, IFN-γ, poly I:C, and LPS treatments of day 7 immature moDCs were associated with ∼5- to 12-fold induction of 1αOHase as compared with nontreated controls. Immunocytochemical staining of monocytes, immature and mature moDCs for the 1αOHase enzyme demonstrated a similar increase in protein expression, with dense staining, in particular of mature moDCs (Figs. 1⇓C and 3B). In parallel, 1α25(OH)2D3 synthesis, as estimated by thin layer chromatography of cellular extracts incubated in the presence of the [3H]25(OH)D3 precursor, was absent in monocytes and increased dramatically following differentiation into immature moDC (Fig. 1⇓, D and E). CD40 ligation-induced moDC maturation further boosted 1α25(OH)2D3 synthesis by >3-fold as compared with immature moDCs or mock-treated cells (Fig. 1⇓, D and E, and data not shown). moDC 1αOHase mRNA expression and 1α25(OH)2D3 synthesis was similar to that of GM-CSF-propagated macrophages before and following activation (Fig. 1⇓E and data not shown). In additional experiments, we confirmed that moDC generated under serum-free conditions also up-regulated 1αOHase gene transcription and activity, indicating that this phenomenon is reproducible under alternate culture conditions (data not shown). Thus, moDCs develop the capacity to synthesize 1α25(OH)2D3 in vitro to an extent which is comparable to macrophages and which is enhanced upon terminal differentiation induced by a diverse set of stimuli.
25(OH)D3 inhibits DC differentiation, maturation, and allostimulatory potential
Loss of the 1α-hydroxyl group from 1α25(OH)2D3 is associated with an ∼500-fold reduction in the binding affinity of ligand to VDR (19) and, thus, no VDR-mediated biological activity is demonstrable at physiological concentrations of 25(OH)D3 (20). The ability of human moDCs to generate 1α25(OH)2D3 from the inactive precursor 25(OH)D3 suggested that the latter could have inhibitory effects upon the differentiation or function of human moDCs. To test this hypothesis we exposed moDC cultures from the first day of culture to 10−8–10−7 M (10–100 nM) 25(OH)D3, representative of concentrations below or within the normal physiological range (19) and evaluated the effect upon DC differentiation, maturation, allostimulatory potential, and survival. In the same experiments, as a positive control, DCs were propagated in the presence of 10−8 M (10 nM) 1α25(OH)2D3, the concentration at which maximal inhibitory effects are observed (5, 6). By day 7, immature DCs cultured in the presence of GM-CSF and IL-4 were CD14-negative and had acquired surface CD1a expression (Fig. 2⇓, A and B). As in previous studies, DCs differentiated in the presence of 10−8 M 1α25(OH)2D3 showed a marked decrease in CD1a expression, while CD14 remained elevated. Addition of 10−8–10−7 M 25(OH)D3 at the initiation of culture also resulted in marked inhibition of CD1a expression at all concentrations tested. Furthermore, we observed a concentration-dependent failure to down-regulate CD14 with relatively minor effects observed at 10−8 25(OH)D3 but significant inhibition at 10−7 M 25(OH)D3, equivalent to that observed in the presence of 10−8 M 1α25(OH)2D3. Consistent with this, day 7 immature moDCs that had been cultured in the presence of 10−8–10−7 M 25(OH)D3 had impaired allostimulatory function as compared with controls (data not shown). As shown in Fig. 2⇓, C and D, and in accordance with other studies (5, 6), DCs generated in the presence of 10−8 M 1α25(OH)2D3 failed to up-regulate CD83 expression in response to CD40 ligation to the same degree as controls. Again, similar concentration-dependent effects were observed following the addition of 10−8–10−7 M 25(OH)D3, with greater inhibition at 10−7 M 25(OH)D3, equivalent to that observed in the presence of10−8 M 1α25(OH)2D3. In addition, DCs cultured in the presence of 10−8 M 1α25(OH)2D3 and higher concentrations of 25(OH)D3 both failed to up-regulate other molecules involved in the development of accessory functions, including CD80, CD86, and HLA-DR (Fig. 2⇓, C and D). Similar results were observed when DCs were matured in the presence of TNF-α (data not shown). Furthermore, as compared to controls, DCs differentiated in the presence of both 10−8 M 1α25(OH)2D3 and 10−7 M 25(OH)D3 demonstrated an almost complete failure in their capacity to produce IL-12p70 and IL-12p40 in response to CD40 ligation (Fig. 2⇓E) and accordingly had weak APC activity in an allogeneic mixed leukocyte reaction (Fig. 2⇓F). At the concentrations used and under the specific conditions described above, we observed no effects of 25(OH)D3 upon day 9 DC viability as evaluated by annexin V-FITC (FITC)-propidium iodide (PI) staining (control-FITC−PI− 82.4 ± 3.4%; 10−8 M 1α25(OH)2D3-FITC−PI− 80.0 ± 3.0%; and 10−7 M 25(OH)D3- FITC−PI− 84.0 ± 1.6%, n = 6). In conclusion, these results suggest that the active synthesis of 1α25(OH)2D3 from its inactive precursor restricts the differentiation of immature moDCs in vitro and their subsequent sensitivity to maturation stimuli. This phenomenon appears to operate at physiological concentrations of 25(OH)D3.
Regulation of 1αOHase
The above data are indicative of a link between immature moDC maturation and the up-regulation of 1αOHase expression. Integral to the maturation program of DCs in response to divergent inflammatory stimuli is the activation and cooperative interaction of the p38 MAPK and NF-κB pathways (21, 22). Therefore, we reasoned that specific inhibition of each of these pathways would also inhibit the up-regulation of 1αOHase expression. As shown in Fig. 3⇓, A and B, inhibition of the p38 MAPK pathway using the pyridinyl imidazole compound SB203580 resulted in a major reduction in the expression of the 1αOHase induced by TNF-α as demonstrated by immunocytochemical staining. Western blots of total cell lysates were performed in parallel to confirm that SB203580 inhibited p38 MAPK phosphorylation (Fig. 3⇓C). Furthermore, inhibition of nuclear NF-κB translocation using the proteosomal inhibitor TLCK was also associated with a failure to up-regulate 1αOHase expression induced by TNF-α (Fig. 3⇓, A and B). We confirmed in these experiments that TLCK effectively inhibited TNF-α-induced increases in nuclear NF-κB binding (Fig. 3⇓D). In contrast, SB203580 had no effect upon nuclear NF-κB binding, confirming that under the conditions used in these experiments, its effects were limited to inhibition of the p38 MAPK pathway alone. We confirmed, as reported previously (23), that inhibition of both these pathways inhibited up-regulation of cell surface CD83, but had no effect on moDC viability (data not shown). In additional experiments (Fig. 3⇓E), we confirmed that disabling either the p38 MAPK or NF-κB pathways led to almost complete inhibition of the transcriptional up-regulation of 1αOHase upon maturation.
In proximal renal tubule cells, 1αOHase gene expression is subject to exquisite feedback inhibition upon exposure to 1α25(OH)2D3, an effect that is dependent upon VDR expression (24). To test the effect of 1α25(OH)2D3 upon 1αOHase gene expression, immature moDCs were matured by CD40 ligation in the presence or absence of 10−8 M 1α25(OH)2D3. As shown in Fig. 3⇑E, 1α25(OH)2D3 had little effect upon the expression of 1αOHase mRNA. The lack of effect was also observed if 1α25(OH)2D3 was added >24 h before the application of the maturation stimulus, indicating that this was not due to slower kinetics of inhibition (data not shown). The resistance of 1αOHase gene expression to the negative regulatory effects of 1α25(OH)2D3 is also observed in other cells derived from extrarenal sites, such as keratinocytes (25) or macrophages (26). In these examples, the synthesis of 1α25(OH)2D3 may be controlled by the competing effects of another enzyme, 24(OH)ase, which inactivates 1α25(OH)2D3. 24(OH)ase is itself induced by 1α25(OH)2D3 through a VDR-mediated mechanism (27, 28). As shown in Fig. 3⇑F, in the absence of 1α25(OH)2D3, human moDCs do not express significant levels of 24(OH)ase. In contrast, exposure of differentiating moDCs to 1α25(OH)2D3 increased 24(OH)ase gene expression dramatically (up to 3000-fold) at early time points (day 3 of culture), but notably the degree of induction then fell by ∼6-fold on days 7 and 9. This fall in 24(OH)ase expression did not represent the consequence of 24(OH)ase-mediated metabolism of 1α25(OH)2D3, because the active ligand was continually replenished in the medium. Importantly, 25(OH)D3 had little effect upon 24(OH)ase gene expression on day 3, but consistent with its conversion to active ligand, induced 24(OH)ase expression at later time points to a degree similar to exogenous 1α25(OH)2D3.
In summary, 1αOHase up-regulation in moDCs is associated with the p38 MAPK-and NF-κB-dependent maturation of DCs, but its transcription is resistant to the negative regulatory effects of 1α25(OH)2D3. A degree of control for 1α25(OH)2D3 synthesis may be imparted by the induction of an inactivating enzyme, 24(OH)ase, particularly at early time points during differentiation.
VDR expression is down-regulated during DC differentiation
A major arbiter of the dendritic cellular response to either exogenously or endogenously derived 1α25(OH)2D3 ligand will be the level of expression of VDR and/or the presence of coactivators that are recruited to the nuclear receptor complex. One indirect measure of the integrity of this pathway within cells is the 1α25(OH)2D3-mediated induction of 24(OH)ase transcription (8). Our finding that the induction of 24(OH)ase transcription by 1α25(OH)2D3 in moDCs was greatest at early time points during DC differentiation, then followed by a significant decline, suggested the possibility that VDR levels change during DC differentiation. To evaluate VDR expression we performed real time RT-PCR from RNA extracts of moDC cultures at a number of time points. The DC-derived amplified product was identical in size to the product from the HL-60 cell line that is known to express VDR (not shown). Immature and mature moDCs demonstrated about a 4-fold reduction in VDR gene transcription as compared with the starting population (Fig. 4⇓A). These changes were mirrored by similar reductions in the level of VDR protein expression as evaluated by nuclear binding of the [3H]1α25(OH)2D3 ligand which showed a >2-fold reduction in VDR expression in immature and mature moDCs (Fig. 4⇓B). In contrast to 1αOHase, there was no significant difference between VDR expression (either at the transcript or protein level) in immature and mature DCs. Thus, the down-regulation of VDR was not clearly linked to the terminal differentiation of DCs. Consistent with this interpretation, the decline in VDR transcription was not arrested by specific inhibition of p38 MAPK or NF-κB pathways (data not shown). We also confirmed that neither 1α25(OH)2D3 nor 25(OH)D3 treatment of moDCs was associated with significant differences in VDR expression as compared with controls (data not shown).
The above experiments indicated that as moDCs differentiate, the actions of 1α25(OH)2D3 upon DC function could be dependent upon a number of factors, including the expression of VDR, 24(OH)ase activity, and endogenous 1α25(OH)2D3 synthesis. Other factors, such as changes in the expression of partner receptors/coactivators to VDR may also modulate the DC response to 1α25(OH)2D3. To examine how the net influence of these multiple factors influenced the response of moDCs to 1α25(OH)2D3, 10−8 M 1α25(OH)2D3 was added to media on days 1–3, 4–6, or 7–9 following the initiation of moDC culture. After each 72 h, the cells in each culture were washed and resuspended in fresh media with or without the addition of 1α25(OH)2D3. As shown in Fig. 4⇑C, the effect of 1α25(OH)2D3 ligand upon DC allostimulatory function was reduced the later it was added to developing DC cultures. In parallel, there was a progressive reduction in the inhibitory effect of 1α25(OH)2D3 upon expression of cell surface costimulatory molecule expression (Fig. 4⇑D). Thus, as moDCs differentiate in culture, they develop an increasing capacity to synthesize 1α25(OH)2D3, but in parallel a reduced ability to respond to its inhibitory actions.
In this work, we have demonstrated that human moDCs can synthesize the immunosuppressive steroid hormone 1α25(OH)2D3, suggesting autocrine and/or paracrine regulation of DC function. Synthesis of 1α25(OH)2D3 increases markedly as DC differentiate from monocyte precursors. However, and in direct contrast to monocytes, VDR expression is reduced in immature DCs. These data are consistent with a model in which differential regulation of ligand and nuclear receptor contributes to a relative resistance of immature and mature DCs to the inhibitory effect of ligand as compared with their monocyte precursors.
Recent evidence supports the existence of an autocrine element to the homeostatic control of DC development and function (29, 30, 31). For example, certain maturation stimuli induce the autocrine synthesis of IL-10, which appears to bind selectively to its receptor on the surface of immature but not mature DCs, the latter having lost cell surface expression of the receptor (30). Consistent with this, blocking IL-10 enhances sustained NF-κB activation and maturation of DCs (32). In another example, exposure of immature moDCs to LPS or TNF-α in the presence of serum leads to activation of the extracellular signal-regulated protein kinase-signaling pathway which is a negative regulator of DC maturation (33). These potential autoregulatory phenomena and the work in this study suggest that as DCs integrate positive proinflammatory signals inducing maturation, they simultaneously begin to generate negative signals that limit the extent to which they or other developing DCs can propagate an immune response. Our data suggests that active 1α25(OH)2D3 synthesis by DCs is able to limit the full acquisition of APC function by inhibiting costimulatory molecule expression and the secretion of bioactive IL-12p70. The lack of specific in vitro inhibitors of 1αOHase activity means that it has not been possible to exclude the possibility that the effects upon DC function observed in this study result, in part, from a direct effect of 25(OH)D3. However, loss of the 1α-hydroxyl group from 1α25(OH)2D3 is associated with an ∼500-fold reduction in the binding affinity of ligand to VDR (19) and no biological activity is demonstrable at physiological concentrations of 25(OH)D3 (20), suggesting that this is unlikely.
The potential role of such inhibitory signals may be to reduce the possibility of bystander activation of the immune response. By maintaining DCs in an immature state or modulating their survival (5, 6), negative feedback mediated by 1α25(OH)2D3 could conceivably contribute to the maintenance of peripheral tolerance to self-Ags. The phenotype of VDR knockout mice which have enlarged lymph nodes containing increased numbers of mature DCs is consistent with a nonredundant physiological role for this ligand in limiting DC activation and/or survival (7). Furthermore, transfer of vitamin D3 analog-treated donor-derived murine bone marrow-derived DCs can induce tolerance to alloantigens (7). Indeed, systemic administration of 1α25(OH)2D3 prolongs allograft survival, an effect associated with greatly reduced IL-12 production and costimulatory molecule expression by graft-associated DCs, and the induction of a CD4+CD25+ T cell population with regulatory properties (34). The generation of high local concentrations of 1α25(OH)2D3 upon ligation of the CD40 receptor, as shown in this study, suggests that one potential target is the interacting CD4+ T cell that up-regulates CD40L following signaling via its TCR. By limiting the capacity of DCs to become activated by CD40 ligation, 1α25(OH)2D3 could conceivably increase the threshold at which licensing occurs, such that low affinity T cells are less likely to generate unwanted immune responses. Alternatively, the actions of 1α25(OH)2D3 may directly affect CD4+ T cells, perhaps by skewing them toward a Th2 phenotype (35), or modulating their sensitivity to Fas-mediated activation-induced cell death (36), or, in conjunction with glucocorticoids, enhancing the development of IL-10-producing regulatory T cells (37).
The ability of any particular cell type to generate 1α25(OH)2D3 will depend upon regulation of 1αOHase expression/activity, access to the substrate (25(OH)D3) and the competing effects of inhibitory enzymes, such as 24(OH)ase. This study reaffirms important differences between renal and extrarenal sites of 1α25(OH)2D3 production. First, DCs and macrophages (38), but not proximal renal tubular cells, (39) respond to a number of proinflammatory stimuli by increasing synthesis of 1α25(OH)2D3. Consistent with this, inhibition of intracellular p38 MAPK and NF-κB pathways that promote terminal differentiation of DCs (23) prevents the transcriptional up-regulation of 1αOHase. Definition of the minimal promoter region of the 1αOHase gene in DCs or macrophages has yet to be determined. Thus, while the 1αOHase gene has a putative NF-κB binding sequence in its promoter region (40), it is quite possible that the stimulatory effects of proinflammatory signals upon its expression are indirect rather than direct. The failure of inflammatory mediators to enhance proximal renal tubular cell expression of 1αOHase, may reflect a relative lack of the cellular apparatus required to respond to such signals. In this context it is noteworthy that differential expression of Toll-like receptors in cell lines isolated from the proximal and distal nephron has been implicated in the variable transcriptional up-regulation of 1αOHase in response to LPS (39). A second difference between renal and extrarenal 1αOHase regulation is that in contrast to proximal renal tubular cells (24), both DCs and macrophages (26) demonstrate relative insensitivity to the negative regulatory effects of 1α25(OH)2D3. Variations in the absolute levels of VDR between the cell types or the ability of VDR ligand to recruit coactivators/partner receptors may explain this difference. Indeed, it is noteworthy that IFN-γ-activated macrophages can recruit complexes containing 1α25(OH)2D3 into other pathways that directly antagonize the binding of VDR-ligand complexes to vitamin D3-responsive cis elements (41). Whether such mechanisms are operative in human DCs and interfere with VDR-mediated repression of 1αOHase expression remains to be determined. Third, whereas 25(OH)D3 complexed to vitamin D binding protein is taken up by receptor-mediated endocytosis in proximal renal tubular cells (42), this means of controlling access of enzyme to substrate does not appear to be operative in moDCs which do not express the transcript for megalin, one of the key endocytic receptors involved in this process (M. Hewison and R. Chakraverty, unpublished observations). Despite this, during only relatively short incubations in vitro, 5–10% of the input 25(OH)D3 is converted to 1α25(OH)2D3, suggesting that that the process whereby vitamin D3 metabolites enter DCs is highly efficient. Whether this is entirely dependent upon passive diffusion (as with other steroid hormones) or is regulated/facilitated by other means is not known.
Both macrophages (26) and DCs (the latter at the level of transcription) demonstrate the capacity to express the enzyme 24(OH)ase which inactivates 1α25(OH)2D3. The role this enzyme plays in limiting the activity of 1α25(OH)2D3 is unclear at present, because 1α25(OH)2D3 (added directly to the culture or generated by differentiating DCs from the precursor molecule 25(OH)D3) is clearly able to inhibit DC development. Indeed, the inhibitory effects of 1α25(OH)2D3 was at its greatest during the early culture period, during which the capacity of DCs to up-regulate 24(OH)ase transcript levels in response to ligand is at its highest. Although we have not formally evaluated DC 24(OH)ase enzyme activity, this finding suggests that inactivation of 1α25(OH)2D3 by 24(OH)ase is not significant under the specific conditions studied here. One possibility is that 24(OH)ase activity is more relevant to lower concentrations of 1α25(OH)2D3 that may exist, for example, at the peripheries of focal areas of local synthesis.
The capacity of cells to respond to 1α25(OH)2D3 will be related to their expression of VDR, the presence or absence of intracellular pathways that transduce VDR-mediated responses or, indeed, their ability to invoke other nongenomic signaling pathways (43). In common with macrophages (38), differentiation of moDCs is associated with down-regulation of the VDR. This may explain in part the relative resistance of immature DCs as compared with monocyte precursors to the inhibitory effects of the ligand. Although immature DCs still retain some sensitivity to 1α25(OH)2D3 (5, 6, 44), mature DCs are almost completely resistant to its inhibitory effects (44) despite similar expression of VDR. Thus, other factors, including expression of partner receptors/coactivators or the integrity of downstream signaling pathways, are likely to contribute to this resistance. The precise factors that regulate VDR cellular levels are not yet defined in detail, although the ligand appears to have an important role in increasing its expression via a posttranscriptional mechanism in nonhemopoietic cells (45). In this regard, tissue-specific major N-terminal differences in human VDR mRNAs have been described that could conceivably contribute to variations in renal cell and myeloid hemopoietic cell responsiveness to ligand (46). We did not identify any significant changes in moDC VDR expression in response to exogenous 1α25(OH)2D3 (data not shown), although we did not determine VDR activity in this particular context. Although data have yet to establish a link between any particular VDR gene polymorphisms and absolute levels of VDR mRNA expression (47) or protein synthesis (48) this potential relationship has not been explored in hemopoietic cells. The findings in this study that VDR is down-regulated as DCs differentiate from monocytes and that transcription and activity fall in parallel provide a potentially useful model system in which regulation of VDR expression could be dissected in detail.
In summary, we have demonstrated that differentiation of human moDCs is associated with the active synthesis of 1α25(OH)2D3. In contrast, differentiation of moDCs is associated with significant reductions in the expression of VDR as compared with monocyte precursors. Thus, generation of 1α25(OH)2D3 appears to exert an autoregulatory function by partially prohibiting the differentiation of monocyte precursors into immature DCs and their subsequent ability to undergo terminal differentiation in response to maturation stimuli. These findings suggest a model in which DC synthesis of 1α25(OH)2D3 acts to limit unwanted immune reactivity. In this regard, we are currently investigating the regulation of 1α25(OH)2D3 synthesis and VDR expression in other human DC subsets. How this phenomenon impacts upon the induction of adaptive immunity in vivo will require the detailed analysis of how VDR or 1αOHase knockout mice (49) generate immune responses to infectious agents, model, or transplanted Ags.
We gratefully acknowledge Moray Campbell and Jasbir Moore (Medical Sciences, Birmingham, U.K.) for advice with real-time RT-PCR analyses.
↵1 R.C. was supported by a Senior Fellowship in Experimental Hematology from the Leukemia Research Fund U.K. A.E. was supported by a Medical Research Council Career Development Award and by Cancer Research U.K. Grant SP2584. This work was also supported by grants from the National Kidney Research Fund U. K. (to R.B.) and Biotechnology and Biological Sciences Research Council Research Grant 6/S14523 (to M.H. and M.K.).
↵2 Address correspondence and reprint requests to Dr. Ronjon Chakraverty at the current address: Transplant Biology Research Center/Bone Marrow Transplantation Section, Massachusetts General Hospital, Harvard Medical School, Massachusetts General Hospital East, Building 149-5102, 13th Street, Boston, MA 02129. E-mail address:
↵3 Abbreviations used in this paper: DC, dendritic cell; 1α25(OH)2D3, 1α25-dihydroxyvitamin D3; moDC, monocyte-derived DC; VDR, vitamin D receptor; 25(OH)D3, 25-hydroxyvitamin D3; 1αOHase, 25(OH)D3 1α-hydroxylase; 24(OH)ase, 25(OH)D3 24-hydroxylase; TLCK, N-α-tosyl-l-phenylalanine chloromethyl ketone; CD40L, CD40 ligand; MAPK, mitogen-activated protein kinase; Ct, cycle threshold; PI, propidium iodide.
- Received October 11, 2002.
- Accepted March 20, 2003.
- Copyright © 2003 by The American Association of Immunologists