Endothelial Cell–Derived Chemerin Promotes Dendritic Cell Transmigration

The correct localization of dendritic cells (DC) to peripheral tissues and secondary lymphoid organs is a crucial event for optimal immune responses (1). Multiple chemotactic signals regulate the trafficking of myeloid DC (mDC) and plasmacytoid DC (pDC), the two main circulating DC subsets (2). Chemotactic agonists active on DC include chemokines, cytokines, formyl peptides, complement fragments, and bioactive lipid molecules (3, 4).

Chemerin was originally described as the product of the Tazarotene-induced gene 2 (Tig2) in stimulated skin cultures of psoriatic patients (5) and subsequently purified from ascetic fluids from ovarian cancer patients and synovial exudates from rheumatoid arthritis patients (6). Chemerin is secreted as a poorly active precursor that is converted into a bioactive agonist following the proteolytic removal of the last 6 or 7 aa from the C terminus (7). Various proteases mediate chemerin processing; these include neutrophil serine proteases and proteases from the coagulation and fibrinolytic cascades (8). Many tissues express chemerin in a constitutive manner, although the nature of the producing cells is still largely unknown (6). ChemR23, the functional chemerin receptor (also known as CMKLR1 in humans and Dez in the mouse), is a G protein–coupled receptor expressed in various leukocyte populations, including mDC, pDC, monocytes, macrophages, and NK cells (9–13). More recently, two other high-affinity chemerin receptors were described, as follows: GPR1, a poor signaling receptor mainly expressed in the CNS, and CCRL2/ACKR5, a member of the atypical chemokine receptor family (also known as LCCR in the mouse) (8, 14). CCRL2 is a nonsignaling receptor that was proposed to bind and concentrate bioactive chemerin to ChemR23-positive cells (15, 16).

Chemerin is normally present in circulation of healthy subjects, and its levels are upregulated in many inflammatory conditions, such as lupus erythematosus, psoriasis, rheumatoid arthritis, and Crohn’s disease (17). We previously reported chemerin immunostaining in decidual cells during early pregnancy, with highest levels present in stromal and extravillous trophoblast cells (18). In addition, chemerin was found to be associated with tubular epithelial cells and renal lymphatic endothelial cells in patients with lupus nephritis but not in normal kidneys (19). Chemerin expression also colocalized with high endothelial venules in lymph nodes and with inflamed endothelial cells in skin biopsies obtained from patients with autoimmune diseases, such as systemic lupus erythematosus, lichen planus, and psoriasis (10, 11, 20). However, at present it is unknown whether endothelial cells produce chemerin or only bind and present on their surface the protein. Of note, in all cases, chemerin⁺ endothelial cells were surrounded by ChemR23⁺ DC postulating a role for the ChemR23/chemerin axis in DC trafficking across endothelial cell barriers.

The aim of the current study was to investigate the regulation of chemerin expression and production by vascular and lymphatic endothelial cells and to evaluate the role of the chemerin/ChemR23 axis in the transmigration of DC subsets across endothelial cell barriers.

DMEM (4.5 g/L glucose), RPMI 1640 medium, heat-inactivated FBS, penicillin/streptomycin, nonessential amino acids, and Na pyruvate were from Lonza (Verviers, Belgium). The 1α,25 dihydroxyvitamin D3 (calcitriol), all trans retinoic acid (RA), BSA (A2934), gelatin type B from bovine skin, and heparin were from Sigma-Aldrich (St. Louis, MO). Recombinant mouse chemerin (chemerin), TNF-α, and IFN-γ were from R&D Systems (Minneapolis, MN). Endothelial cell growth supplement was from Biomedical Technologies; 2-ME was from Life Technologies (Scotland, U.K.); CpG was from InvivoGen (San Diego, CA); and Flt3L, GM-CSF, CCL3, CCL19, and CXCL12 were from PeproTech (Rocky Hill, NJ). LPS (Escherichia coli strain 055:B5) was from Difco Laboratories (Detroit, MI). Sarcoma 180 (S180) cell line was from American Type Culture Collection.

The mouse lung capillary endothelial cell line (1G11) was grown in complete medium (DMEM, 20% FBS, 1% nonessential amino acids, 1 mM Na pyruvate, 100 U/ml penicillin and streptomycin, freshly added heparin, and endothelial cell growth supplement at final concentration of 100 μg/ml), as described (21). The mouse lymphatic endothelial cell line (MELC) was cultured in complete medium supplemented with 10% supernatant from sarcoma 180 cells (growth medium), as described (22). Confluent cells were passed routinely at a split ratio of 1:3 after trypsin/EDTA digestion. Both endothelial cell lines were grown on flasks precoated with 1% gelatin. Mouse embryonic fibroblasts and 3T3L1 cells were prepared and cultured, as described (23, 24).

Eight- to 12-wk-old C57BL/6J mice (Charles River Laboratories) and ChemR23 knockout (KO; provided by M. Parmentier [Université Libre de Bruxelles, Brussels, Belgium]) were used. CD34⁺ bone marrow–positive cells were purified by positive immunoselection and cultured, as previously described (25, 26), with murine GM-CSF (40 ng/ml) and Flt-3L (100 ng/ml) to generate mDC, or with Flt3L (200 ng/ml) only to generate pDC. Cells were split every 2–3 d, and used at day 9. DC were matured in the presence of TNF-α (20 ng/ml), LPS (100 ng/ml), or CpG (2 μg/ml) for the indicated time points.

PBMCs were isolated from buffy coats of normal donors by Ficoll gradient (Ficoll-Paque Premium; GE Healthcare, Life Sciences) and were magnetically sorted with blood pDC Ag BDCA-4 cell isolation kits (Miltenyi Biotec), as previously described (27).

DC were matured in the presence of TNF-α (20 ng/ml) or LPS (100 ng/ml) for 24 h and, after blocking Fc-nonspecific binding with anti-CD16/32 (clone 2.4G2), incubated with an anti–ChemR23-PE (clone 477806) mAb or rat IgG2b-PE isotype-matched mAb (R&D Systems). The 1G11 cells were stimulated with RA (5 μM) or with IFN-γ/TNF-α/LPS (50 ng/ml, 20 ng/ml, 1 μg/ml, respectively) for 18 h. After Fc blocking, the cells were stained with an anti-CCRL2 mAb (28), washed with PBS (Life Technologies), and then incubated with an anti-mouse Alexa 488 mAb (Invitrogen, Life Technologies, Monza, Italy). Samples were read on a Particle Analysing System cytofluorimeter (Partec, Muenster, Germany), and results were analyzed by FlowJo software (Tree Star).

Mouse endothelial cell lines were grown in six-well culture plates (BD Labware) till confluence. Then medium was replaced with medium with 0.2% BSA without serum and growth factors. Cells were stimulated with RA (5 μM), 1α,25 dihydroxyvitamin D3 (calcitriol, 1 μM), and TNF-α (20 ng/ml) for reported times. DC and endothelial cell total RNA was isolated using TRIzol reagent (Invitrogen), according to the manufacturer’s specifications. Total RNA (1 μg) was reverse transcribed using a high-capacity cDNA reverse-transcription kit from Applied Biosystems, following manufacturer’s instructions. RT-PCR was performed on cDNA samples using a Power SYBR Green PCR master mix (Applied Biosystems, Warrington, U.K.) and specific primers. RT-PCRs were performed on a 7900HT Fast Real-Time PCR System machine according to manufacturer’s guidelines (Applied Biosystems). The primer pairs used were as follows: mouse ChemR23 (forward: 5′-CCATGTGCAAGATCAGCAAC-3′, reverse: 5′-GCAGGAAGACGCTGGTGTA-3′), CCRL2 (forward: 5′-TGTGTTTCCTGCTTCCCCTG-3′, reverse: 5′-CGAGGAGTGGAGTCCGACAA-3′), mCCR1 (forward: 5′-CTGCCCCCCCTGTATTCTCT-3′, reverse: 5′-GACATTGCCCACCACTCCA-3′), mCCR2 (forward: 5′-CTACGATGATGGTGAGCCTTGTC-3′, reverse: 5′-AGCTCCAATTTGCTTCACACTG-3′), mCCR7 (forward: 5′-TGGTGGTGGCTCTCCTTGTC-3′, reverse: 5′-CCTCATCTTGGCAGAAGCACA-3′), mCCL3 (forward: 5′-CATATGGAGCTGACACCCCG-3′, reverse: 5′-TCTTCCGGCTGTAGGAGAAGC-3′), mCCL4 (forward: 5′-GCCCTCTCTCTCCTCTTGCT-3′, reverse: 5′-GAGGGTCAGAGCCCATTG-3′), mCCL5 (forward: 5′-TGCTCCAATCTTGCAGTCGT-3′, reverse: 5′-ACACACTTGGCGGTTCCTTC-3′), mCXCL10 (forward: 5′-CGTCATTTTCTGCCTCATCCTG-3′, reverse: 5′-CCGTCATCGATATGGATGCAGT-3′), mCXCL12 (forward: 5′-CTGTGCCCTTCAGATTGTTG-3′, reverse: 5′-TAATTTCGGGTCAATGCACA-3′), hβ-actin (forward: 5′-TCACCCACACTGTGCCCATCTACGA-3′, reverse: 5′-CAGCGGAACCGCTCATTGCCAATGG-3′), and mGAPDH (forward: 5′-CGTGTTCCTACCCCCAATGT-3′, reverse: 5′-TGTCATCATACTTGGCAGGTTTCT-3′).

The primers/fluorogenic (FAM) probe sets Mm 00503581_gH and Hs 99999901_s1 from Applied Biosystems (Branchburg, NJ) were used to amplify mouse chemerin and 18S, respectively. In this case, a TaqMan Universal PCR master mix was used (Applied Biosystems).

Supernatants of endothelial cells were collected at the indicated time points, centrifuged, and stored at −20°C until ELISAs. The supernatants were concentrated 10 times using a vivaspin concentrator unit (3000 MWCO; Sartorius Stedim Biotech) and examined for the presence of chemerin by ELISA, according to manufacturer’s instructions (Mouse Chemerin Quantikine ELISA kit; R&D Systems).

The 1G11 cells (7.5 × 10⁴) were seeded on glass coverslips precoated with 1% gelatin in complete medium and then stimulated with RA (5 μM) for 24 h. After 1% paraformaldehyde fixing, the slides were saturated with 1% BSA in PBS and incubated with an anti-chemerin Ab (Bioss, Woburn, MA) or an anti–VCAM-1 (clone M-K-2; Antibodies On-Line, Atlanta, GA), followed by anti-rabbit Alexa 488 and anti-rat Alexa 594 (Invitrogen) Abs, respectively. Nuclei were counterstained with DAPI. Analysis was performed using a Zeiss Axio Observer Z1 microscope equipped with Apotome system and a Plan-Apochromat 100×/1.4 NA oil objective.

Endothelial cells were grown to confluence on 0.1% gelatin-coated transwell inserts in 24-well costar chambers (5 μm pore size; Corning). When indicated, the endothelial cells were exposed to different stimuli in RPMI 1640 medium containing 0.2% BSA (migration medium) for 18 h. For the transmigration assays, 100 μl DC (0.5 × 10⁶ cells/ml ⁵¹Cr-labeled mouse DC; 2 × 10⁶ cells/ml human DC) in migration medium was placed in the upper chamber and 600 μl chemoattractant or control medium was added to the lower chamber. DC were allowed to migrate for 90 min (mouse DC) or 4 h (human DC) at 37°C in a 5% CO₂ atmosphere. Mouse DC migration was evaluated as the percentage of radioactivity recovered in the lower compartment relative to input. Values are expressed as net migration (% radioactivity sample − % radioactivity control). Human DC were collected in the lower chamber and counted. The results were expressed as percentage of cell input in the upper chamber. To block ChemR23, human pDC were incubated (30 min at 4°C) with an anti-ChemR23 mAb (clone 1H2, IgG2a, 3 μg/ml) donated by M. Parmentier (6) and then added to the upper chamber. Chemerin was added to endothelial cells for 90 min at 37°C in the upper chamber and then removed before addition of DC to be tested.

To analyze DC recruitment, we used an in vitro flow chamber assay. For this purpose, glass capillaries (2 × 0.2 mm; VitroCom, Mountain Lakes, NJ) were assembled into microflow chambers, as described previously (29). For adhesion studies, a combination of recombinant murine P-selectin (20 μg/ml), chemerin (10 μg/ml), and integrin-ligand murine rVCAM-1 (CD106, 15 μg/ml) and murine rCCL3 (10 μg/ml) from R&D Systems was used. After overnight incubation, the flow chambers were blocked with 5% casein (from bovine milk; Sigma-Aldrich, Munich, Germany) in PBS for 2 h at room temperature and flushed with PBS. The DC suspension (2 × 10⁶/ml) was flushed through the flow chamber using a high-precision syringe pump (Harvard Apparatus, Holiston, MA) at a flow rate of 1 μl/min, resulting in a shear stress of ∼1 dyne/cm². Results were evaluated after 10 min of perfusion. DC adhesion to 1G11 cells monolayer, grown on ibidi chambers, was investigated under shear stress conditions. The 1G11 cells were incubated with RA (5 μM) overnight or with CCL3 (100 ng/ml) or chemerin (100 ng/ml) for 90 min. The chambers were connected to an intramedic polyethylene tube (inner diameter 0.58 mm; outer diameter 0.965 mm; BD Biosciences) and perfused with mDC at the concentration of 0.5 × 10⁶/ml and a flow rate of 0.28 ml/min, resulting in a shear stress of ∼0.5 dyne/cm² based on the manufacturer’s description (ibidi Application Note #11; www.ibidi.com) (30).

Measurement of the binding of mouse rVCAM-1/CD106 Fc chimera (R&D Systems) to DC was performed in the following manner: 3 × 10⁵ DC in HBSS, 1 mM HEPES, and 0.25% BSA (pH 7.4) were incubated with the master mix containing murine rVCAM-1/CD106 Fc chimera (400 μg/ml), anti-human IgG Fcγ-specific biotin (eBioscience, San Diego, CA), PE-Cy5 streptavidin (BD Biosciences, San Jose, CA) plus murine CCL3 (100ng/ml) or murine chemerin (100 ng/ml) for the indicated time points at 37°C. The binding reaction was stopped by the addition of FACS lysing solution (BD Biosciences), and the cells were washed twice in ice-cold HBSS. Then the samples were acquired by FACSCalibur (BD Biosciences) and analyzed using CellQuest software.

Results are expressed as mean ± SEM. Statistical significance was determined by Student t test or one-way ANOVA, as appropriate. Differences were considered significant when p < 0.05.

First, the expression level of ChemR23 in mouse bone marrow–derived mDC and pDC was investigated by RT-PCR. Similarly to the data obtained with mouse tissue pDC (31) or with human DC (10), the expression of ChemR23 was easily detected in both mDC and pDC. As expected, based on previous data obtained with human DC, mRNA and protein levels were rapidly downregulated following activation with LPS and TNF-α, or CpG, with LPS being already effective after 1-h incubation (Fig. 1A, 1B, Supplemental Fig. 1A, and data not shown). The expression of ChemR23 was functional, because chemerin was able to promote in a dose-dependent manner the migration of DC subsets with a peak of activity observed at 100 pM chemerin. Conversely, chemerin was not active on activated DC nor in DC derived from ChemR23-deficient bone marrow precursors (Fig. 1C, 1D, and data not shown).

Previous data have shown by immunohistochemistry that, in human pathological conditions, chemerin expression is associated with activated blood and lymphatic endothelial cells (10, 11, 19), but no evidence is available on the nature of the signals involved in chemerin regulation. Two mouse endothelial cell lines, one of lymphatic origin (MELC) (22) and the other of vascular derivation (1G11) (21), were used to better understand the possible role of chemerin production by endothelial cells. A constitutive basal expression of chemerin was detected in both cell lines (Fig. 2A–D). Both vascular and lymphatic endothelia were then stimulated with RA, calcitriol, and a variety of proinflammatory stimuli, such as TNF-α, LPS, IFN-γ, and IL-1β. Among the different agonists, only RA was able to induce the upregulation of chemerin mRNA expression (∼4-fold increase over control) in both MELC and 1G11 cells following overnight stimulation with the optimal concentration of 5 μM (Fig. 2A, 2B, and data not shown). This effect was paralleled by the concomitant secretion of chemerin, as evaluated by ELISA with 0.076 ± 0.006 and 0.879 ± 0.28 ng/10⁶ cells (n = 5) released in 24 h by MELC and 1G11 cells, respectively (Fig. 2C, 2D). No statistically significant release of chemerin was observed with calcitriol and with any of the proinflammatory agonists tested. Thus, these results indicate that RA is active in inducing chemerin production in both lymphatic and vascular endothelial cells in vitro, although the levels of chemerin released were ∼1 log higher in 1G11 cells. When compared with mouse embryonic fibroblasts and with differentiated 3T3L1 adipocytes, 1G11 cells were, respectively, 3.1- and 4.2-fold more efficient in chemerin production (data not shown). The action of RA on 1G11 cells was apparently selective for chemerin, because no mRNA induction of other DC bioactive CC (CCL3, CCL4, and CCL5) and CXC (CXCL10 and CXCL12) chemokines was detected under the same experimental conditions. This selective regulation of chemerin RA was in contrast with the action of TNF-α, a prototypic proinflammatory cytokine, which was able to induce chemokine production, but not chemerin, in both endothelial cell types (Fig. 2E, 2F).

Fig. 3 shows that chemerin could be detected by immunocytochemistry on the plasma membrane of RA-activated endothelial cells, with a distribution similar to that of VCAM-1. On the contrary, no immunostaining could be observed in resting 1G11 cells. This result demonstrates that at least part of the chemerin produced by activated endothelial cells remains associated to the plasma membrane.

Several reports suggested chemerin as a relevant chemotactic factor for myeloid and plasmacytoid DC transmigration across endothelial cells (32). To further explore this hypothesis, we tested the ability of endothelial cell–derived chemerin to induce DC transmigration using MELC and 1G11 cells stimulated with RA in the absence of any added chemotactic factor. Fig. 4A and 4B show that both vascular and lymphatic RA-stimulated endothelial cells were able to sustain the migration of mDC and pDC with efficacy comparable to that of 100 pM recombinant chemerin. As expected, TNF-α–, LPS-activated mDC, or CpG-activated pDC were unable to transmigrate across the RA-activated endothelial monolayer but efficiently migrated in response to CCL19 (Fig. 4A, 4B). Two different approaches were used to formally demonstrate the role of membrane-bound chemerin in DC transmigration across RA-stimulated endothelial cells. First, Fig. 4C shows that ChemR23 KO DC did not migrate across activated 1G11 cells. Second, Fig. 4D shows that the transmigration of human pDC across RA-stimulated cells was completely blocked in the presence of a specific anti-ChemR23 blocking mAb (1H2) (10, 19, 20). Conversely, no effect was observed with an irrelevant isotype-matched control mAb. The action of the anti-ChemR23 mAb was specific because pDC transmigration in response to CXCL12 was not affected by the treatment. These results formally prove that membrane-bound endothelial cell–derived chemerin promote DC transmigration.

Recently, it was reported that the atypical chemotactic receptor CCRL2 binds and concentrates chemerin at the endothelial cell surface (33). Therefore, it was investigated whether a similar mechanism could function in mouse vascular 1G11 endothelial cells. To this goal, vascular endothelial cell monolayers were treated with either RA or a mixture of proinflammatory stimuli (IFN-γ/TNF-α/LPS) known to induce CCRL2 expression (33). As shown in Fig. 5A, in both experimental conditions CCRL2 expression was induced at the mRNA and protein levels in 1G11 cells. The mixture of proinflammatory stimuli (IFN-γ/TNF-α/LPS) also increased ICAM-1 expression but not the expression of VCAM-1 and several other chemokines (Fig. 5B). To functionally evaluate the contribution of CCRL2 expression on the ability of endothelial cells to support chemerin-induced DC transmigration, resting or proinflammatory agonist-activated 1G11 cell monolayers were preincubated with chemerin for 90 min, and then supernatant was removed and DC was added to the transwells for the transmigration assay. As shown in Fig. 5C, the basal net migration of DC across activated endothelial cells was not different from that across resting endothelial cells, but the addition of chemerin to cytokine-activated endothelial cells induced DC transmigration more efficiently (2.3-fold increase) than to resting endothelial cells. These results strongly support the concept that CCRL2 expressed on endothelial cell membrane binds chemerin and promotes the transmigration of ChemR23⁺ DC.

Inside-out signaling generated by activated chemotactic receptors promotes leukocyte transmigration through the induction of the high-affinity integrin conformation (34, 35). To gain further insights on the possible role of chemerin in DC transmigration, mDC rolling and adhesion were quantified under dynamic conditions using a microflow chamber at a shear stress of 1 dyne/cm², resembling the physiological shear stress found in postcapillary venules in vivo (36). Flow chambers were coated with a combination of P-selectin, VCAM-1, and chemerin, or CCL3 as reference chemokine acting on mDC (3). Fig. 6A reports the number of cells rolling and undergoing adhesion at 10 min of perfusion and shows that chemerin caused a 3-fold increase in cell adhesion to the immobilized substrate, an effect similar to that obtained in the presence of CCL3. These results indicate that chemerin is able to induce the rapid activation of β₁ integrin affinity. This conclusion is further supported by the increased ability of chemerin-activated mDC to bind soluble VCAM-1/CD106 Fc chimera protein, as evaluated by flow cytometry. Once again, the effect of chemerin was comparable to that observed with an optimal (100 ng/ml) concentration of CCL3 (Fig. 6B). Finally, the ability of chemerin to induce adhesion of mDC was further exploited using flow chambers coated with 1G11 cells subsequently stimulated with RA. Fig. 6C shows that also under dynamic conditions DC could efficiently bind to activated endothelial cells in a chemerin-dependent manner. Indeed, ChemR23 KO DC could perfectly respond to CCL3 but failed to adhere to RA-activated endothelial cells.

This study identifies vascular and lymphatic endothelial cells as an important source of bioactive chemerin and provides evidence to support the role of the chemerin/ChemR23 axis in DC transmigration across endothelial cell barriers.

Chemerin is a chemotactic agonist that becomes activated following proteolytic processing. Active chemerin binds with high-affinity ChemR23, a heptahelic G protein–coupled chemotactic receptor involved in the recruitment of DC, macrophages, and NK cells (3, 8). Although the chemerin inactive precursor is normally found in the plasma of healthy donors, increased levels of chemerin are detected in many pathological conditions, including infectious and metabolic diseases (8, 37). The nature of the cells responsible for chemerin production, as well as the signals involved in chemerin regulation, is poorly understood. In previous reports, we described that chemerin immunoreactivity could be detected on the apical side of endothelial cells in lymph nodes and in the skin of patients with autoimmune diseases, such as systemic lupus erythematosus, lichen planus, and psoriasis (10, 11, 19, 20). Both blood and lymphatic endothelium were positive by immunohistochemistry, but the nature of the chemerin-producing cells is still elusive.

This study was performed to directly address the question as to whether endothelial cells are capable of chemerin production and can support the transmigration of ChemR23⁺ DC. For this purpose, two mouse endothelial cell lines were used to investigate chemerin expression and regulation, namely MELC, of lymphatic origin (22) and 1G11, of vascular derivation 1G11 (21), previously isolated and characterized by our group. Among the different agonists tested, RA was identified as the only agonist able to induce chemerin production both at the mRNA and protein level. Blood and lymphatic endothelial cells produced chemerin in the supernatant at low levels under basal conditions, and the production increased after an overnight stimulation, with 1G11 cells releasing ∼10-fold higher levels than MELC cells. These results are in agreement with the original description of chemerin as a Tazarotene (a RA-specific retinoid)-inducible gene (Tig2) in the skin of psoriatic patients (5). Conversely, as previously reported with human endothelial cells (10), proinflammatory agonists, such as TNF-α, IL-1β, IFN-γ, and LPS, were not active. Similarly, calcitriol, the strongest agonist able to induce chemerin production in human skin fibroblasts (20), was ineffective in both types of endothelial cells, suggesting a cell-specific regulation of chemerin production. RA-producing DC are localized in the lung, skin, and intestine and in their draining lymph nodes, and the expression of aldehyde dehydrogenase, the enzyme that controls RA production, is increased following infections and stimulation of TLRs (38). Thus, these results suggest that chemerin production by endothelial cells stimulated by RA released by aldehyde dehydrogenase–positive DC may represent a positive feedback mechanism of DC to promote the recruitment of ChemR23⁺ cells. RA-stimulated endothelial cells promoted DC transmigration in the absence of any added chemotactic agonist. The migration was comparable to that induced after the preincubation of the endothelial cell monolayer with recombinant chemerin and was completely blocked in the presence of a specific anti-ChemR23 mAb. These results imply two different concepts, as follows: first, chemerin released by activated endothelial cells undergoes proteolytic activation by endothelial cell–derived proteases; second, at least some of the released protein becomes membrane associated and available for the recognition by ChemR23⁺ DC.

Prochemerin C-terminal processing was shown to occur in the presence of different proteases, such as those released by activated neutrophils, or generated during the coagulation and fibrinolytic cascades (39, 40), or by circulating carboxypeptidases (41). The production of bioactive chemerin was reported in different cell lines transfected with the chemerin full-length cDNA, including CHO-K1, COS-7, and HEK293 cells, suggesting the ability of membrane-associated proteases to process prochemerin (6). The finding that DC can undergo ChemR23-dependent transmigration across RA-activated endothelial cells implies that both MELC and 1G11 cells are able to process and secrete bioactive chemerin in the absence of the contributions of other inflammatory cells.

Immobilization of chemotactic agonists on the surface of endothelial layers is crucial for maintaining the chemotactic gradient under shear stress conditions and for leukocyte extravasation (42, 43). Chemerin is positively charged, and it is likely to bind negatively charged heparin or sulfated glycosaminoglycans expressed on the endothelial cell membranes. In addition, chemerin was reported to bind the atypical chemotactic receptor CCRL2/ACKR5 as a mechanism to concentrate bioactive chemerin on the endothelial cell surface (16, 33). CCRL2 is constitutively expressed by MELC and 1G11 cells and can be further upregulated by RA and by a mixture of TNF-α, IFN-γ, and LPS acting in a synergistic manner (16, 33). Conceivable with CCRL2 induction, activated 1G11 endothelial cells support more efficiently than resting cells chemerin-induced DC transmigration. Because activation of 1G11 cells did not result in the upregulation of DC active chemokines (e.g., CCL3, CCL4, CCL5, CXCL10, and CXCL12) or in a general upregulation of integrin expression, it is tempting to speculate that the increased transmigration of DC is due to the improved binding of chemerin to CCRL2 expressed by activated endothelial cells.

In the multistep process of leukocyte extravasation, chemotactic factors play a crucial role in promoting firm adhesion of leukocyte to endothelial cells through the activation of integrin adhesive properties. The engagement of chemotactic receptors generates an inside-out signaling leading to the increase integrin affinity and clustering (35, 44). In our experimental conditions, chemerin was able to induce a rapid increase of the binding of soluble VCAM-1/CD106 Fc chimeric protein in DC, a sign of the increase of β₁ integrin-binding affinity. In addition, to better understand the potential role of chemerin in DC extravasation, we performed adhesion assays under dynamic conditions. Flow chambers assays were performed in the presence of immobilized P-selectin, VCAM-1, and chemerin at the shear stress of 1 dyne/cm². This experimental approach revealed that chemerin was as effective as CCL3, a reference chemokine, in inducing DC arrest on coated glass capillary surfaces, and similar results were obtained using flow chambers coated with RA-activated endothelial cells. Taken together, these two experimental approaches indicate that chemerin, in addition to the previously documented ability to induce β₁ integrin clustering (13), can also promote in DC the high-affinity conformational state of β₁ integrin and induce VCAM-1–dependent cell arrest.

In conclusion, this study reports that chemerin is produced by activated vascular and lymphatic endothelial cells and becomes associated with the endothelial cell membrane through the possible interaction with membrane proteoglycans and CCRL2. Endothelial cell–derived chemerin promotes β₁ integrin activation and directional migration of DC across activated endothelial cell monolayer. This mechanism is apparently tightly regulated. First, the activation of TLRs by infectious or autoimmune stimuli is responsible for the upregulation of inflammatory cytokines that will induce the production of RA by aldehyde dehydrogenase–expressing DC and the upregulation of CCRL2 in endothelial cells (33, 38). Second, RA will further upregulate the expression of CCRL2 and the production of prochemerin by activated endothelial cells. Finally, endothelial cell–associated proteases will proteolytically generate biologically active chemerin. Taken together, these results show that the ChemR23/chemerin axis represents a new pathway for activated endothelial cells to promote DC extravasation and provides functional implications to the previously described observation of chemerin immunoreactivity of endothelial cells in pathological tissues (32).

The authors have no financial conflicts of interest.