The activation of the complement system is a key initiating step in the protective innate immune-inflammatory response against injury, although it may also cause harm if left unchecked. The structurally related soluble complement inhibitors C4b-binding protein (C4BP) and factor H (FH) exert a tight regulation of the classical/lectin and alternative pathways of complement activation, respectively, attenuating the activity of the C3/C5 convertases and, consequently, avoiding serious damage to host tissues. We recently reported that the acute-phase C4BP isoform C4BP lacking the β-chain plays a pivotal role in the modulation of the adaptive immune responses. In this study, we demonstrate that FH acts in the early stages of monocyte to dendritic cell (DC) differentiation and is able to promote a distinctive tolerogenic and anti-inflammatory profile on monocyte-derived DCs (MoDCs) challenged by a proinflammatory stimulus. Accordingly, FH-treated and LPS-matured MoDCs are characterized by altered cytoarchitecture, resembling immature MoDCs, lower expression of the maturation marker CD83 and the costimulatory molecules CD40, CD80, and CD86, decreased production of key proinflammatory Th1-cytokines (IL-12, TNF-α, IFN-γ, IL-6, and IL-8), and preferential production of immunomodulatory mediators (IL-10 and TGF-β). Moreover, FH-treated MoDCs show low Ag uptake and, when challenged with LPS, display reduced CCR7 expression and chemotactic migration, impaired CD4+ T cell alloproliferation, inhibition of IFN-γ secretion by the allostimulated T cells, and, conversely, induction of CD4+CD127low/negativeCD25highFoxp3+ regulatory T cells. Thus, this novel noncanonical role of FH as an immunological brake able to directly affect the function of MoDCs in an inflammatory environment may exhibit therapeutic potential in hypersensitivity, transplantation, and autoimmunity.
The complement system is an ancestral innate immunity defense mechanism present in eukaryotic organisms not only in the fight against invading infectious agents, but also removing dangerous and apoptotic cells, debris, and particles from the body. Nevertheless, complement action is tightly regulated to avoid serious damage of self-tissue. In fact, uncontrolled or excessive complement activation generates toxic effector compounds that, when not properly regulated, trigger the development of a variety of inflammatory and autoimmune diseases (1, 2).
Factor H (FH), the main soluble regulator of the alternative pathway, is synthesized mainly in the liver and secreted into the bloodstream. This pathway plays a central role generating proinflammatory complement activation products in vivo (3). Like other members of the regulators of complement activation family, FH share 60-aa globular domains stabilized by two internal disulfide bonds termed complement control protein modules (CCPs), short consensus repeats, or Sushi domains. FH holds 20 homologous CCP domains arranged in tandem, conforming a 150-kDa single-chain glycoprotein that operates in the fluid phase and on host cellular surfaces. Thus, the region encompassing CCP1–4 of the protein (FH1–4) prevents the formation and promotes dissociation of the C3 convertase and, together with factor I, mediates proteolytic inactivation of C3b, an important opsonin and a major component of the convertase complexes that drive the complement activation cascade (4–6). Moreover, multiple mutagenesis, binding and functional reports, along with the last structural studies (7–10) show that the C-terminal recognition region formed by CCP19–20 of FH (FH19–20) binds to surface-bound C3b and can efficiently discriminate polyanionic host surfaces versus foreign surfaces and protect the former from complement attack. In fact, mutations in the C terminus of FH and binding of deficiency of FH-related plasma proteins and autoantibody-positive form of hemolytic uremic syndrome–associated autoantibodies defined the C-terminal recognition region as a hot spot associated with atypical hemolytic-uremic syndrome (aHUS), leading to disturbances in the physiologic interaction of FH with host endothelial cells (11). Conversely, many pathogens as well as some cancer cells are capable to hijack FH function to its own benefit to avoid complement-mediated lysis (12, 13).
Apart from or as a consequence of its essential role in complement regulation, other relevant functions dependent on interaction with endogenous ligands (pentraxins, extracellular matrix proteins, and DNA) or with receptors (integrins, malondialdehyde epitopes, and anionic phospholipids) have recently been suggested for FH such as modulation of local complement-mediated inflammation (14), protection from oxidative stress (15), modulation of platelet structure and function (16), anticoagulation (17), and regulation of cellular immune-inflammatory responses by interaction with a variety of cells from the immune system (18, 19).
Besides a CCP-based modular composition, FH also shares evolutionary and functional similarities with C4b-binding protein (C4BP), the main soluble inhibitor of the classical and lectin complement pathways (20). We recently reported that a minor isoform of C4BP (C4BPα7β0; C4BP lacking the β-chain [C4BP(β−)]), upregulated under proinflammatory conditions, induces a semimature and anti-inflammatory state in monocyte-derived dendritic cells (MoDCs) activated by proinflammatory stimuli. The resulting tolerogenic MoDCs were characterized by low surface expression of CD83, CD80, CD86, CD40, and CCR7, inability to release proinflammatory Th1 cytokines (IL-12, TNF-α, IFN-γ, IL-6, and IL-8), and, conversely, increased anti-inflammatory cytokine expression (IL-10 and TGF-β1). Moreover, C4BP(β−)–treated MoDCs failed to enhance allogeneic T cell proliferation and Th1 polarization and, instead, promoted regulatory T cell (Treg) generation (21).
Given the important role of FH in immune homeostasis, we speculated whether FH could also influence the differentiation/activation and response of MoDCs. In this study, we report that, analogously to C4BP(β−), FH is also able to induce an anti-inflammatory and tolerogenic state in MoDC activated by proinflammatory stimuli reminiscent to myeloid-derived suppressor cells (MDSCs), which seem to play an essential immunoregulatory role under different inflammatory conditions (22). Furthermore, a molecular dissection study indicated that the surface binding region of FH (FH19–20) is necessary to induce the tolerogenic phenotype in MoDCs. Given the fundamental role of DCs in immune surveillance and their capability to bridge innate and adaptive immunity, the emerging role of the soluble complement inhibitors C4BP(β−) and FH controlling inflammatory DC function may open new perspectives in pathological conditions characterized by overreactive immune responses such as autoimmunity and transplantation.
Materials and Methods
FH purified from human plasma (Biopur, Reinach, Switzerland) was used throughout the study. The 150-kDa protein was estimated to be >97% pure by SDS-PAGE. FH deletion constructs CCP1–7 (FHL1), CCP8–20 (FH8–20), CCP15–20 (FH15–20), and CCP19–20 (FH19–20) were expressed as histidine-tagged proteins in Sf9 insect cells using the pBSV-8His vaculovirus expression vector and purified by Ni2+-NTA agarose chromatography as previously described (4). The C4BP isoforms C4BP(β−) and C4BP containing the β-chain [C4BP(β+)] were also isolated from pooled human plasma, as previously described (21).
Cell culture and FH or C4BP treatment
RPMI 1640 was supplemented with 100 μg/ml streptomycin, 100 IU/ml penicillin, 2 mM l-glutamine (all from Invitrogen, Carlsbad, CA) and 10% heat-inactivated FBS (Linus, Cultek, Spain) (complete medium), unless otherwise stated.
PBMCs were obtained from buffy coat preparations collected from healthy donors from the Blood and Tissue Bank (Barcelona, Spain) after Ficoll-Paque density centrifugation (GE Healthcare Bio-Sciences, Uppsala, Sweden). For surface phenotype determination, monocytes were plated at 1 × 106 cells/ml in 60-mm culture plates (Corning), in RPMI 1640 medium without serum, and allowed to adhere for 2 h at 37°C in 5% CO2. The nonadherent cells were removed by washing in PBS. The final population contained >75% of monocytes, as demonstrated by flow cytometry of anti-CD14–stained isolates. For all functional assays, including the DC/T cell cocultures, monocytes were purified using colloidal superparamagnetic microbeads conjugated with monoclonal mouse anti-human CD14 Abs (MACS; Miltenyi Biotec, Auburn, CA; or EasySep Human CD14 Positive Selection Kit, Stemcell Technologies, Grenoble, France) and counted using Perfect Count microspheres (Cytognos SL, Salamanca, Spain). The purity of CD14+ cells was tested by CD14 staining and flow cytometry analysis (>90% CD14+ cells).
MoDCs were generated supplementing the monocyte cultures with complete RPMI 1640 medium plus GM-CSF (800 UI/ml) and IL-4 (500 UI/ml) (both from Gentaur, Kampenhout, Belgium) at days 0 and 3 of culture. At day 5, immature MoDCs (iDCs) were further stimulated for 24 or 48 h, depending on the assay, with either 5 μg/ml LPS (Escherichia coli 055.B5; Sigma-Aldrich, Copenhagen, Denmark) or for 48 h with 2 μg/ml recombinant human sCD40 Ligand/TRAP (ProSpec, Rehovot, Israel) to obtain mature MoDCs (mDCs).
CD3+ T cells were isolated from PBMCs by negative selection using EasySep Human T Cell Enrichment Kit (Stemcell Technologies). CD3+ T cells were >90% pure, as assessed by CD3 staining through flow cytometry.
Unless otherwise stated, either FH or C4BP(β+)/C4BP(β−) isoforms were added at 2, 5, or 10 μg/ml, at days 0, 3, and 5 (at the last time point, combined with LPS or CD40L) (differentiation and maturation). For control and comparison purposes, in some assays, we included a parallel, analogous treatment with the immunomodulator vitamin D3 (calcitriol, Calcijex; Abbott Laboratories) at 1 nM.
Inhibition of glycosaminoglycan (GAG) sulfation in MoDCs was achieved by adding sodium chlorate (Sigma-Aldrich) at the indicated concentrations concomitantly with FH along the DC differentiation and maturation stages.
Unfractionated sodium heparin (∼15 kDa; ≥180 IU/mg) (Sigma-Aldrich) at concentrations ranging from 100 to 1000 μg/ml was added either alone or concurrently with FH to assess its involvement on MoDC differentiation and maturation.
Blockade of CR3 and CR4 complement receptors was accomplished by 1-h preincubation of monocytes with 10 μg/ml either mouse anti-human CD11b (clone ICRF44) or mouse anti-human CD11c (clone 3.9) mAbs (both from Southern Biotechnology Associates, Birmingham, AL) in RPMI 1640 medium without serum, followed by 24-h coincubation with FH at 5 μg/ml in the above-described complete MoDC differentiation medium. A mouse IgG1κ (clone 15H6) was used as isotype control mAb.
Double staining using the fluorescent dyes Annexin V (Annexin V-PE Apoptosis Detection Kit I; BD Pharmingen, San Diego, CA) or 7-aminoactinomycin D (BD Pharmingen) and flow cytometry analysis were employed to assess the viability/apoptosis status of MoDCs.
Abs and flow cytometry
Cell-surface phenotypes were analyzed using the following mAbs: FITC-conjugated (anti–HLA-DR [Immu-357], anti-CD83 [HB15a], anti-CD14 [RMO52]), PE-conjugated (anti-CD40 [MAB89], anti-CD1a [BL6], anti-CD80 [MAB104], anti-CD86 [HA5.2B7]) (all from Beckman-Coulter), Alexa Fluor 488–conjugated anti-CCR7 (TG8/CCR7) (BioLegend, San Diego, CA), and PerCP-conjugated anti-CD3 (BD Pharmingen). The respective isotype controls were: FITC-conjugated (anti-IgG1 [4E02], anti-IgG2b [H2], anti-IgG2a [7T4-1F5]), PE-conjugated (anti-IgG1 [4E02] and anti-IgG2b [H2]), Alexa Fluor 488–conjugated anti-IgG2aκ (MOPC-173), and PerCP-conjugated anti-IgG1κ were from the same commercial sources. After washing with PBS, cells were subsequently stained with 3 μl mAb/1 × 105 cells in 100 μl FACS buffer (PBS containing 1% BSA and 0.1% sodium azide) for 20 min at room temperature. To exclude dead cells and debris, MoDCs were gated according to forward scatter and side scatter parameters. Stained cells were analyzed using an FACSCalibur (BD Biosciences, Franklin Lakes, NJ) equipped with CellQuestPro software (BD Biosciences).
Alternatively, all functional assays were evaluated using an FACScanto II flow cytometer equipped with FACSDiva software (BD Biosciences). Subsequent analyses used FlowJo software (Tree Star, Ashland, OR).
MoDCs (1 × 106/condition) were harvested at day 7, and the mRNA was extracted using the RNeasy RNA Isolation kit (Qiagen) and incubated with RNase-free DNase I (Ambion, Austin, TX) according to the manufacturer’s protocol. A two-step real-time RT-PCR technique was used to determine the relative mRNA levels of human CCR7, precursor of miR-155 (BIC-1), IDO (IDO), and mitochondrial superoxide dismutase 2 (SOD-2). Reverse-transcription reactions were performed with 500 ng total RNA using the Omniscript RT kit (Qiagen). Quantification of mRNA levels was performed by real-time PCR with the use of the LightCycler technology (Roche Molecular Biochemicals, Indianapolis, IN). The following primers were used: CCR7-forward (5′-TGGGCATCTGGATACTAGC-3′), CCR7-reverse (5′-AAGAAAGGGTTGACGCAGC-3′); IDO-forward (5′-GGTCATGGAGATGTCCGTAA-3′), IDO-reverse (5′-ACCAATAGAGAGACCAGGAAGAA-3′); BIC-1-forward (5′-AACCTACCAGAGACCTTACC-3′), BIC-1-reverse (5′-ATGCTTCTTTGTCATCCTCC-3′); SOD2-forward (5′-GACAAACCTCAGCCCTAAC-3′), SOD2-reverse (5′-ACACATCAATCCCCAGCAGT-3′), yielding products of 435, 227, 296, and 248 bp, respectively. These gene-specific primer pairs were designed using Oligo 4.0 and Primer 3 software (MBI, Cascade, CO) and selected to prevent primer-dimer formation.
All samples were normalized with the use of the following primer set for the constitutively expressed human cyclophilin A gene (CypA; PPIA): CypA-forward (5′-CTCCTTTGAGCTGTTTGCAG-3′) and CypA-reverse (5′-CACCACATGCTTGCCATCC-3′). All primers were purchased from Bonsai Technologies (Copenhagen, Denmark).
PCR amplifications were performed in a 20-μl volume containing 2 μl ready-to-use reaction mix, 10× DNA Master SYBR Green I (Roche Molecular Biochemicals); MgCl2 (3 mM for CCR7, 4 mM for BIC-1 and SOD-2, and 5 mM for IDO); 0.15 μM each primer; 5% DMSO; and 75 ng cDNA as template. The amplification program used an initial denaturation at 95°C for 10 min, followed by 45 cycles: 95°C for 1 s; 58°C (CCR7 and SOD-2)/60°C (BIC-1 and IDO) for 5 s; and 72°C for 10 s.
The relative levels of TGF-β1 were assessed using TaqMan technology. Briefly, total RNA was converted to single-stranded cDNA using the high-capacity cDNA archive kit according to the manufacturer’s instructions (Applied Biosystems, Carlsbad, CA). Each TaqMan gene expression assay contained a forward and reverse primer both for TGF-β1 (Hs00998133_M1) and PPIA (Hs99999904_M19) (Applied Biosystems) chosen for normalization. The real-time RT-PCR amplifications were run on an ABI Prism 7900Ht sequence Detection System (Applied Biosystems). Thermal cycling conditions were as follows: 2 min at 50°C; 10 min at 95°C; 40 cycles of denaturation at 95°C for 15 s; and annealing and extension at 60°C for 1 min.
The reproducibility of the assays was verified, and the expression of the five genes was shown to be within the linear range at the chosen cell concentration.
Scanning electron microscopy
Monocytes were seeded on glass slides covered with either poly-l-lysine (25 μg/ml) or fibronectin (42 μg/ml), differentiated for 5 d in complete RPMI 1640 medium supplemented with 800 U/ml GM-CSF, 500 U/ml IL-4, in the presence or absence of FH, and further stimulated with LPS for 48 h in the same medium as previously described. The resulting MoDCs were fixed in 1% paraformaldehyde and 1.25% glutaraldehyde in cacodylate buffer for 2 h. Finally, the cells were postfixed in 1% osmium tetroxide and dehydrated with graded series of ethanol followed by acetone. After dehydration, the cells were dried in a critical-point dryer and coated with gold before observation by scanning electron microscopy (Zeiss DSM940A; Carl Zeiss). To quantify the percentage of MoDCs (iDC-like [few short dendrites, round shape] or mDC-like [numerous long dendrites, elongated shape]) under a given condition, 100 cell profiles were recorded by a blinded scorer using random scans at a fixed magnification (×1750).
Endocytosis of iDCs and mDCs was assessed using DQ-OVA (1 mg/ml, DQ-OVA; Molecular Probes, Leiden, the Netherlands) and Lucifer Yellow CH (10 mg/ml; Sigma-Aldrich) as previously described (21). Briefly, 2 × 105 cells/ml were resuspended in 100 μl PBS and incubated either with 4 μl BODIPY FL-conjugated DQ-OVA at 37°C or at 0°C for 15 min (receptor-dependent endocytosis) or with 6 μl Lucifer Yellow CH at 37°C or at 0°C for 2 h (fluid-phase endocytosis). The incubations were stopped by adding 1 ml cold FACS buffer. The cells were washed two times with cold FACS buffer and their fluorescence analyzed using flow cytometry.
MoDC cytokine secretion
Concentrations of IL-12p70, TNF-α, IFN-γ, IL-10, IL-6, and IL-8 were determined in supernatants from MoDCs, either untreated or treated with FH, using the Th1/Th2 Flow cytomix Multiplex kit (Bender-MedSystems, Vienna, Austria) according to the manufacturer’s instructions. Alternatively, the presence of IL-12 in several functional assays, or the IL-10/IL-12 ratio from untreated, FH-, FHL1-, and FH19–20-treated MoDCs were assessed through IL-10– and/or IL-12p70–specific ELISAs (Diaclone, Besancon Cedex, France).
MoDCs differentiated and matured (LPS for 48 h) in the presence or in the absence of FH were tested for migration toward the CCL21 chemokine using transwell assays. Briefly, the lower chambers of transwell plates (polycarbonate filters of 8.0-μm pore size; Costar, Corning, NY) were filled with 400 μl complete RPMI 1640 medium with or without CCL21 (200 ng/ml). A total of 1 × 105 DCs in 100 μl complete RPMI 1640 medium were added into the upper chamber, and cells were incubated at 37°C for 2 h. We verified the absence of filter-associated MoDCs by removing nonmigratory cells from the upper chamber and upper side of the membrane and examining the underside by crystal violet staining. Cells migrated into the lower chambers were harvested and counted with a flow cytometer, acquiring events for a fixed time period of 2 min. The migration assays for all stimulation conditions were performed in triplicate wells. Values are given as percentage of migrated cells relative to the untreated mDCs (100%).
Allogeneic CD3+ T cells (1 × 105/well) were stimulated in vitro with FH-treated, or vitamin D3-treated, and LPS-activated MoDCs in a 96-well round-bottom plate at various DC/T cell ratios (1:40, 1:80, and 1:160) and cultured in X-VIVO 15 medium (Biowhittaker, Walkersville, MD) supplemented with 2% human AB serum, 100 μg/ml streptomycin, 100 IU/ml penicillin, and 2 mmol l-glutamine. Alloproliferation was measured after 5 d of incubation. At day 4, the cells were pulsed with [3H]thymidine (1 μCi/well; Amersham, Freiburg, Germany), followed by incubation for another 16 h. Labeled cells were then harvested onto glass-fiber filters (Harvester 96; TomTec), and the T cell proliferation rate was determined by the amount of [3H]thymidine incorporation using a scintillation counter (1450 MicroBeta Trilux plate reader; Wallac, Turku, Finland). Results are reported as the mean cpm ± SD of thymidine incorporation in quadruplicate culture wells.
Intracellular cytokine staining
T cells were labeled with the intracellular fluorescent dye CFSE (Molecular Probes), plated in 96-well round-bottom plates (1 × 105 cells/well) and stimulated for 5 d with allogeneic MoDCs (1:40 DC/T cell ratio). Cell proliferation was determined by the sequential loss of CFSE fluorescence of cells, as detected by flow cytometry. Then, the cells were stimulated with 50 ng/ml PMA plus 500 ng/ml ionomycin for 5 h in the presence of 10 μg/ml brefeldin A (all from Sigma-Aldrich). After stimulation, cells were washed with PBS, fixed, permeabilized using an IntraStain kit (DakoCytomation, Glostrup, Denmark), and incubated for 30 min at room temperature with anti-human IFN-γ allophycocyanin-conjugated mAb (Miltenyi Biotec). Cells were washed and analyzed by flow cytometry.
Determination of CD4+CD127 low/negativeCD25high and Foxp3+ T cells
Negatively selected CD3+ T lymphocytes were plated (1 × 105 cells/well) in 96-well round-bottom plates and cocultured at a 1:40 DC/T cell ratio. After 6 d of coculture without restimulation and any supplemental cytokines, we used flow cytometry to determine the percentage of Tregs defined as CD4+, CD127 low/negative, CD25high, and intracellular Foxp3+ (Human Regulatory T Cell Staining Kit; eBioscience, San Diego, CA).
Unless otherwise indicated, three technical replicates were performed from each independent experiment. Results are presented as means ± SD. DC variables under different experimental conditions respect to a reference condition (usually mDCs or iDCs) were compared using the ANOVA and Bonferroni correction for multiple comparisons, considering p < 0.05 as significant.
FH modifies the morphology and downregulates the phenotypic maturation of MoDCs
We first evaluated the impact of FH in the differentiation of MoDCs by assessing the expression of different monocyte and DC surface markers, including CD14, HLA-DR, CD40, CD80, CD83, CD86, and CD1a. When added concomitantly with the DC differentiation factors GM-CSF and IL-4, FH did not significantly modify the expression of the above surface markers at the stage of iDCs (day 5 after incubation) (data not shown). Nevertheless, upon LPS stimulation, FH-treated MoDCs significantly prevented the upregulation of the costimulatory molecules CD40, CD80, and CD86 and of the maturation marker CD83 in a dose-dependent manner compared with untreated mDCs (Figs. 1A, 2). Conversely, expression of HLA-DR, CD14, and CD1a was not significantly altered by FH treatment (Fig. 2). A side-by-side comparison of the effect of FH and the C4BP(β−) isoform over MoDCs revealed an analogous outcome of both soluble complement inhibitors in terms of surface marker expression (Supplemental Fig. 1). This distinctive cell-surface marker pattern induced by FH was essentially maintained when the MoDCs were alternatively matured using CD40L as proinflammatory stimulus (Supplemental Fig. 2).
In addition, although there were no significant scatter modifications between iDCs and mDCs, as previously reported (23, 24), FH-conditioned MoDCs produced a shift toward decreased scatter compared with mDCs (Fig. 1A), indicating that FH had a direct effect on MoDC morphology. A more comprehensive analysis of the MoDC shape by scanning electron microscopy revealed that, before LPS exposure, 95% of the untreated iDCs were essentially round with few short projections, whereas after 48 h of LPS-mediated maturation the dendritic morphology became evident in 80% of the analyzed cells, with numerous long cytoplasmic projections. Conversely, MoDC treatment with FH fully prevented the porcupine-like DC morphology resulting upon LPS induction (72% of the analyzed cells had iDC appearance) (Fig. 1B).
The yield of the FH-treated and LPS-matured MoDCs obtained (>85%) was slightly although significantly reduced compared with mDCs. Nevertheless, these cells remained viable throughout the differentiation/maturation process, as assessed by Annexin V/7-aminoactinomycin D staining, with <17% of apoptotic cells evidenced at 48 h after LPS-mediated MoDC maturation. Conversely, an analogous treatment with the reference immunomodulator vitamin D3 significantly reduced MoDC viability and yield (Supplemental Fig. 3).
C4BP and FH hold unusually stable activities regarding complement inhibition (25). Additionally, FH needed to be exposed to harsh denaturing conditions (98°C; 30 min.) to hamper its capacity to confer a semimature phenotype on MoDCs (Fig. 3).
Together, these data are evidence that FH, analogously to the C4BP(β−) isoform, has the potential to modify proinflammatory MoDC differentiation and/or maturation as judged by changes in both their morphology and the expression pattern of various DC surface markers.
FH affects the expression profile of LPS-induced MoDCs
We next assessed the effect of FH on relevant transcripts shaping the molecular signature of LPS-matured MoDCs. Thus, we analyzed by quantitative RT-PCR (RT-qPCR) the expression of the genes coding for the immunomodulatory factor TGF-β1, involved in Treg generation, the immunoregulatory enzyme IDO, involved in tryptophan metabolism, the mitochondrial antioxidant enzyme SOD-2, and miR-155 (BIC-1), an important miRNA involved in immune function and part of a negative-feedback loop downmodulating inflammatory cytokine production in response to LPS stimulation. As reported (26), expression of IDO, SOD-2, and BIC-1 were upregulated, whereas TGF-β1 was downregulated, upon LPS-mediated MoDC maturation. In contrast, FH-conditioned MoDCs suppressed the subsequent induction of IDO and BIC-1 by LPS, while maintaining TGF-β1 expression, which is normally reduced by LPS (Fig. 4). Thus, all four molecular biomarkers analyzed reached transcript levels comparable to those from iDCs.
FH downregulates the endocytic activity of MoDCs
The Ag internalization capacity of MoDCs was assessed at 37°C by flow cytometry of both self-quenching DQ-OVA (mannose receptor-mediated endocytosis marker) and Lucifer Yellow CH (macropinocytosis marker). The high endocytic activity of iDCs was significantly reduced by treatment with FH (Fig. 5A). Uptake of both DQ-OVA and Lucifer Yellow CH was further inhibited by incubation at 0°C (gray histograms), confirming that the fluorescence shifts from both dyes under different conditions were due to active internalization rather than nonspecific binding. Likewise, FH-treated MoDCs behaved analogously to untreated mDCs after LPS-induction, showing a reduced endocytic capacity for soluble proteins (Fig. 5B). These data indicate that FH alter the endocytic capacity of MoDCs.
FH inhibits the release of inflammatory cytokines by LPS-induced MoDCs
We next assessed whether the effect of FH on the MoDC phenotype was accompanied by changes in the release of cytokines (IL-12p70, IL-10, IL-8, IL-6, TNF-α, and IFN-γ). Compared to untreated iDCs, secretion of each of the proinflammatory cytokines was upregulated when iDCs were matured with LPS. In contrast, pretreatment with FH prevented the release of IL-12p70 and significantly decreased the secretion of TNF-α, IFN-γ, IL-8, and IL-6 and, conversely, increased the production of the anti-inflammatory cytokine IL-10 (Fig. 6). Thus, Th1 proinflammatory cytokine production upon LPS-mediated MoDC stimulation was significantly diminished in FH-treated MoDCs.
FH alters the chemotaxis of MoDCs
The expression of the chemokine receptor CCR7 in DCs is inhibited by TGF-β1 (27). Accordingly, FH treatment significantly suppressed the induction of CCR7 at both the transcriptional and the translational level upon LPS-mediated activation of MoDCs (Fig. 7A, 7B). Reduced surface CCR7 expression, in turn, significantly decreased the migration of LPS-activated MoDCs toward the chemokine CCL21 (Fig. 7C). In contrast, LPS-mediated activation of untreated MoDCs (mDCs) induced maximal migration in response to CCL21.
MoDCs exposed to FH fail to stimulate allogeneic T cell proliferation
Given that FH was found to impact on phenotypic maturation and the amount of inflammatory cytokines released by MoDCs, we next examined the T cell immunostimulatory capacity of MoDCs exposed to FH at different MoDC/T cell ratios. When untreated MoDCs were matured with LPS, maximal allogeneic T cell proliferation was observed. In contrast, mature MoDCs preincubated with FH failed to enhance CD3+ T cell proliferation, approaching the levels observed using iDCs (Fig. 8).
MoDCs exposed to FH elude differentiation toward a Th1-polarizing phenotype and, conversely, promote adaptive Treg generation from allogeneic T cells
To assess the functional outcome of T cells allostimulated with FH-treated MoDCs, the resulting T lymphocytes were restimulated with PMA plus ionomycin and IFN-γ production was measured by intracellular staining. A significant reduction in IFN-γ production (∼50%) was evident from T cells exposed to FH-treated and LPS-matured MoDCs relative to T cells exposed to mDCs, approaching the IFN-γ levels achieved by T cells allostimulated with iDCs (Fig. 9A, top panel, Fig. 9B). When only the CSFElow (proliferating) T cell population was analyzed, FH-conditioned MoDC-allostimulated T cells showed again a significant decrease in IFN-γ production relative to untreated MoDC-allostimulated T cells (Fig. 9A, bottom panel, Fig. 9C).
Finally, we further investigated the presence of adaptive Tregs, defined as CD4+CD127low/negativeCD25highFoxp3+, under the above MoDC/T cell coculture conditions. Blast T cells exposed to FH-treated MoDCs showed a significant increase in the percentage of CD4+CD127low/negativeCD25highFoxp3+ cells, analogously to that achieved by blast T cells allostimulated with iDCs (Fig. 10). Taken together, the above results support the FH-mediated induction of a tolerogenic, anti-inflammatory phenotype in MoDCs.
The C-terminal binding region (CCP19–20) of FH determines its tolerogenic activity over MoDCs
Given the modular composition of FH, including 20 CCP domains arranged linearly, and the functional involvement of the CCP domains in the inhibition of the alternative complement pathway, with CCP1–4 encompassing the complement regulatory region and CCP18–20 engaging the central surface binding region, we aimed to dissect the major region(s) responsible for the immunomodulatory or tolerogenic activity of FH over MoDCs. Thus, recombinant FHL1 (including CCP1–7) and truncated variants FH8–20 (comprising CCP8–20), FH15–20 (comprising CCP15–20), and FH19–20 (comprising CCP19–20) (Fig. 11A) were tested for their ability to modify the activation phenotype of LPS-induced MoDCs. Consequently, FHL1 conditioning of MoDCs did affect neither upregulation of the CD83 maturation marker nor skewing of cytokine secretion toward a proinflammatory profile (low IL-10/IL-12 ratio) induced by LPS (Fig. 11B–D). Hence, FHL1 lacked the immunomodulatory activity characteristic of full-length FH. In contrast, all progressive FH deletion variants—FH8–20, FH15–20, and FH19–20—significantly blocked the upregulation of the CD83 maturation marker (Fig. 11B, 11C) and increased the IL-10/IL-12 ratio toward an anti-inflammatory phenotype upon LPS induction (Fig. 11D). Thus, the surface-binding region of FH (domains CCP19–20) appears necessary for the immunomodulatory activity of FH in MoDCs.
FH acts in the early stages of monocyte differentiation toward MoDCs
To delineate which stage of the differentiation and/or maturation process of MoDC was affected by the tolerogenic activity of FH, we performed several time-course experiments. We first added FH consecutively to each of the 5-d differentiation process from monocytes to iDCs mediated by IL-4/GM-CSF, followed by LPS-mediated MoDC maturation, and assessed their phenotypic outcome (Fig. 12A). Both CD83/CD86 surface expression and IL-12 production were found significantly reduced after LPS-mediated MoDC maturation only when FH was added in the early monocyte to iDC differentiation process. FH added from day 2 of differentiation onwards did not affect the normal differentiation and maturation phenotype of MoDCs. An analogous inverse experiment in which we consecutively removed FH from initially exposed monocytes to each of the 5-d monocyte to iDC differentiation process, followed by LPS-mediated MoDC maturation (Fig. 12A, bottom panels), confirmed that 24-h FH exposure at the initial stage of monocyte to iDC differentiation was necessary and sufficient to confer an anti-inflammatory and tolerogenic phenotype to MoDCs.
An equivalent experiment was additionally performed to evaluate the outcome of FH exposure only at the level of mDCs. Analysis of CD83/CD86 surface marker expression and IL-12 production evidenced the lack of FH anti-inflammatory and tolerogenic activity when acting over LPS-matured MoDCs (Fig. 12B).
FH immunomodulatory activity is mediated neither by CR3 or CR4 complement receptors nor by surface GAG interaction
To gain further insight about the molecular mechanism mediating the tolerogenic activity of FH over MoDCs, we focused on the potential role of the highly abundant GAGs and of complement receptors CR3 (CD11b/CD18) and CR4 (CD11c/CD18) present on the monocyte/MoDC surface. It has been recently shown that both CR3 and CR4 function as FH receptors on monocyte-derived macrophages (28). Thus, we exposed monocytes to blocking mAbs against integrins CD11b and CD11c, followed by 24-h coincubation with FH at the monocyte to iDC early differentiation stage. Blockade neither of CD11b nor of CD11c was able to abolish the activity of FH on LPS-matured MoDCs in terms of surface CD83/CD86 expression and IL-12 production, suggesting that the corresponding complement receptors do not affect the FH-induced tolerogenic phenotype in MoDCs (Supplemental Fig. 4A).
Tissue-specific host recognition by FH is mediated by differential activities of its two polyanionic GAG-binding regions (29). Thus, we first evaluated whether soluble high m.w. heparin (a highly sulfated form of the polyanion heparan sulfate) might be able to alter the FH-mediated tolerogenic phenotype in MoDCs. Heparin alone had no apparent effect in MoDC differentiation and/or maturation, whereas coincubation of FH and increased heparin concentrations failed to reverse the tolerogenic activity of the former regarding CD83/CD86 surface expression and IL-12 secretion (Supplemental Fig. 4B). An analogous result was obtained using bemiparin, a low m.w. form of heparin (data not shown). Moreover, exposure of the monocyte/MoDC culture to nontoxic concentrations of sodium chlorate, which competitively inhibits the formation of 3′-phosphoadenosine 5′-phosphosulfate, preventing GAG sulfation, was unable to affect the FH-induced MoDC phenotype regarding surface CD83/CD86 expression and IL-12 production (Supplemental Fig. 4C). In contrast, inhibition of this high-energy sulfate donor by sodium chlorate efficiently suppressed the expression of the IL-4–induced DC marker CD1a (30) (Supplemental Fig. 4C, bottom left panel). Furthermore, surface GAG digestion with heparinase III and chondroitinase ABC was again incapable to reverse the FH-induced tolerogenic phenotype in MoDCs (data not shown). Taken together, the above data indicate that cell-surface GAG do not mediate the immunomodulatory actions of FH over MoDCs.
The complement system constitutes a key arm of the innate immune response in the fight against pathogen infection. To maintain immune homeostasis, the main fluid-phase complement inhibitors C4BP and FH share complementary activities in the regulation of excessive complement activation and offer protection against its deleterious consequences into the host by modulating inflammation and the development of adaptive immune responses and by disposal of cellular debris and apoptotic cells (31). Interestingly, we have recently shown that the α7β0 isoform of C4BP [C4BP(β−)] was able to directly modulate the biological behavior of MoDCs toward a semimature, anti-inflammatory, and tolerogenic phenotype (21). Although FH has been suggested to be involved in adhesion, phagocytosis, and modulation of cell activation through binding to specific receptors in host cells (18), it remained unknown whether FH could also directly affect the function of MoDCs.
In this study, we demonstrate for the first time, to our knowledge, that FH, the major fluid-phase regulator of the central protein C3b in the alternative pathway of complement activation, is also able to induce an anti-inflammatory and tolerogenic phenotype in MoDCs activated by proinflammatory stimuli, a model of inflammatory DCs able to initiate Th1 cell differentiation (32–34). Thus, analogously to that previously reported for C4BP(β−), FH skewed the LPS-induced maturation of MoDCs and induced new morphological and functional features reminiscent of the monocytic subtype of MDSCs (22), resulting in a reduced capacity to activate T cells. In fact, although nontoxic, FH conditioning of MoDCs prevented upregulation of costimulatory molecules, such as members of the B7 family (CD80 and CD86) and CD40, and the maturation marker CD83 upon LPS engagement of TLR4, further precluding the NF-κB–mediated polarization of the T cell response toward Th1.
It has also been reported that LPS activation is essential for inducing migratory and Ag-presenting activity in tolerogenic DCs (35). In contrast to other immunomodulatory agents such as IL-10, dexamethasone, or vitamin D3 (36), surface HLA-DR expression was not reduced by FH treatment, similarly to that previously observed for C4BP(β−), suggesting that the Ag-presentation capability remains viable in FH-conditioned MoDCs. Conversely, C4BP(β−) and FH conditioning diverged regarding their capability to influence Ag uptake in MoDCs. Whereas C4BP(β−) treatment did not affect the Ag internalization capacity of immature MoDCs (21), FH treatment significantly reduced their endocytic activity. Moreover, FH conditioning of MoDCs activated by LPS triggered the downregulation of the CCR7 surface receptor and, consequently, significantly reduced their migration in response to the CCL21 ligand. Accordingly, it has been shown that CCR7 controls the cytoarchitecture, rate of the endocytosis, survival, migratory speed, and maturation of DCs (37). Nevertheless, the reduction of surface CCR7 and endocytic activity of FH-conditioned MoDCs should not compromise their immunomodulatory activity. For instance, the well-characterized immunoregulatory neuropeptide vasoactive intestinal peptide, which has shown therapeutic potential in autoimmune disorders, also causes a decrease in the expression of the costimulatory molecules CD80 and CD86, induce significant production of anti-inflammatory cytokines such as IL-10, and inhibits the phagocytic activity of DCs (38, 39).
Furthermore, analysis of selected transcripts and secreted cytokines involved in DC biology confirmed that both C4BP(β−) (21) and FH were able of reverting LPS-induced proinflammatory Th1 polarization of MoDCs (showing reduced expression of IL-12, TNF-α, IFN-γ, IL-6, and IL-8) and, instead, promoted the development of an anti-inflammatory profile (yielding increased expression of TGF-β1 and IL-10). It is well known that IL-10 blocks IL-12 synthesis by DCs, downregulates the expression of costimulatory molecules, and potentiates their tolerogenicity (40). Moreover, IL-10 and TGF-β1 synergistically generate tolerogenic DCs capable of inducing anergy in Ag-specific memory CD4+ T cells and differentiating them into IL-10–producing Tregs (41). Concordantly, we have demonstrated that the cross-talk between allogeneic T cells and FH-conditioned MoDCs: 1) induce T cell hyporesponsiveness and prevent Th1 polarization, characterized by a significant reduction of IFN-γ secretion; and 2) promote the generation of CD4+CD127low/negativeCD25highFoxp3+ cells (adaptive Tregs).
From the foregoing, it seems evident that there might be many similarities in the responses induced by C4BP(β−) and FH over MoDCs, although these responses are not identical, in agreement with recent reports comparing different immunosuppressive agents (42). Our ongoing comparative analysis of genome-wide transcriptional profiles from LPS-stimulated MoDCs conditioned by both fluid-phase complement inhibitors will allow the comprehensive assessment of key common and differential genes and signaling pathways involved in the development of the tolerogenic phenotype.
In addition, the physiological significance of the newly described immunomodulatory role of FH is intriguing. Mutations in the C-terminal region (host ligand binding region) of FH predispose individuals to aHUS and other microangiopathies (11). Through a functional analysis employing rFH variants covering the whole molecule, we have shown that CCP19–20, which binds to C3b and polyanions such as glycosaminoglycans (GAGs) on host cells to mediate cell-surface protection, also has a major role in the tolerogenic activity of FH over MoDCs. These modules are thereby critical for attachment of the protein in the specific context to self-surface. This binding profile is further exploited by many microbial pathogens that similarly bind FH via the C-terminal CCP19–20 (43). The use of recombinant FHL1 and deletion mutants shows that the immunomodulatory activity is specific for FH and is not present in FHL1, the related human plasma protein, which shares the N-terminal but lacks the C-terminal CCP domains and thereby the main recognition region. Interestingly, a recent report has shown that the CCP6–8 and CCP19–20 regions of FH differentially mediate binding in different host tissues. CCP19–20 binds poorly to the GAG heparin compared with CCP6–8, which might imply additional binding of the C terminus of FH to an as-yet-unknown receptor responsible for the immunomodulatory activity of FH (29). Indeed, FH has been suggested to bind specific adhesion molecules in immune cells such as the integrins CR3, CR4, and αIIbβ3 or l-selectin (18). Moreover, regarding the most thoroughly characterized interaction, FH–CR3, it has been shown that CCP19–20 holds a major binding site for CR3, a regulator of human DC immunostimulatory function (44, 45). However, according to our results, neither CR3 nor CR4 seem to be involved in the noncanonical immunomodulatory activity of FH over MoDCs. Alternatively, tissue-specific host recognition by FH has been shown to be mediated by its two (CCP6–8 and CCP19–20) GAG-binding regions (29). Nevertheless, we have shown in this study that the tolerogenic activity displayed by FH in the presence of MoDCs does not involve GAGs. Additionally, in a physiological setting, we cannot discard the influence of other known FH ligand proteins such as SIBLING proteins, adrenomedullin, pentraxins (46), or malondialdehyde modified protein adducts (47), the last two showing also CCP19–20 binding, in the overall immunomodulatory actions of FH. Further studies are underway to identify the monocyte/DC surface receptor able to interact with FH, either directly or indirectly, and induce a tolerogenic phenotype in MoDCs.
The above evidence argues for a tissue-specific recognition and local immunomodulatory action of FH in the sites of inflammation or injury. Indeed, inflammatory DCs differentiate in situ from monocytes recruited to the sites of inflammation (48). Thus, FH might represent an immunomodulatory differentiation factor targeting inflammatory MoDCs. Moreover, the source of local FH might be other than that released by the liver into the circulation. To note, attempts at liver transplantation in patients with FH mutations have been unsuccessful (49). Indeed, FH has been shown to be released by different cell types such as monocytes, fibroblasts, endothelial cells, keratinocytes, mesenchymal stem cells (MSCs), and retinal pigment epithelial cells (18). Thus, locally released FH in tissues (approaching concentrations of 2–10 μg/ml, according to our in vitro results) may help maintaining an anti-inflammatory environment and to modulate the adaptive immune response by inducing tolerogenic DCs. For example, it has been recently reported that MSCs, which inhibit DC differentiation/maturation (50), constitutively secrete FH. In addition, production of FH by MSCs is augmented by inflammatory cytokines TNF-α and IFN-γ in a dose- and time-dependent manner (51). Thus, locally produced FH in the MSC niche environment outside the vasculature could reach sufficient local concentrations to be of physiological significance, contributing to MSCs potent and broad anti-inflammatory and immunosuppressive activities.
Regarding pathological scenarios, the C-terminal region of FH is a hotspot for disease-associated mutations, which have been linked to increased risk for the development of aHUS, early onset drusen, and age-related macular degeneration (AMD) (52–54). Moreover, it has been recently reported that the treatment with human FH rapidly reverses the renal abnormality developed in FH-deficient mice (55). Our findings embrace the tantalizing possibility that the noncanonical immunomodulatory activity of FH over DCs, a critical link between innate and adaptive immunity because their major role as APCs able to induce potent T cell–mediated immune responses, could contribute to the overall outcome not only of FH-related diseases (aHUS and AMD) but also of other immune-inflammatory diseases. For instance, regarding AMD, there remains enough clinical and experimental support of the concept of immune engagement that goes far beyond the role of FH as complement inhibitor. Indeed, evidence supporting the adaptive immune system is directly implicated in both mouse and human (56). Thus, several FH mutations affecting its immunomodulatory activity might lead to loss of local control of many aspects of the immune system, and the release of inflammatory factors would lead to injury and loss of retinal pigment epithelium cells. Moreover, it has been reported that alternative pathway complement activation has a critical role in asthma, and strategies to deliver the complement regulatory region of FH specifically to the site of inflammation provide greater protection than that afforded by endogenous regulators (57). Thus, additional protection conferred by FH immunomodulation over inflammatory DCs might have been underestimated, because DCs are instrumental in the development of both regulatory tone and mucosal immunity in asthma (58). Furthermore, it has been recently shown that MDSCs are implicated in resistance to streptozotocin-induced diabetes in the absence of complement C3 (59), arguing for a potential role of FH-mediated immunomodulation.
Finally, it remains to be established whether FH itself or FH-conditioned MoDCs hold promise for future DC-based tolerance-inducing strategies and how will they perform compared with tolerogenic DCs induced by other mediators (60). Additional therapeutic approaches using animal models of autoimmunity and/or transplantation will be needed to ascertain whether the phenotypic and functional signature of FH might be particularly suited to prime Ag- and tissue-specific Tregs for clinical translation, targeting a specific immunoinflammatory process (61).
It will be also interesting to decipher whether the FH-related protein family members (FHR1–5) (62), of which functions are not fully understood, also hold immunomodulatory activity over DCs or interfere with FH function.
In summary, we have identified a novel immunomodulatory activity of FH, unrelated to complement regulation, able to influence adaptive immunity by directly inducing a tolerogenic phenotype on inflammatory DCs. Challenges ahead include the identification of the putative FH surface receptor responsible for the immunomodulatory activity, the assessment of the molecular mechanism involved in FH-conditioned DC transformation toward a tolerogenic state, and the physiological relevance of this FH-mediated immunomodulatory activity in health and disease. Notwithstanding, our findings could open new avenues for the evaluation of the therapeutic potential of FH or FH-related peptides, either by direct administration or by adoptive transfer of ex vivo-conditioned DCs, in the modulation of dysregulated immune-inflammatory processes such as hypersensitivity, autoimmune pathologies, and transplant rejection.
We thank Frida Mohlin (Lund University), Esther Castaño (Cytometry Service, Serveis Cientificotècnics de la Universitat de Barcelona), and Núria Cortadellas (Scanning Electron Microscopy Service, Serveis Cientificotècnics de la Universitat de Barcelona) for invaluable technical assistance.
This work was supported by the Ministerio de Economía y Competitividad (Madrid, Spain) through ISCIII Grants PI10/1073 and PI13/01490, and cofunded by FEDER funds/European Regional Development Fund (ERDF)-a way to build Europe, Grant 12/1210 from La Marató de TV3 Foundation, and Grant 2014SGR541 (Generalitat de Catalunya) to J.M.A. J.M.A. is sponsored by the Researchers Consolidation Program from the Sistema Nacional de Salut-Departament de Salut Generalitat de Catalunya (CES06/012). S.R.d.C. is supported by the Ministerio de Economia y Competitividad (SAF2011-26583), the Ciber de Enfermedades Raras, and the Fundación Renal Iñigo Alvarez de Toledo. A.M.B. is supported by the Swedish Research Council (K2009-68X-14928-06-3) and the Swedish Cancer Foundation. F.E.B. is supported by Grant 2014SGR804 (Generalitat de Catalunya) to support emerging research groups.
The online version of this article contains supplemental material.
Abbreviations used in this article:
atypical hemolytic-uremic syndrome
age-related macular degeneration
C4BP containing the β-chain
C4BP lacking the β-chain
complement control protein module
myeloid-derived suppressor cell
monocyte-derived dendritic cell
mesenchymal stem cell
superoxide dismutase 2
regulatory T cell.
The authors have no financial conflicts of interest.