Expression of 11β-hydroxysteroid dehydrogenase type 1 permits regulation of glucocorticoid bioavailability by human dendritic cells

Synthetic derivatives of naturally occurring glucocorticoids (GCs) have major anti-inflammatory and immunosuppressive properties that make them suitable for the treatment of autoimmune disease or the prevention of allograft rejection.¹  Rational design of novel synthetic GC derivatives or of strategies that enable their optimal use requires an improved understanding of how exogenous and endogenous GCs exert their effects upon the immune system. Studies in vitro which use GCs at supraphysiologic concentrations suggest that major cellular targets are T cells^2,3  and dendritic cells (DCs).^4-9  In the latter case, GCs inhibit in vitro differentiation of DCs from their precursor cells^4-6,9  and impair their capacity to undergo terminal differentiation or generate proinflammatory cytokines.^5-9  The role of endogenous GCs in modulating DC development or function however is unknown.

Crucial to any concept that seeks to explain how endogenous GCs influence DC development is an appreciation of how GC levels are regulated in vivo. Endogenous GCs exist either as active 11-hydroxy ligands (eg, cortisol) that bind glucocorticoid receptor α (GRα) or inactive 11-keto derivatives which do not (eg, cortisone).^10,11  Fine-tuning of active GC bioavailability in peripheral tissues is regulated by prereceptor interconversion of inactive and active ligands mediated by 2 isozymes of 11β-hydroxysteroid dehydrogenase (11β-HSD) that catalyze the so-called “cortisol-cortisone” shuttle.^10,11  Inactivation of cortisol by dehydrogenation to cortisone is mediated by 11β-HSD2. In the kidney, this activity prevents competition between GCs and mineralocorticoids for binding to nonselective mineralocorticoid receptors.¹²  In the placenta, inactivation of cortisol by 11β-HSD2 protects the fetus from the effects of maternal GCs.¹³  The other enzyme in this group, 11β-HSD1 has a more ubiquitous distribution, including liver, gut, and connective and lymphoid tissues.^11,14  In contrast to the actions of the type 2 isozyme, 11β-HSD1 acts primarily to amplify the effect of GCs by reducing inactive cortisone to cortisol.¹⁰  This cortisone reductase activity is functionally coupled to the provision of reduced nicotinamide adenine dinucleotide phosphate (NADPH) by hexose-6-phosphate dehydrogenase (H6PDH), an enzyme that colocalizes with 11β-HSD1 on the endolumenal surface of the endoplasmic reticulum (ER).^15-17 

We and others have previously reported that human DCs are involved in prereceptor amplification of another immunosuppressive steroid, vitamin D₃, by 1α-hydroxylation of inactive 25-hydroxyvitamin D₃ to active 1α,25-dihydoxyvitamin D₃.^18,19  This prompted us to examine whether human DCs can also participate in prereceptor regulation of GC bioavailability and, if so, under what conditions. We report here that differentiation of immature DCs from monocyte precursors is associated with dramatic induction of 11β-HSD1 reductase activity that permits conversion of inactive cortisone to active cortisol. Synthesis of cortisol from cortisone added at physiologic concentrations leads to autocrine-negative regulation of DC differentiation from monocyte precursors. Prereceptor amplification of glucocorticoid bioavailability is maintained or increased by signals associated with innate immune activation but is lost following CD40 ligation, a signal associated with cognate antigen recognition.

Human monocyte-derived DCs were prepared as described previously.¹⁹  Briefly, adherent mononuclear cells were cultured in Iscoves Modified Dulbecco medium (Gibco-BRL, Rockville, MD) plus 5% human AB serum (HD Supplies, Buckinghamshire, United Kingdom) plus 2 mM l-glutamine plus 100 μg/mL streptomycin plus 100 μg/mL penicillin for 7 days in the presence of 800 U/mL granulocyte-macrophage colony-stimulating factor (GM-CSF; PeproTech, London, United Kingdom) and 1000 U/mL interleukin 4 (IL-4; R & D Systems, Abingdon, United Kingdom) in 75-cm² flasks. In certain experiments DCs were propagated in macrophage serum-free media (Gibco-BRL). Cortisol or cortisone [10^-9 M to 10^-6 M] or vehicle (Sigma, St Louis, MO) were added in some experiments either at the beginning of culture or at time points indicated in the text. Day 7 immature DCs were matured for 48 hours in the presence of one of the following maturation stimuli: G28.5 CD40 antibody 5 μg/mL (Professor J. Gordon, Department of Immunology, University of Birmingham) plus goat anti–mouse immunoglobulin crystallizable fragment (Fc) cross-linker 5 μg/mL (Pierce, Rockford, IL), tumor necrosis factor α (TNF-α) 50 ng/mL (PeproTech), lipopolysaccharide (LPS) Escherichia coli 1 μg/mL (Sigma), polyinosinic polycytidylic acid/polydeoxyinosinic polydeoxycytidylic acid (poly I:C/poly dI:dC) 25 μg/mL (Amersham, Piscataway, NJ), Saccharomyces cerevesiae zymosan 10 μg/mL (Sigma), Pam₃CysSerLys₄ (Pam₃CSK₄) 100 ng/mL (InvivoGen, San Diego, CA), and Salmonella typhimurium flagellin 100 ng/mL (InvivoGen). In controls, mouse immunoglobulin G1 (IgG1) MOPC21 5 μg/mL (Sigma) plus goat anti–mouse immunoglobulin Fc cross-linker 5 μg/mL (Sigma), or phosphate-buffered saline were substituted for the various maturation stimuli. To test the effects of inhibiting 11β-HSD activity, we added 10^-5 M 18β-glycyrrhetinic acid (Sigma) in some experiments.

Cell-surface phenotype was analyzed using the following monoclonal antibodies: anti–CD1a–fluorescein isothiocyanate (FITC) (HI149, IgG1), anti–CD14-phycoerythrin (PE) (M5E2, IgG2a), anti–CD83-PE (HB15E, IgG1), anti–CD86-PE (IT2.2, IgG2b), anti–HLA-DR-Cychrome (G46.6, IgG2a), and the isotype controls labeled with the appropriate fluorochromes (MOPC21, IgG1; G155-178, IgG2a; and 27-35, IgG2b), which were obtained from BD Pharmingen (San Diego, CA). Staining of the samples was measured on a fluorescence-activated cell sorting (FACS) Coulter Epics XL (Beckman Coulter, Fullerton, CA), using System II software (Beckman Coulter).

Cytospins of monocytes and immature DCs were stained for 1 hour at room temperature with sheep anti–11β-HSD1 antibody diluted to 1/50 (The Binding Site, Birmingham, United Kingdom). Detection was performed using secondary FITC-conjugated rabbit antisheep antibody (Abcam, Cambridge, United Kingdom). Specificity of staining was confirmed by blocking experiments in which antibody was preincubated with immunizing peptide. Staining was examined using an Axiovert 100 confocal microscope (Zeiss, Hertfordshire United Kingdom) and × 63 C-Apochromat objective (numerical aperture 0.55). Images were acquired using the LSM 4 software (Zeiss). To detect protein expression by Western blot, cells were lysed and sonicated in 9 M urea buffer, 50 nM Tris(tris(hydroxymethyl) aminomethane)–HCl (pH 7.3) and 0.15 M β-mercaptoethanol, and then extracts were fractionated by 12.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Detection was performed using anti–11β-HSD1 primary antibody diluted 1:200 in 5% nonfat milk overnight at 4°C. Secondary detection was with peroxidase-conjugated goat anti–sheep IgG (Sigma) diluted 1:2000 for 2 hours at room temperature. In parallel, protein loading was evaluated using mouse anti–β-actin, followed by peroxidase-conjugated anti–mouse IgG (both from Sigma). Protein bands were visualized using an enhanced chemiluminescent detection system according to the manufacturer's instructions (Amersham).

Expression of mRNA for 11β-HSD1/2 was determined by quantitative reverse transcriptase–polymerase chain reaction (RT-PCR). Primers and probes, together with the methods for RNA extraction, cDNA amplification, and reaction conditions, have been described previously.^17,20 

Cells were washed and then resuspended in 500 μL RPMI 1640 without serum and incubated with either 10^-7 M cortisone or 10^-7 M cortisol supplemented with tracer [³H]-cortisone or [³H]-cortisol, respectively, for 1 hour at 37°C. The reaction was terminated by freezing at -20°C. Steroids present in the samples were extracted in dichloromethane and separated using thin-layer chromatography in a solvent tank made up of a 12:1 ratio of chloroform to ethanol. The steroids produced were scanned on a Bioscan 3000 image analyzer (Lablogic, Sheffield, United Kingdom), and activity was expressed as picomoles produced per hour per 10⁶ cells. In separate experiments, DC cultures were incubated with 10^-6 M cortisone or cortisol in serum-free media, and supernatant cortisol levels were measured by a competitive binding assay kit according to manufacturer's instructions (Immunodiagnostic Systems, Newcastle, United Kingdom).

Activity assays were carried out on whole-cell extracts of monocytes and DCs by freezing and thawing the cells in homogenization buffer, followed by disruption of the cell membranes using a Glycerol/Triton X-100 mix (1:1 ratio vol/vol). H6PDH activity was analyzed in 1 mL total volume of glycine buffer with 200 μg protein, galactose-6-phosphate (2 mM), and cofactor (0.5 mM NADP⁺). Relative H6PDH activities were measured by absorbance at 340 nm.

DCs were incubated for 48 hours 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 levels were evaluated using a sandwich enzyme-linked immunosorbent assay (ELISA) kit (BD Pharmingen, San Diego, CA) according to the manufacturer's instructions.

Statistical analysis was performed using the Student t test (Instat version 2.04a computer program; GraphPad Software, San Diego, CA).

To determine the expression of 11β-HSD1 and 11β-HSD2 mRNA in human DCs, we incubated monocytes in the presence of IL-4 and GM-CSF for 7 days, prepared RNA, and performed real-time RT-PCR. Differentiation of monocyte precursors to DCs was accompanied by a marked increase in 11β-HSD1 mRNA expression, increasing by a median of 1502-fold (range, 419- to 6790-fold, n = 10) in immature DCs (Figure 1A). In contrast, 11β-HSD2 mRNA expression remained very low throughout DC differentiation with levels less than 0.1% of its type 1 counterpart in immature DCs. Immunofluorescent staining for the 11β-HSD1 enzyme demonstrated a similar increase in protein expression with dense cytoplasmic staining in immature DCs (Figure 1B).

As part of these initial studies, we also evaluated GR expression in differentiating DCs, because GR levels are an important determinant of GC-induced signaling.²¹  We found that GRα transcription (Figure 1C) and intracellular staining (data not shown) were readily detectable and equivalent in monocytes and immature DCs. We also examined expression of GRβ (a splice variant receptor of GR), which can potentially modulate GC signaling by antagonizing the activity of the GRα-ligand complex.²²  GRβ mRNA levels, however, remained constant and very low throughout DC differentiation at 3 to 4 logs lower than that of GRα (Figure 1C).

Next, we evaluated the ability of immature DCs to interconvert cortisone and cortisol using thin-layer chromatography of products generated by intact cells incubated in the presence of 10^-7 M [³H]-cortisol or [³H]-cortisone. Consistent with their lack of 11β-HSD1 expression, monocytes lacked both reductase and dehydrogenase activity (Figure 2A). In contrast, immature DCs demonstrated marked enhancement of reductase activity as manifested by their capacity to generate active cortisol from inactive cortisone (Figure 2A). A small increase in dehydrogenase activity was also observed (4:1 ratio of reductase to dehydrogenase activity) in immature DCs. Reductase activity was observed by day 2 of culture, indicating that this function is acquired early as DCs differentiate from monocyte precursors (Figure 2B). Similar up-regulation of 11β-HSD1 mRNA expression and reductase activity was observed in DCs cultured in serum-containing and serum-free conditions (data not shown).

Using a competitive binding assay, we detected a progressive accumulation of cortisol in culture supernatants as detected by competitive binding assay (Figure 2C). This accumulation was significantly inhibited when differentiating DCs were cotreated with 18β-glycyrrhetinic acid (GA), an inhibitor of 11β-HSD activity. By contrast, in cultures that were supplemented with cortisol rather than cortisone, GA had no effect on the levels as measured at the end of culture on day 7, confirming no net conversion of cortisol to cortisone. No accumulation of cortisol was observed in media alone controls, indicating no de novo synthesis of cortisol from cholesterol.

Although 11β-HSD1 is a bidirectional enzyme in whole-cell preparations, reductase activity is lost in cell extracts,¹⁰  indicating that the specific intracellular location of the enzyme is important. The catalytic domain of 11β-HSD type 1 faces into the lumen of the ER where it colocalizes with and demonstrates functional coupling to H6PDH, which generates the NADPH that drives its reductase activity.^15-17  In concert with the dramatic increases in 11β-HSD1 activity, we also observed increases in H6PDH mRNA expression (15.6-fold; range, 3.6- to 50.6-fold; n = 7) and activity (Figure 2D) upon differentiation of monocytes to immature DCs.

Taken together, these data indicate that differentiation of immature DCs is associated with the acquisition of both the enzymatic apparatus and NADPH cofactor provision required for prereceptor amplification of GC bioavailability.

According to several reports, differentiation of human DCs from monocyte precursors is exquisitely sensitive to inhibition by dexamethasone added at concentrations that are achievable therapeutically.^5,6,9  These observations cannot be extrapolated directly to normal physiologic effects of circulating GCs for 2 reasons. First, unlike endogenous GCs, dexamethasone is not metabolized by the cortisol-cortisone shuttle.²³  Second, the dexamethasone concentrations used in these studies were in excess of endogenous GC concentrations usually encountered physiologically.¹⁰  Our finding that human monocyte-derived DCs participate in the endogenous synthesis of active GCs allowed us to examine the cellular response to naturally occurring GCs at their physiologic concentrations and under conditions in which they can be metabolized.

The apparent K_m (Michaelis constant) for dehydrogenase and K_m for reductase activity of 11β-HSD1 are 3 × 10^-6 M and 2.5 × 10^-7 M, respectively. As shown in Figure 2A, 11β-HSD1 reductase activity predominates within intact DCs when exposed to physiologically relevant concentrations of glucocorticoids (10^-7 M). Although cortisone possesses negligible affinity for the GRα,¹⁰  we reasoned that DC-mediated conversion of cortisone to cortisol could be associated with autocrine- and/or paracrine-negative regulation of DC differentiation and function. To test this hypothesis we exposed CD14⁺CD1a^- monocyte precursors from the first day of culture to 10^-7 M cortisol or cortisone, representative of concentrations at the upper end of their normal physiologic ranges and evaluated the effect upon immature DC differentiation. By day 7, immature DCs cultured in the presence of GM-CSF and IL-4 alone had lost surface CD14 expression and had acquired surface CD1a expression (Figure 3B). In contrast, DCs differentiated in the presence of 10^-7 M cortisol were also CD14^- but showed a marked reduction in the acquisition of CD1a expression. Significantly, addition of 10^-7 M cortisone at the initiation of culture also resulted in marked inhibition of CD1a expression, and this effect was relieved by the addition of GA. Addition of GA did not abrogate the inhibitory effects of cortisol. The inhibitory effects of both cortisol and cortisone were dose dependent with lesser effects observed at 10^-8 M and 10^-9 M concentrations (not shown). In contrast to studies using supraphysiologic concentrations of GCs,^5,9  we did not observe retention of CD14 surface expression in cells cultured in the presence of either cortisol or cortisone at the concentrations tested. No effect of cortisol or cortisone upon GRα or GRβ expression was detected in parallel experiments (data not shown). These data support the hypothesis that autocrine generation of cortisol mediated by 11β-HSD1 has a negative regulatory effect upon DC differentiation from precursor cells. Furthermore, this process operates at physiologically relevant GC concentrations.

We next considered how cells cultured under the above conditions would respond to activating stimuli that induce terminal differentiation and enhance immunostimulatory potential. Thus, monocytes were cultured in the presence of IL-4/GM-CSF ± 10^-7 M cortisol for 7 days and then exposed to TNF-α for 48 hours. Cells that had been exposed to cortisol failed to up-regulate the DC-specific maturation marker CD83 to the same degree as control DCs (Figure 4A). Furthermore, cortisol-treated cells demonstrated a marked failure to develop immunostimulatory potential in response to maturation stimuli as judged by reduced allostimulatory activity (Figure 4B) and reduced IL-12p70 production (Figure 4C). The addition of GA did not reverse cortisol-mediated inhibition. Cells differentiated in the presence of 10^-7 M cortisone from the initiation of culture also failed to undergo full terminal differentiation or to develop significant immunostimulatory potential (Figure 4A-C). However, in this case, cotreatment of cells with GA partially reversed the inhibitory effects observed following cortisone treatment alone. Thus, data in Figures 3 and 4 indicate that 11β-HSD1–dependent prereceptor amplification of active GCs by DCs is associated with autocrine- and paracrine-negative regulation of monocyte differentiation into immature DCs. Differentiating cells subjected to this negative feedback are rendered unable to develop their full immunostimulatory potential following subsequent exposure to maturation stimuli.

In the above studies, we evaluated 11β-HSD1 activity under conditions associated with differentiation of monocyte precursors to immature DCs. Next, we evaluated how 11β-HSD1 activity was influenced by activation signals associated with terminal differentiation of DCs. In other cell types, exposure to proinflammatory mediators enhances the expression and activity of 11β-HSD1.^24,25  To examine whether this was also the case for human DCs, we exposed immature DCs to signals associated with innate or adaptive immune activation and then measured 11β-HSD1 activity by thin-layer chromatography of products following brief incubation with [³H]-cortisol or [³H]-cortisone. To model the effect of innate-activating signals, immature DCs were activated for 48 hours in the presence of the proinflammatory cytokine, TNF-α (Figure 5A) or through direct recognition of pathogen-associated molecular patterns (PAMPs) that stimulate Toll-like receptors (TLRs)²⁶  (Table 1). Activation through TNF-α maintained but did not increase 11β-HSD1 reductase activity (Figure 5A). Similarly, poly I:C (a TLR3 agonist) and E coli LPS (a TLR4 agonist) had little effect upon 11β-HSD1 reductase activity. In contrast, significant increases in reductase activity were observed following exposure to the yeast wall derivative, zymosan (signals through TLR2/6) and S typhinurum flagellin (a TLR5 agonist). There was also a trend for increased reductase activity following exposure to Pam₃CSK₄, a synthetic tripamitoylated lipopeptide that mimics the acylated amino terminus of bacterial lipoproteins (signals through TLR1/2). Thus, DC maturation induced by signals associated with innate immune activation is associated with maintenance of or increases in 11β-HSD1 reductase activity.

We next considered the effect of signals associated with adaptive immunity upon DC 11β-HSD1 activity. Cognate interactions between naive T cells and DCs lead to up-regulation of CD40L by T cells and activation of the DCs via engagement of the CD40 receptor.²⁷  Strikingly, we found that CD40 ligation induced by agonistic anti-CD40 antibody and anti-Fc cross-linker was associated with a marked 74.1% ± 4.3% decrease in 11β-HSD1 cortisone reductase activity compared with control antibody plus anti-Fc cross-linker (Figure 5A). We observed similar reductions in reductase activity following exposure of DCs to mouse fibroblasts transfected with CD40L (54.6% ± 18.4% decrease compared with reductase activity following coincubation with nontransfected controls, n = 3). This shift occurred in the absence of any changes in DC viability (not shown), significant down-regulation of 11β-HSD1 transcription (Figure 5B), or protein expression (Figure 5C). Using densitometry to correct for protein loading, the ratio for expression of 11β-HSD1 following CD40 ligation versus control was 0.93 ± 0.11 (n = 3). Decreased 11β-HSD1 reductase activity following CD40 ligation was not associated with any increase in 11β-HSD1 type 2 expression as indicated by a failure to up-regulate 11β-HSD1 type 2 mRNA expression or cortisol dehydrogenase activity (Figure 5A). Thus, although the mechanism remains to be determined, CD40 ligation sharply down-regulates the capacity of monocyte-derived DCs to generate cortisol.

Previous reports have demonstrated that susceptibility to GC-mediated inhibition is evident both during the early phase of DC culture⁹  and during the maturation phase.^5,7  To identify at which points during differentiation DCs are susceptible to autocrine-negative regulation mediated by local increases in GC bioavailability, we exposed cells to 10^-7 M cortisone ± GA from days 1 to 3, 4 to 6, or 7 to 9 during DC culture. From day 7, DCs were matured with TNF-α. As shown in Figures 2B and 5A, significant 11β-HSD1 type 1 reductase activity is evident from day 2 of culture, peaks between days 3 and 7, and is maintained during TNF-α–mediated maturation. On day 9 cells were washed extensively, and their immunostimulatory potential evaluated in an allogeneic-mixed leukocyte reaction. Cortisone inhibited the allostimulatory potential of DCs when added at each of these time points, whereas cotreatment with GA partially relieved the effect (Figure 6A). Thus, autocrine inhibition of DC differentiation and function through increased GC bioavailability potentially operates throughout DC differentiation.

Terminal differentiation of DCs is associated with resistance to the effects of dexamethasone.^4,6  We hypothesized therefore that prereceptor amplification of GC by mature DCs would not be associated with autocrine inhibitory effects. To test this, we exposed TNF-α–matured DCs to cortisone or vehicle for 48 hours and then examined the cells by flow cytometry. In contrast to the GA-reversible effects observed when cortisone is added at the initiation of culture (Figure 6B), addition of cortisone to TNF-α–matured DCs cells had little effect upon CD83, CD86, or HLA-DR expression (Figure 6C). A similar lack of effect was observed following addition of cortisol (data not shown). TNF-α–matured DCs treated as in Figure 6C with cortisone or cortisol, washed extensively, and then coincubated with allogeneic CD4⁺ T cells in the absence of GC retained similar immunostimulatory potential as compared with vehicle-treated control mature DCs (not shown). Thus, although TNF-α–matured DCs are proficient at amplifying local cortisol concentrations, they are themselves resistant to its effects.

We have demonstrated for the first time that immature DCs are proficient in the prereceptor amplification of GC bioavailability. Generation of active GCs acts as an autocrine-negative regulator of DC differentiation from monocyte precursors, and, significantly, this checkpoint operates at physiologic concentrations of circulating GCs. The enhancement of GC bioavailability by mature DCs depends upon the context of their maturation. DCs matured in the presence of signals associated with innate immune activation maintain or increase cortisone reductase activity. In sharp contrast, DCs undergoing terminal differentiation in response to CD40 ligation, a key signal associated with adaptive immune activation, demonstrate a marked reduction in the capacity to generate active cortisol.

Serum concentrations of non–protein-bound cortisol vary significantly from 10^-9 M at the diurnal nadir to 10^-7 M at the diurnal peak.¹⁰  Cortisone, is largely unbound to protein and has somewhat less variable levels, which at 10^-8 to 10^-7 M are often in excess of endogenous cortisol.¹⁰  The levels of GCs within immune microenvironments, however, may not always fully reflect those found in serum. For example, high GC concentrations are detected in gut-associated lymphoid tissue or in the lung following infection.^14,28  This implies that immune cells entering or exiting sites of inflammation traffic across gradients of widely differing GC concentrations. In such contexts, fine-tuning of GR signaling may be important in regulating the cellular response to endogenous GCs. Our data suggest that this regulation takes place in DCs at the prereceptor level via 11β-HSD1 activity.

Significantly, we found that 11β-HSD1 reductase activity is already set at high levels in immature DCs. High levels of active GC generation in the steady state may be particularly important in cells that possess such high immunostimulatory potential in order to raise the threshold for DC-induced immune activation. Our data suggest that autocrine-negative regulation of DC differentiation from precursor cells may be one means by which this threshold is raised. Several other pathways that have also been implicated in autocrine-negative regulation of DC differentiation and function, including the synthesis or release of thrombospondin,²⁹  1α,25-dihydoxyvitamin D_3,¹⁹  prostaglandin E₂ (PGE₂),³⁰  or IL-10.³¹  In some cases, negative regulatory mediators act cooperatively by enhancing the generation of other mediators (eg, the enhancement of thrombospondin generation by IL-10 or PGE₂²⁹ ). We have not, however, observed any effect of exogenous 1α,25-dihydoxyvitamin D₃ or PGE₂ upon the cortisone reductase activity of immature DCs (L.F. and R.C., unpublished data, September 2004). Although to date we detected no evidence of cross-regulation in this system, active GCs generated by DCs may have additive or synergistic effects with other immunosuppressive mediators. For example, active GCs demonstrate synergy with 1α,25-dihydoxyvitamin D₃ in inhibiting DC differentiation³²  or in the generation of T cells with an IL-10–producing regulatory phenotype.³³ 

This multifaceted restraint of DC differentiation and function is presumably released under inflammatory conditions associated with infection or allograft rejection. We found that activation of DCs via innate stimuli was associated with either maintenance or increases in the levels of 11β-HSD1 reductase activity. These data are consistent with the concept that DCs activated via stimuli associated with innate immunity retain their capacity to amplify GC bioavailability. In the initial phases of an inflammatory response, it is possible that such activity might be required to prevent excessive recruitment of DCs from precursor pools or to limit through paracrine effects, the recruitment or activation of other immune cells. In contrast to DC stimulation via TLR3 or TLR4, PAMP recognition through TLR2 and TLR5 was associated with increases in 11β-HSD1 reductase activity. This differential effect of the various TLR agonists is of interest because DCs activated via TLR2 or TLR5 may preferentially drive T helper 2 (Th2) responses,^34-36  whereas DCs activated through TLR3 or TLR4 bias toward Th1 polarization.^37,38  For example, in contrast to human DCs activated via TLR3 or TLR4, DCs activated via TLR2 are unable to produce IL-12 that instructs Th1 differentiation.^34,39  Although any one pathogen is likely to activate more than one TLR family member in vivo,²⁶  it will be of interest to determine the extent to which TLR-induced changes in DC regulation of GC bioavailability have the potential to influence T-cell differentiation.

Although innate signals may be required for complete “licensing” of DCs,⁴⁰  the coupling of cognate recognition to CD40 cross-linking is thought to be an important checkpoint that prevents unwanted immune reactivity.²⁷  In contrast to maturation induced by innate signals, terminal differentiation induced by ligation of the CD40 receptor blocks DC generation of cortisol. This reduction in activity takes place in the absence of any significant change in enzyme expression. Furthermore, H6PDH activity remained equivalent to that observed in immature DCs (L.F. and R.C., unpublished data, January 2004), suggesting that the fall in reductase activity is not consequent upon a failure to generate reducing equivalents within the ER. Thus, although the mechanism underlying CD40-mediated decreases in 11β-HSD1 reductase activity remains unclear, it is possible that the resulting reduced local concentrations of active GCs lower the threshold for DC-induced immune activation. For example, on the basis of known effects of GCs, such reductions may lower the threshold for T-cell activation,²  or generate conditions that are permissive for Th1 effector differentiation.³  Precise testing of these concepts will require in vivo model systems in which the effect of DC 11β-HSD1 activity upon the induction of adaptive T-cell responses can be tested.

In agreement with previous reports,^4,6  we have found that terminal differentiation of DCs is associated with resistance to the inhibitory effects of GCs. This finding has 2 implications regarding the functions of 11β-HSD1 activity in DCs. First, it suggests that autocrine-negative regulation of DC development or function acts primarily by blocking differentiation from precursor cells or immature DCs but has lesser effects upon mature DCs. Second, it implies that paracrine effects predominate in situations in which mature DCs retain 11β-HSD1 reductase activity, as for example when they are activated by innate stimuli. The resistance of mature DCs to the effects of GCs are akin to the resistance of mature DCs to other immunosuppressive mediators.^19,41  In the case of autocrine-negative regulation mediated by IL-10 or 1α,25-dihydroxyvitamin D₃, this resistance may be mediated in part by down-regulation of the respective surface or nuclear receptors.^19,31  However, a similar mechanism does not appear to operate in the case of autocrine regulation mediated by active GCs since total GR expression was constant throughout differentiation (Figure 1C; our unpublished data). Furthermore, we have detected no evidence of increased expression of either GRβ or the mineralocorticoid receptor (both of which may act as receptor “decoys”^22,42 ) in mature DCs. These findings suggest that other GR-independent factors such as postreceptor accessory pathways (eg, invoked by nuclear factor-κB activation) contribute to GC resistance.²¹ 

Our data are consistent with a model in which prereceptor amplification of GC bioavailability by immature DCs imposes restraint upon the differentiation of monocyte precursors to DCs. In response to innate immune activation, DCs continue to generate cortisol, but this capacity is lost upon ligation of the CD40 receptor as might occur following cognate DC–T-cell interactions. Subsequent falls in local GC concentrations may, thus, generate conditions that are permissive for specific T-cell activation. Detailed examination of the integrity of the 11β-HSD1 pathway in tissues or lymphoid organs at differing phases of the immune response may permit the design of therapies to target DCs with enzyme substrates or inhibitors with the respective aims of inhibiting or promoting immunity.

We thank Jean Villard, Andrew Tager, Thomas Fehr, and Paul Stewart for their careful review of the manuscript. We also thank Garry Sedgewick for technical assistance and all members of the laboratory of Professor Paul Moss for their valuable input.