Leukocytes are rapidly recruited to sites of inflammation via interactions with the vascular endothelium. The steroid hormone dehydroepiandrosterone (DHEA) exerts anti-inflammatory properties; however, the underlying mechanisms are poorly understood. In this study, we show that an anti-inflammatory mechanism of DHEA involves the regulation of developmental endothelial locus 1 (DEL-1) expression. DEL-1 is a secreted homeostatic factor that inhibits β2-integrin–dependent leukocyte adhesion, and the subsequent leukocyte recruitment and its expression is downregulated upon inflammation. Similarly, DHEA inhibited leukocyte adhesion to the endothelium in venules of the inflamed mouse cremaster muscle. Importantly, in a model of lung inflammation, DHEA limited neutrophil recruitment in a DEL-1–dependent manner. Mechanistically, DHEA counteracted the inhibitory effect of inflammation on DEL-1 expression. Indeed, whereas TNF reduced DEL-1 expression and secretion in endothelial cells by diminishing C/EBPβ binding to the DEL-1 gene promoter, DHEA counteracted the inhibitory effect of TNF via activation of tropomyosin receptor kinase A (TRKA) and downstream PI3K/AKT signaling that restored C/EBPβ binding to the DEL-1 promoter. In conclusion, DHEA restrains neutrophil recruitment by reversing inflammation-induced downregulation of DEL-1 expression. Therefore, the anti-inflammatory DHEA/DEL-1 axis could be harnessed therapeutically in the context of inflammatory diseases.
Activation of the endothelium is integral to leukocyte recruitment into inflamed tissues (1, 2). Upon activation by proinflammatory cytokines, such as TNF, endothelial cells orchestrate inflammation and leukocyte recruitment, which is mediated by a cascade of leukocyte–endothelial adhesive interactions (2–4). This cascade is initiated by selectin-mediated rolling and deceleration of leukocytes on the endothelial surface. Rolling triggers integrin activation, and activated integrins (primarily of the β2 family) promote firm adhesion of leukocytes to the activated endothelium, a prerequisite step for the subsequent leukocyte extravasation (5, 6).
Developmental endothelial locus 1 (DEL-1; also designated EGF-like repeats and discoidin domains 3 [EDIL3]) is a glycoprotein secreted by endothelial and other cells and has anti-inflammatory properties (7–16). DEL-1 interferes with β2-integrin–dependent adhesion of leukocytes to endothelial ICAM-1, thereby restraining leukocyte recruitment (8, 9). Consistently, genetic deletion of DEL-1 causes elevated leukocyte infiltration under different inflammatory conditions in mice (8, 9, 12, 15, 17–20). Inflammatory cytokines, such as IL-17 and TNF, inhibit endothelial DEL-1 expression, thereby facilitating leukocyte recruitment and inflammation (9, 17, 21). The IL-17–dependent downregulation of DEL-1 expression is reversed by D-series resolvins (RvDs) (21). However, little is known about other factors regulating DEL-1 expression.
Dehydroepiandrosterone (DHEA; 5-androsten-3β-hydroxy-17-one) and its sulfate ester are abundant circulating steroid hormones in human adults, whereas their concentration declines with age and in inflammatory diseases, such as arthritis and systemic lupus erythematosus (22–26). In humans, DHEA is produced in the adrenal cortex, the gonads, and the CNS (27–30). In tissues, DHEA displays anti-inflammatory properties, including inhibition of leukocyte recruitment (31, 32). DHEA can bind to nuclear receptors, such as estrogen receptor α and β (33, 34). Moreover, it was shown to bind to G protein–coupled receptors in endothelial and neuronal cells (35, 36). Additionally, it binds and activates the nerve growth factor (NGF) receptor, tropomyosin-related kinase A (TRKA), in neuronal and microglial cells, thereby triggering downstream AKT signaling (30, 37, 38). However, its exact mechanisms of action, especially in the context of recruitment regulation, remain largely unknown (39, 40).
In the current study, we demonstrate that DHEA mitigates leukocyte adhesion efficiency in the LPS-induced cremaster muscle inflammation model and reduces neutrophil recruitment in the LPS-induced lung inflammation model. Mechanistic studies revealed that DHEA counteracts the inhibitory effect of TNF on endothelial DEL-1 expression, suggesting that DEL-1 might mediate the antirecruitment effect of DHEA. Consistent with this notion, the anti-inflammatory effect of DHEA in the lung inflammation model is lost in DEL-1–deficient animals. Furthermore, we show that DHEA restores the TNF-mediated reduction in DEL-1 expression in endothelial cells by a mechanism that involves the TRKA receptor and PI3K/AKT signaling. These findings support an anti-inflammatory role of DHEA through restoration of endothelial DEL-1 expression under inflammatory conditions.
Materials and Methods
Intravital microscopy of the cremaster muscle
Eight to twelve-week-old male C57BL/6 mice (purchased from Janvier Labs, Le Genest-Saint-Isle, France) were injected i.p. with 2 mg DHEA (Sigma-Aldrich, Munich, Germany) diluted in PBS containing 4.5% ethanol and 1% BSA or the same amount of control vehicle diluent (4.5% ethanol, 1% BSA in PBS), as previously described (37). Thirty minutes after DHEA injection, 50 ng of LPS (Escherichia coli O111:B4; Sigma-Aldrich) were injected intrascrotally. Intravital microscopy was performed 3.5 h later. The cremaster muscle preparation was performed as previously described (41). Briefly, the scrotum of the mouse was incised, the cremaster muscle was exteriorized, additional tissue was removed, and the muscle was then opened through a longitudinal incision and mounted onto a self-customized stage. During intravital microscopy, the cremaster muscle was constantly superfused with warm superfusion buffer (41). Intravital microscopy was conducted on a BX51WI microscope (Olympus, Center Valley, PA) equipped with a 40× saline immersion objective (MplanFI/RI, 0.8 numerical aperture; Olympus) and a charge-coupled device camera (Kappa CF8 HS). VirtualDub (version 1.9.11) was used for recording of postcapillary venules. Leukocyte rolling was assessed as a percentage of rolling leukocytes relative to the number of leukocytes passing the vessel (rolling flux fraction), and leukocyte adhesion efficiency as was assessed as the number of adherent leukocytes per square millimeter relative to the WBC count per microliter. Adherent leukocytes were defined as nonmoving cells or cells with a displacement smaller than one cell diameter during 1 min of observation (42). Leukocyte rolling velocities were measured as averages over a 1-s time window using the Fiji software (43). Vessel diameters were measured using a digital image processing system (44). Centerline RBC velocity in cremaster muscle microvessels was obtained using a dual photodiode and a digital on-line cross-correlation program (CircuSoft Instrumentation, Hockessin, Germany). Wall shear rates were assessed as 4.9 (8 v/d), where v is the mean blood flow velocity and d is the diameter of the vessel (45). Blood cell numbers were determined through IDEXX ProCyte Dx Hematology Analyzer (IDEXX Europe B.V.). Experiments were approved by the Regierung of Oberbayern, Germany.
LPS-induced lung inflammation
LPS-induced lung inflammation was performed in DEL-1+/+ wild-type (WT) and DEL-1−/− mice. DEL-1−/− mice were previously described (8). Eight to twelve-week-old male mice were injected i.p. with DHEA (80 mg/kg) diluted in PBS containing 4.5% ethanol and 1% BSA or the same amount of control vehicle solution (4.5% ethanol, 1% BSA in PBS), followed by a second i.p. injection of the same dose of DHEA or vehicle solution after 24 h. One and a half hours later, anesthetized mice received LPS intranasally (0.2 mg/kg E. coli 0111:B4; InvivoGen, San Diego, CA). After 24 h, mice were anesthetized prior to bronchoalveolar lavage (BAL) collection and euthanasia. The total cell number in the BAL fluid was determined in a blinded manner by using a hemocytometer, and the percentage of recruited neutrophils was determined by flow cytometry analysis using Ly-6G– allophycocyanin (1A8; BD Pharmingen, Heidelberg, Germany) and CD11b–Alexa Fluor 488 (M1/7; BD Pharmingen) Abs. Experiments were approved by the Landesdirektion Sachsen, Dresden, Germany.
Cell lines, primary cells, and cell treatments
HEK-293T cells (American Type Culture Collection, Manassas, VA) were cultured in DMEM (Life Technologies) supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C and 5% CO2. HUVECs (Lonza, Basel, Switzerland) were cultured on 0.2% gelatin-coated plates in Endothelial Cell Growth Medium-2 (Lonza) at 37°C and 5% CO2.
For in vitro stimulations, HUVEC were prestarved in Endothelial Cell Growth Basal Medium-2 (EBM-2; Lonza) containing 1% FCS (Invitrogen, Carlsbad, CA) for up to 24 h, depending on the assay. DHEA was freshly dissolved in ethanol at 10 mM prior to each experiment. Cells were pretreated for 30 min with DHEA (100 nM or as indicated in the figure legends) or vehicle control (ethanol), followed by addition of LPS (100 ng/ml, E. coli 0111:B4; InvivoGen) or human rTNF (10 ng/ml; R&D Systems, Wiesbaden-Nordenstadt, Germany), as described in the figure legends. In certain experiments, the cells were pretreated with NGF (100 ng/ml; Merck Millipore, Darmstadt, Germany) in lieu of DHEA. In experiments investigating the receptor involved or downstream signaling molecules, prior to addition of DHEA, the cells were earlier treated with one of the following reagents; TRKA inhibitor 648450 (Merck Millipore), AKT inhibitor MK2206 (Cayman Chemical, Ann Arbor, MI), PI3K inhibitor LY294002 (Sigma-Aldrich) or its inactive analogue LY303511 (Sigma-Aldrich), or DMSO (as control), as described in the figure legends. In all experiments, sequential treatments were performed without intermediate washing steps; for instance, treatment with TNF was performed in the presence of aforementioned pretreatments. All treatments were performed in 1% FCS EBM-2.
Quantitative real-time PCR
Total RNA extraction, cDNA synthesis and quantitative real-time PCR (qPCR) were performed as previously described (21, 46–48). Total RNA was extracted using the NucleoSpin RNA isolation kit (Macherey-Nagel, Duren, Germany) according to the manufacturer’s instructions. RNA was reverse transcribed with the iScript cDNA Synthesis kit (Bio-Rad Laboratories, Munich, Germany). qPCR analysis was performed using the SsoFast Eva Green Supermix (Bio-Rad Laboratories), a CFX384 real-time System C1000 Thermal Cycler (Bio-Rad Laboratories) and the Bio-Rad CFX Manager 3.1 software. In other experiments, total RNA was extracted using the PerfectPure RNA cell kit (5 Prime; Fisher, Gaithersburg, MD). RNA was reverse transcribed using the High-Capacity cDNA Archive kit (Applied Biosystems, Foster City, CA) and real-time PCR with cDNA was performed using the ABI 7500 Fast System, according to the manufacturers protocol (Applied Biosystems). The relative amounts of mRNA were quantified with the comparative ΔΔCt method (49). The internal control for normalization was GAPDH mRNA. TaqMan probes were purchased from Life Technologies (Carlsbad, CA) and primers for detection and quantification of genes were from Invitrogen or Life Technologies.
Gene silencing by RNA interference
HUVEC were transfected with small interfering RNA (siRNA) (20 nM; Dharmacon, Lafayette, CO) in combination with Lipofectamine RNAiMAX reagent (Life Technologies) according to the manufacturer’s instructions. For DEL-1 silencing, HUVEC were transfected with On Target plus SMART pool siRNA (Dharmacon) against EDIL3 (the gene encoding DEL-1; hereafter referred to as DEL-1) and control-treated cells were transfected with nontargeting control siRNA pool (Dharmacon). C/EBPβ silencing was described elsewhere (21).
Static cell adhesion assay
Assays were performed according to a previously described protocol (50). Nunc-Immuno MicroWell 96-well plates (Sigma-Aldrich) were coated with 10 μg/ml ICAM-1 (R&D systems) in PBS overnight at 4°C, washed twice with PBS and blocked with 3% BSA (Sigma-Aldrich) in PBS for 1 h in RT. Human neutrophils were isolated from peripheral heparinized blood of healthy volunteers (the procedure was approved by the ethics committee of the Technische Universität Dresden), using double gradient separation (Histopaque 1119 and 1077; Sigma-Aldrich) (11). Erythrocytes were lysed with RBC lysis buffer (eBioscience, San Diego, CA). Neutrophils were stained with 4 μM BCECF (Life Technologies) for 15 min and washed with PBS. Cells were then pretreated with 100 nM DHEA or equivalent amount of ethanol (control vehicle) in RPMI 1640 0.1% BSA medium for 30 min at 37°C and plated onto precoated wells (5 × 104 cells/well) in the presence of 50 ng/ml PMA (Sigma-Aldrich) or equivalent amount of DMSO (as control) in 0.1% BSA RPMI 1640 medium without any intermediate washing steps. Following an incubation period of 30 min, the wells were washed with PBS. Fluorescence was measured before (cell input) and after washing (adherent cells) using a SynergyTM HT multidetection microplate reader (Biotek Instruments, Winooski, VT). The percentage of adhesion is defined as the fluorescence of adherent cells divided by the fluorescence of cell input (51).
Human neutrophils were isolated from peripheral blood of healthy volunteers as described in the static adhesion assay, treated with RBC lysis buffer and washed with HBSS (Invitrogen). Neutrophils were thereafter incubated with 100 nM DHEA or equivalent amount of ethanol (control vehicle) in HBSS containing 1 mM Ca2+, 1 mM Mg2+ and 10 mM HEPES for 20 min at 37°C. Next, they were incubated additionally (without an intermediate washing step) with 10 nM fMLF (Sigma-Aldrich) or equivalent amount of DMSO (control vehicle) for 10 min at 37°C in the presence of Alexa Fluor 488-conjugated mAb24 Ab (BioLegend, San Diego, CA), which binds the open high-affinity conformation of the human αLβ2 integrin (52, 53). Cells were washed with cold HBSS containing 1 mM Ca2+, 1 mM Mg2+ and 10 mM HEPES, centrifuged and mAb24 binding was analyzed by flow cytometry with a FACSCanto II flow cytometer (BD Biosciences, San Jose, CA) and the FACSDiva 6.1.3 software.
HUVEC were treated with DHEA (100 nM) or vehicle control (equal amount of ethanol) followed by addition of TNF (10 ng/ml) or LPS (10 ng/ml) (i.e., TNF or LPS treatment was done in the presence of DHEA [or vehicle control]). HUVEC were then fixed with 0.5% formaldehyde solution for 3 min at 4°C, washed with PBS and incubated for 30 min with anti-ICAM-1-PE (HA58; BD Biosciences), anti-VCAM-1- allophycocyanin (51-10C9; BD Biosciences) or anti-E-selectin- allophycocyanin (68-5H11; BD Biosciences) diluted in PBS containing 0.1% BSA under shaking conditions on ice. Next, cells were washed with PBS, detached with accutase (Thermo Fisher Scientific, Schwerte, Germany) and diluted in PBS supplemented with 2% FCS. Samples were analyzed by flow cytometry (FACSCanto II flow cytometer; FACSDiva 6.1.3 software).
Detection of human DEL-1 protein in culture supernatants was performed by a sandwich ELISA, as described (21), using serial dilutions of recombinant human DEL-1 (R&D systems) for standard curve generation.
Western blot analysis
Western blot was performed as previously described (37). HUVEC were lysed in RIPA lysis buffer supplemented with 1 mM sodium orthovanadate, 2 mM PMSF and protein inhibitor mixture (Santa Cruz Biotechnology, Heidelberg, Germany), centrifuged and the total protein concentration was determined with BCA Protein Assay Kit (Thermo Fisher Scientific). Protein lysates were diluted with NuPAGE LDS sample buffer (Thermo Fisher Scientific) and NuPAGE sample reducing agent (Thermo Fisher Scientific) and boiled at 95°C for 3 min. Equal amounts of cellular protein were loaded in NuPAGE 4–12% polyacrylamide gradient Bis-Tris Protein Gels (Invitrogen) and transferred onto nitrocellulose membranes (GE Healthcare, Freiburg, Germany). Membranes were washed in TBS-T buffer, blocked in TBS-T 5% nonfat milk (BD Biosciences) and immunoblotted with anti–phospho-AKT (Ser473) (#4060; Cell Signaling, Danvers, MA) and anti-vinculin (#4650; Cell Signaling) in TBS-T 5% BSA, followed by incubation with HRP-conjugated secondary Ab (R&D Systems). In other experiments, membranes were immunoblotted with anti–phospho-TRKA (Tyr490) (ab1445; Abcam, Cambridge, MA) in TBS-T 5% BSA, followed by incubation with HRP-conjugated secondary Ab. After extensive washing, membranes were reblotted with anti-vinculin (#4650; Cell Signaling) followed by incubation with secondary Ab. Signal detection was performed using SuperSignalTM West Pico or Femto Chemiluminescent Substrates (Thermo Fisher Scientific) and detected with the Fusion FX7 Multimaging System image analyzer (Peqlab, Erlangen, Germany).
Chromatin immunoprecipitation assay followed by qPCR
Chromatin immunoprecipitation (ChIP) analysis of C/EBPβ binding to the DEL-1 promoter in HUVEC was performed using the SimpleChIP Plus Enzymatic Chromatin IP Kit with magnetic beads (Cell Signaling), as previously reported (21). Cross-linked chromatin was immunoprecipitated with nonimmune rabbit IgG or rabbit IgG Ab to C/EBPβ (Santa Cruz Biotechnology). Nonimmunoprecipitated cell extracts were used as input samples. For qPCR, customized TaqMan-specific primers flanking a C/EBPβ binding site in the DEL-1 promoter (−328 to −589 bp) were used (Thermo Fisher Scientific). ChIP-qPCR data obtained with the C/EBPβ Ab or nonimmune IgG were normalized using the percentage input method that normalizes to chromatin input based on the following equation: % Input = 100 × 2(Ct[adjusted input] − Ct[IP]) (54, 55).
Luciferase reporter assay
The luciferase reporter assay and the construct of human DEL-1 promoter/luciferase reporter plasmid (hDEL-1-promoter-Luc) are described elsewhere (21). HEK-293T cells or HUVEC were seeded on 96-well plates at a density of 5 × 103 cells per well and cotransfected with hDEL-1-promoter-Luc and pGL3 firefly luciferase reporter plasmid (Promega, Madison, WI) as an internal transfection control, using 4D-Nucleofector V4XC (Lonza) for HEK293T cells or FuGENE HD transfection reagent (Promega) for HUVEC. In experiments in which cells were treated with DHEA or NGF, HEK-293T cells were additionally transiently cotransfected with pCMV6-Entry Tagged Cloning control vector (Origene, Rockville, MD) or with NTRK1 cDNA expression plasmid (RC213091; Origene), using TurboFectin 8.0 transfection reagent (Origene), to overexpress TRKA (HEK-293TTRKA). Six hours after transfection, the cells were pretreated either with DHEA (or ethanol; vehicle control) or with NGF (or medium-only control). These pretreatments were followed without an intermediate washing step by stimulation with TNF, as described in the figure legends. HEK-293T cells express TNFR1 and low levels of TNFR2 (56, 57). Luciferase assay was performed using the Dual-Glo Luciferase Assay System and GloMax-Multi Detection System (Promega) according to the manufacturer’s instructions.
A Mann–Whitney U test or Student t test was used for the comparison of two groups. One-way ANOVA followed by Dunnett or Tukey multiple-comparison test was used for multiple group comparisons. A p value <0.05 was considered to be statistically significant. Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, La Jolla, CA).
DHEA inhibits leukocyte adhesion
We initially examined whether DHEA affects leukocyte–endothelial interactions in vivo by performing real-time intravital microscopy in inflamed mouse cremaster muscle venules. To this end, mice were injected i.p. with DHEA or control vehicle solution followed by intrascrotal injection of LPS. After the endotoxin challenge, the cremaster muscle was exposed, and leukocytes adherent to the endothelium were monitored. Intravital microscopy revealed that leukocyte adhesion efficiency, assessed as the number of adherent leukocytes per square millimeter vessel surface area relative to WBC count per microliter of blood, was reduced in DHEA-treated mice, as compared with control vehicle–treated animals (Fig. 1A). Consistent with its antiadhesive action, DHEA treatment increased the leukocyte rolling flux fraction (Fig. 1B) and leukocyte rolling velocity (Fig. 1C, 1D). Circulating WBC numbers and hemodynamic parameters, such as shear rates, vessel diameter, and centerline velocity, did not significantly change upon DHEA treatment (Table I). These data suggest that DHEA inhibits leukocyte adhesion to the endothelium in vivo.
|.||Vessel Diameter (μm) .||Centerline Velocity (μm/s) .||Shear Rate (s−1) .||WBC (Cells/μl) .|
|Control||33.6 ± 1.1||2019.0 ± 132.8||1506.3 ± 115.6||5423.3 ± 464.1|
|DHEA||32.4 ± 0.8||1747.8 ± 104.6||1325.1 ± 74.7||6233.3 ± 569.1|
|.||Vessel Diameter (μm) .||Centerline Velocity (μm/s) .||Shear Rate (s−1) .||WBC (Cells/μl) .|
|Control||33.6 ± 1.1||2019.0 ± 132.8||1506.3 ± 115.6||5423.3 ± 464.1|
|DHEA||32.4 ± 0.8||1747.8 ± 104.6||1325.1 ± 74.7||6233.3 ± 569.1|
Hemodynamic parameters observed in cremaster muscle postcapillary venules of WT mice stimulated with intrascrotal injection of LPS. WBC indicates systemic WBC count (n = 6 mice per group).
DHEA does not directly affect β2-integrin activation or endothelial adhesion molecule expression
Adhesive interactions between β2-integrins on leukocytes and their endothelial counterreceptors are key steps in leukocyte adhesion. To this end, we first examined whether the inhibitory effect of DHEA on leukocyte adhesion is associated with altered leukocyte β2-integrin activation. DHEA did not affect PMA-induced neutrophil adhesion to ICAM-1 (Fig. 2A). Consistently, DHEA did not affect β2-integrin activation. Specifically, DHEA failed to influence β2-integrin activation in fMLF-activated neutrophils, as assessed by flow cytometry using mAb24 (Fig. 2A), an Ab that recognizes an activation-dependent epitope on β2-integrins (58).
Next, we asked whether DHEA could alter endothelial cell adhesion molecule expression. However, surface expression of E-selectin, ICAM-1, or VCAM-1 in LPS- or TNF-treated HUVEC was not affected by DHEA (Fig. 2C–E). These data collectively suggest that the inhibitory effect of DHEA on leukocyte adhesion is neither because of regulation of β2-integrin activation in leukocytes nor because of reduction of adhesion molecule expression in endothelial cells.
DHEA reverses the inhibitory effect of TNF on DEL-1 expression
In the absence of direct mechanisms that could explain the inhibitory effect of DHEA on leukocyte adhesion, we interrogated alternative indirect mechanisms of DHEA. Because DEL-1 is an inhibitor of leukocyte–endothelial adhesion (8), we sought to investigate whether DHEA acts upstream of DEL-1. Specifically, we determined whether DHEA could regulate endothelial DEL-1 expression under inflammatory conditions. Given the importance of TNF on endothelial cell activation in inflammatory conditions (59), we first examined the effect of TNF on DEL-1 production. To this end, HEK-293T cells were transfected with a hDEL-1-promoter-Luc plasmid and then treated or not with TNF. TNF significantly reduced luciferase activity in HEK-293T cells compared with untreated cells (Fig. 3A), thus confirming the validity of this experimental model for mechanistic studies. Similarly, TNF reduced the DEL-1 promoter activity in HUVEC transfected with a hDEL-1-promoter-Luc plasmid (Fig. 3B). Consistently, TNF reduced DEL-1 mRNA expression in HUVEC (Fig. 3C), as we observed before (17). Moreover, TNF diminished the levels of secreted DEL-1 protein in HUVEC, as assessed by ELISA of culture supernatants (Fig. 3D). The validity of the DEL-1 ELISA was confirmed by the reduced DEL-1 concentrations in supernatants of HUVEC pretreated with siRNA against DEL-1 (Fig. 3E). Importantly, DHEA pretreatment reversed the inhibitory effect of TNF on DEL-1 promoter activity (Fig. 3F) and restored DEL-1 secreted protein levels in a dose-dependent manner in HUVEC (Fig. 3G). In fact, the endothelial DEL-1 protein expression in TNF-stimulated HUVEC was fully restored in the presence of a DHEA concentration of 100 nM (Fig. 3G). On the contrary, DHEA did not affect DEL-1 mRNA expression in the absence of TNF (Fig. 3H), suggesting that DHEA regulates DEL-1 expression only under inflammatory conditions. Moreover, DHEA posttreatment did not reverse the inhibitory effect of TNF on DEL-1 expression, suggesting that DHEA acts in a preventive manner (data not shown).
TNF inhibits DEL-1 expression via a C/EBPβ–dependent mechanism
To understand the mechanism, by which DHEA regulates DEL-1, it was imperative to first investigate how TNF downregulates DEL-1 expression. Because we have previously shown that the transcription factor C/EBPβ is recruited to the DEL-1 promoter and maintains constitutive expression of DEL-1 (21), we examined whether TNF could regulate DEL-1 expression via C/EBPβ. Indeed, siRNA-mediated C/EBPβ silencing in HUVEC (Fig. 4A) reduced constitutive DEL-1 mRNA expression and secretion (Fig. 4B, 4C). TNF did not further decrease DEL-1 mRNA or protein levels in C/EBPβ-deficient HUVEC (Fig. 4B, 4C), revealing that the C/EBPβ transcription factor is a critical target of the inhibitory effect of TNF on DEL-1 expression.
DHEA reverses TNF-mediated downregulation of DEL-1 via the PI3K–AKT–C/EBPβ pathway
We next sought to elucidate the mechanisms mediating the promotional effect of DHEA on DEL-1 expression in TNF-stimulated endothelial cells. Given that DHEA activates and exerts anti-inflammatory actions through the PI3K/AKT pathway (36, 37), we initially identified that DHEA induces AKT phosphorylation in HUVEC (Fig. 5A). Next, we asked whether the PI3K/AKT pathway mediates the effect of DHEA on DEL-1 expression. Because C/EBPβ is implicated in the regulation of DEL-1 expression by TNF, we examined the effect of DHEA on C/EBPβ binding on the DEL-1 promoter in TNF-treated HUVEC pretreated with the PI3K inhibitor LY294002 or its inactive control (LY303511) by ChIP. TNF significantly diminished C/EBPβ binding to the DEL-1 promoter, although this effect was reversed by DHEA. PI3K inhibition with LY294002 abrogated the reversal effect of DHEA (Fig. 5B). Moreover, PI3K inhibition with LY294002 or AKT inhibition with MK2206 abolished the effect of DHEA on restoring DEL-1 mRNA expression in TNF-treated HUVEC (Fig. 5C), thus revealing a critical role of the DHEA–PI3K/AKT signaling pathway in the restoration of the endothelial DEL-1 expression under inflammatory conditions.
DHEA restores DEL-1 expression in endothelial cells through TRKA
We next set out to identify the receptor mediating the promotional effect of DHEA on DEL-1 expression in TNF-treated HUVEC. In this regard, the neurotrophin receptor TRKA is expressed in endothelial cells and can activate the PI3K/AKT signaling pathway in response to DHEA in other cell types (37, 38, 60, 61). We thus reasoned that TRKA may mediate the effect of DHEA on DEL-1 expression in TNF-stimulated endothelial cells. To this end, we first showed that DHEA induced TRKA phosphorylation in HUVEC (Fig. 6A). Next, we performed a luciferase assay in HEK-293T cells cotransfected with the hDEL-1-promoter-Luc construct and a plasmid overexpressing human TRKA (HEK-293TTRKA). DHEA reversed the TNF-mediated decrease in DEL-1 promoter activity in HEK-293TTRKA but failed to do so in HEK-293T cells (Fig. 6B), which do not express TRKA (38, 62). Moreover, a TRKA inhibitor abolished the effect of DHEA on DEL-1 mRNA expression in TNF-treated HUVEC (Fig. 6C). Consistent with these findings, the prototype ligand of TRKA, NGF, also restored DEL-1 mRNA expression in TNF-treated HUVEC and DEL-1 promoter activity in HEK-293TTRKA cells transfected with hDEL-1-promoter-Luc plasmid (Fig. 6D, 6E). These data collectively suggest that DHEA restores DEL-1 expression in TNF-activated endothelial cells via the TRKA receptor.
DHEA inhibits neutrophil recruitment in WT but not DEL-1–deficient mice
We next determined the biological significance of the ability of DHEA to rescue endothelial DEL-1 expression under inflammatory conditions. Specifically, we tested whether DHEA could inhibit neutrophil recruitment in an in vivo model of inflammation. We engaged the model of LPS-induced lung inflammation, in which neutrophil recruitment is regulated by DEL-1 (8). DEL-1–sufficient (DEL-1+/+ or WT) and –deficient (DEL-1−/−) mice received i.p. injections of DHEA or vehicle control, followed by intranasal administration of LPS, and the number of infiltrated neutrophils was determined in the BAL fluid. WT mice pretreated with DHEA displayed reduced numbers of recruited neutrophils in the alveolar space as compared with vehicle-treated animals (Fig. 7A). In contrast, the anti-inflammatory effect of DHEA was abrogated in DEL-1−/− mice, suggesting that the inhibitory effect of DHEA on neutrophil recruitment requires the presence of DEL-1 (Fig. 7B).
Endothelial cell–derived DEL-1 can regulate the initiation of inflammation by controlling leukocyte–endothelial adhesion and, thereby, leukocyte recruitment (8, 9). Reduced DEL-1 expression is observed in several inflammatory conditions, such as human and murine periodontitis, multiple sclerosis, and its murine counterpart, experimental autoimmune encephalomyelitis (EAE), inflammation-related adrenal gland dysfunction, and in human and murine inflammatory lung pathologies (7, 14, 15, 17–19, 63). Whereas reduced DEL-1 expression leading to increased inflammatory cell recruitment may be beneficial in acute infections, decreased DEL-1 levels in a chronic setting may exacerbate inflammatory pathologic conditions as seen in naturally occurring periodontitis in mice (9). Thus, identification of factors restoring DEL-1 expression is of particular importance as an anti-inflammatory therapeutic approach.
In this study, we demonstrate that the steroid hormone DHEA can restore the inflammation-mediated decrease in endothelial DEL-1 expression, thereby inhibiting leukocyte recruitment and, hence, inflammation. Although DHEA diminished leukocyte adhesion efficiency, resulting in increased leukocyte rolling flux fraction and rolling velocity, these effects of DHEA were not mediated by direct interference with leukocyte–endothelial cell interactions. Indeed, DHEA did not regulate β2-integrin–dependent adhesion or the expression of endothelial cell adhesion molecules but, rather, the expression of the β2-integrin antagonist DEL-1. The requirement of DEL-1 for the anti-inflammatory effects of DHEA was conclusively demonstrated in LPS-induced lung inflammation, in which DHEA inhibited acute neutrophil recruitment in DEL-1–sufficient but not –deficient mice.
Endothelial DEL-1 expression is diminished by acute inflammation (8, 9, 21), in which TNF plays a major role as an activator of the endothelium (64). In this study, we show that the TNF-mediated downregulation of endothelial DEL-1 expression (8, 17) occurs at the transcriptional level through blockage of C/EBPβ binding to the DEL-1 promoter. DHEA can reverse the inhibitory effect of TNF on endothelial DEL-1 expression through a pathway that involves the TRKA receptor and downstream PI3K/AKT signaling pathway, thereby restoring C/EBPβ binding to the DEL-1 promoter. Our findings thus provide a mechanistic understanding of the anti-inflammatory action of DHEA. Consistently, NGF, the prototype ligand of TRKA, displayed a similar regulatory effect on DEL-1 expression. Additionally, it would be of interest to know if other steroids with known anti-inflammatory actions may share the regulatory effect of DHEA on DEL-1 expression under inflammatory conditions (65).
Steroid hormones may regulate inflammation, although their synthesis is strongly affected by inflammation (66–68), thus suggesting an intimate cross-talk between inflammatory responses and steroid hormones. For instance, DHEA sulfate serum levels are reduced in autoimmune diseases, such as rheumatoid arthritis, Sjörgen syndrome, and systemic lupus erythematosus (69–72). Moreover, DHEA levels are reduced in the CNS of multiple sclerosis patients, whereas DHEA administration downregulates inflammation in rodent models of LPS-induced brain inflammation or EAE (32, 37, 73, 74). Similarly, DEL-1 levels are decreased in the CNS of multiple sclerosis patients and in the course of EAE (19). Interestingly, high expression of DEL-1 is observed in immunoprivileged organs, such as the brain, which also exhibit relatively high concentrations of NGF and DHEA (8, 19, 29, 75–77). Intriguingly, aging in humans is associated with increased circulating levels of TNF and decreased levels of DHEA, although an age-related decline of DEL-1 expression is described in mice and humans as well (9, 26, 63, 78). It is thus tempting to speculate that the inverse associations between inflammatory cytokines, such as TNF, and the anti-inflammatory factors DHEA and DEL-1 in the context of inflammatory diseases and aging may be mechanistically linked and may represent an “inflamm-aging” signature. Furthermore, it is possible that the aging-related decline in DEL-1 expression might, in part, be secondary to the reduced levels of DHEA, given our current findings that the latter positively regulate the former.
We have recently shown that the spatial cellular distribution of DEL-1 determines distinct functions of DEL-1. Whereas endothelial cell–derived DEL-1 regulates leukocyte recruitment, macrophage-derived DEL-1 promotes resolution of inflammation by mediating phagocytosis of apoptotic neutrophils (efferocytosis) by macrophages, leading to the reprogramming of the latter toward a proresolving phenotype (11). Given our current findings that DEL-1 acts downstream of DHEA, it can be reasoned that DHEA may have a proresolving action. Thus, future studies could address whether DHEA is present at sufficient levels during the resolution phase and promotes the expression of DEL-1 in macrophages, thereby indirectly contributing to successful inflammation resolution.
In conclusion, our findings introduce the DHEA–DEL-1 axis as a novel immune-endocrine homeostatic mechanism that regulates neutrophil recruitment and inflammation, suggesting that DHEA could provide a novel therapeutic approach in diseases associated with excessive neutrophil recruitment or related with reduced DEL-1 expression.
This work was supported by grants from the Deutsche Forschungsgemeinschaft (AL1686/2-2 and AL1686/3-1 to V.I.A., CH279/6-2 to T.C., SFB/TRR 205 to V.I.A. and T.C., and SP621/5-1 and SFB914 TP B01 to M.S.), the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie Grant Agreement (765704 to V.I.A.), the European Research Council (DEMETINL to T.C.), and by National Institutes of Health grants (DE015254, DE024153, and DE024716 to G.H. and DE026152 to G.H. and T.C.).
Abbreviations used in this article:
developmental endothelial locus 1
experimental autoimmune encephalomyelitis
human DEL-1 promoter/luciferase reporter plasmid
nerve growth factor
quantitative real-time PCR
small interfering RNA
tropomyosin-related kinase A
The authors have no financial conflicts of interest.