IκBζ Is a Transcriptional Key Regulator of CCL2/MCP-1 Free

Resident tissue macrophages (Mϕ) make up the first defense line against infiltrating pathogens. In these cells, pathogen recognition triggers a rapid and extensive cascade of physiological responses, including, besides others, the direct intruder abatement and release of proinflammatory cytokines that initiate a systemic immune response (1–3).

Among the most prominent cytokines, CCL2 (MCP-1), the first discovered member of the CC family of chemokines, mediates the chemotactic recruitment of circulating monocytes to inflammatory sites (4–7). Even though further signal molecules, such as CCL7 and CCL13, exhibit overlapping functions, CCL2 seems to be the only nonredundant monocyte-specific chemokine (3, 5, 8). Its pivotal role is emphasized by observations in many (auto)inflammatory disorders, such as rheumatoid arthritis, multiple sclerosis, atherosclerosis, glomerulonephritis, or inflammatory bowel disease, where increased CCL2 levels have been linked to the onset or progression of disease (3, 8, 9). Strikingly, disruption of the Ccl2 gene in mice causes reduced susceptibility for inflammatory diseases (4, 5, 10, 11). As a consequence, almost all major pharmaceutical companies aim at developing antagonistic small molecules targeting the most relevant CCL2 receptor CCR2 for the therapy of chronic inflammation (12, 13).

Under noninflammatory conditions, the Ccl2 locus is transcribed at low level, but rapid induction of gene expression occurs on exposure of cells to various proinflammatory stimuli (14, 15). The differential Ccl2 gene expression is controlled at the promoter level by a complex network of transcriptional regulators, including NF-κB, C/EBP, AP-1, and Sp-1 (16). Despite such complexity, inflammation-linked, high-level expression appears to be dominantly regulated by NF-κB proteins (16–19). Recent research has shed new light on the mechanisms underlying transcriptional activation by NF-κB, which can regulate hundreds of different target genes. Current efforts focus on understanding of how only certain subsets of NF-κB target genes are selectively induced depending on the cell type, a given stimulus, and environmental parameters (20–22). Such insights may open up novel therapeutic strategies aiming to specifically modulate NF-κB responses in chronic inflammation (20).

Despite the presence of high-affinity binding sites, only a fraction of NF-κB target genes is generally activated in response to an inflammatory stimulus. It was suggested that inflammatory genes can be categorized in two groups, based on their kinetics of induction and the requirement of protein synthesis. Whereas primary response genes are rapidly induced, the expression of secondary target genes is delayed and requires the prior synthesis of additional NF-κB coregulators. In this context, IκBζ, an atypical IκB protein, has been recently identified and implicated in differential NF-κB target gene expression in Mϕ (23–25), even though its physiological function remains largely unknown. The IκBζ-coding Nfkbiz gene is rapidly induced as a primary NF-κB response gene by various inflammatory stimuli. Upon de novo synthesis, protein interactions with p50 NF-κB homodimers seem to mediate the recruitment of IκBζ to target promoters (23, 26). IκBζ is believed to act as a transcriptional coactivator required for the expression of a subset of NF-κB secondary response genes, presumably by facilitating transcription-enhancing nucleosome remodeling (21, 27).

In this article, we identify the murine Ccl2 gene as a novel target of IκBζ-regulated gene expression and show that IκBζ is critical for CCL2 secretion in Mϕ. Furthermore, in vivo evidence in a model of experimental peritonitis suggests that this regulatory mechanism is relevant in systemic inflammation and necessary for the proper recruitment of monocytes to sites of inflammation.

Nfkbiz ^ko/ko (28) and littermate control mice were used at 6–18 wk of age. Mouse work was conducted in accordance with the German law guidelines of animal care, as permitted by regional authorities (“Regierungspräsidium Tübingen,” application no. H6/12).

Female C57BL/6 mice were euthanized by CO₂ asphyxiation. For isolation of peritoneal Mϕ (PMϕ), the abdominal skin was removed, a catheter (24 gauge) was inserted into the peritoneal cavity, and 10 ml ice-cold PBS was injected. After massage of the peritoneum, peritoneal fluids were aspirated, centrifuged at 500 × g for 5 min, and yielded PMϕ were resuspended in 0.5–1.5 ml Mϕ medium containing DMEM/Hams F12, 10% FCS, and MycoZapPlusCL antibiotics (Lonza). Next, cells were seeded in 96-well plates (100 μl cell suspension/well) and cultivated at 37°C for 2 h. Finally, adherent cells were washed four times with culture medium to acquire purified PMϕ. For isolation of bone marrow–derived Mϕ (BMMϕ), femur and tibia were separated and flushed with PBS, and cells were pelleted by centrifugation. After resuspension, 3 × 10⁶ cells/ml were seeded in uncoated tissue culture flasks and cultured under low oxygen conditions (5% CO₂, 5% O₂) in Mϕ medium supplemented with murine M-CSF (30 ng/ml; Immunotools). After 7 d of differentiation, cells were washed twice with PBS. Adherent cells were scraped off and cultured in 96-well plates (2 × 10⁵ cells/cm²) under low oxygen conditions.

Isolated PMϕ and differentiated BMMϕ were incubated in the presence of 20 ng/ml murine IL-4 (Immunotools) or 30 ng/ml murine IFN-γ (Immunotools) for 24 h to achieve alternative or classical Mϕ activation. For proinflammatory stimulation, media of activated and nonactivated cells were supplemented with 25 ng/ml murine IL-1β or TNF-α (Immunotools), 1 μg/ml LPS (Escherichia coli serotype O111:B4; Sigma), 1 μg/ml peptidoglycan (PGN; Bacillus subtilis; Sigma), or 1 μg/ml zymosan A (Zym; Sigma).

Raw264.7 cells were genetically modified to allow inducible expression of the canonical IκBζ isoform (IκBζ_(L); GenBank accession no. NM_001159394.1) using the pINDUCER system (29). In brief, IκBζ_(L) was amplified from murine cDNA applying the primers 5′-CACCATGATCGTGGACAAGCTGCTG-3′ and 5′-CTAGTATGGTGGTGCTCGCTG-3′. The amplicon was cloned into pENTR-D-TOPO using TOPO cloning (Invitrogen), and the coding sequence was transferred to pINDUCER20 using gateway technology (Invitrogen). Upon sequence verification, lentivirus was generated using the Lenti-X system (Clontech). After lentiviral transduction, Raw264.7 cells were cultured in the presence of 500 μg/ml G418 for selection of Raw264.7/TetOn-IκBζ cells.

Mouse embryonic fibroblasts (MEFs) were isolated at embryonic day 10.5 (E10.5) according to standard procedures (30). MEFs and Raw264.7 cells were cultured under low oxygen and standard conditions, respectively. The latter additionally received 0.5 μg/ml G418 to maintain the transduced cell population. Cells were seeded at a density of 10⁵ cells/cm² 24 h before the experiments. Optionally, 2 μg/ml doxycycline (Sigma) was added to induce ectopic IκBζ expression. Subsequently, cells were cultured in the presence of 1 μg/ml LPS for proinflammatory stimulation.

Whole-cell RNA was isolated using the RNeasy mini kit (Qiagen) and reverse-transcribed (QuantiTect kit; Qiagen) as per manufacturer’s instructions. Quantitative PCR (qPCR) was performed in a LightCycler 480 II (Roche) using the Maxima Hot Start Taq DNA polymerase (Thermo) in the manufacturer’s two-step cycling protocol (384-well plates; 10 μl reaction). Transcripts were analyzed applying the following primer pairs: Gapdh (5′-ACCACAGTCCATGCCATCAC-3′, 5′-CACCACCCTGTTGCTGTAGCC-3′), Il6 (5′-AGTTGCCTTCTTGGGACTGA-3′, 5′-TCCACGATTTCCCAGAGAAC-3′), Ccl2 (5′-GAGGAAGGCCAGCCCAGCAC-3′, 5′-TGGGGCGTTAACTGCATCTGGC-3′), Nfkbiz (5′-TATCGGGTGACACAGTTGGA-3′, 5′-TGAATGGACTTCCCCTTCAG-3′), Elam1 (5′-CTCACTCCTGACATCGTCCTC-3′, 5′-ACGTTGTAAGAAGGCACATGG-3′), Tnfa (5′-CCTCAGCCTCTTCTCCTTCCT-3′, 5′-GGTGTGGGTGAGGAGCA-3′). Quantification of reverse-transcribed mRNA was performed using the second derivate maximum-based “Advanced relative quantification algorithm” of Roche’s LightCycler 480 Software (V1.5).

Total RNA was isolated from nonactivated PMΦ, which were either left untreated or stimulated with 1 μg/ml LPS for 4 h, and processed for hybridization on Affymetrix MoGene 1.0 ST arrays according to the manufacturer’s instructions. Single-strand cDNA was prepared using the Ambion WT Expression kit and labeled using the GeneChip WT Terminal Labeling kit. Hybridization, scanning, and generation of probe cell intensity files were performed on an Affymetrix GeneChip fluidics station 450 and GeneChip Scanner 3000 7G using Affymetrix Expression Console 1.1 software. Identification of LPS-regulated genes and comparison of LPS-induced gene expression in wild-type (wt) and knockout cells were performed as described previously (31). Data are published in the GEO database (GEO accession no. GSE43075; http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?token=vhghhumuiqkawje&acc=GSE43075).

Immunoblotting was performed as described previously (32). Anti–β-actin mAb was purchased from Sigma (clone AC-74; 1:2000). Polyclonal rabbit anti-IκBζ Ab was raised against a combination of peptides corresponding to short sequences of murine IκBζ_(L) (N-PGSPGSDSSDFSSTSSVSSC-C, N-QIRRILKGKSIQQRAPPY-C), affinity-purified with these peptides, and used at a concentration of 0.5 μg/ml.

A total of 10⁷ cells/sample were stimulated with 1 μg/ml LPS for indicated periods. Chromatin immunoprecipitation (ChIP) was performed with Diagenode’s Red ChIP kit using polyclonal rabbit anti-IκBζ or anti-histone H3 (trimethyl K4; Abcam) Abs. ChIP efficiencies were determined by qPCR on a LightCycler 480 II (Roche) using the Maxima Hot Start Taq DNA Polymerase in the manufacturer’s two-step cycling protocol (96-well plates; 20 μl volume). The following primer pairs spanning indicated promoter regions of the murine Ccl2 gene were used: Ccl2_-190/-3 (5′-TACCAAATTCCAACCCACAGTTTCTC-3, 5′-CTCTGCACTTCTGGCTGCTCTG-5′), Ccl2_-1379/-1245 (5′-GCCATCACAATGCTGCTAAGGA-3′, 5′-TCGTGGAGACAATGACTGTAAAGG-3′), Ccl2_-2491/-2303 (5′-CCTTTGTTGAGTCATTTCAGATTCTCC-3′, 5′-TCTGTCAACAAAGGATGTTCTTCCC-3′). Primers covering portions of Aktb (5′-TCGATATCCACGTGACATCCA-3′, 5′-GCAGCATTTTTTTACCCCCTC-3′) and Gapdh (5′-TACTAGCGGTTTTACGGGCG-3′, 5′-TCGAACAGGAGGAGCAGAGAGCGA-3′) genes were used as negative controls. Specificities of the primers were verified and PCR efficiencies were determined. Analysis was performed according to manufacturer’s standard procedures. In brief, treatment-specific ChIP efficiencies, expressed as percentage of input, were first determined via the formula: Eff = d⁻¹*(1+E)^{(CpINPUT-CpIP)}, where Eff represents immunoprecipitation- and primer-specific efficiencies, d is dilution factor of respective input DNA, E is respective primer efficiency, CpINPUT is Cp value of respective input sample, and CpIP is immunoprecipitation-specific Cp value. Next, treatment-specific relative occupancies, which define the specific ratio of a distinct signal over the background and are expressed as “Fold Ctrl,” were determined using the equation: O = Eff_{Locus,Treatment}/(Eff_{Gapdh,Treatment} + Eff_Aktb,_Treatment)*2, where O is treatment- and locus-specific relative occupancy and Eff_{Locus,Treatment} is locus- and treatment-specific ChIP efficiency.

Concentrations of CCL2, TNF-α, and IL-6 in peritoneal fluids and culture supernatants were determined by Cytometric Bead Arrays (Becton-Dickinson). Before stimulation of cells in 96-well plates, the culture medium was exchanged with 200 μl fresh medium per well. In case of BMMϕ, cytokine concentrations in culture supernatants were directly compared and expressed as cytokine amounts per volume. Because of donor mouse-specific differences in peritoneal cell counts, cytokine concentrations in supernatants of PMϕ were normalized to the protein content of cell lysates. To this end, 50 μl of 0.2 M NaOH was added to each well after aspiration of the supernatants, and protein concentrations in lysates were measured with the BCA Protein Assay (Thermo).

Male Nfkbiz^−/− and control mice were used at 6–18 wk of age. Animals received an i.p. injection of 500 μl PBS containing 40 μg Zym (Sigma) per gram body weight. Mice were sacrificed 4 h later by CO₂ asphyxiation. Subsequently, peritoneal cavity fluids were harvested and analyzed for cytokine amounts. In addition, peritoneal cells were FACS-phenotyped as described previously (32). After centrifugation (500 × g, 5 min), peritoneal cells were resuspended in 10 ml FACS buffer (PBS containing 2% FCS), repelleted by centrifugation (500 × g, 5 min), and resuspended in FACS buffer supplemented with Mouse SeroBlock FcR (1:100; AbD) to block Mϕ FcRs CD16 and CD32. After 10 min of incubation on ice, an equivalent volume of FACS buffer containing the following fluorochrome-labeled Abs at the indicated dilutions was added: rat anti-mouse F4/80:PE (A3-1; 1:20; AbD), rat anti-mouse CD19:FITC (AbD; 6D5; 1:20), rat anti-mouse Gr-1:Pacific Blue (RB6-8C5; 1:50; BioLegend), rat anti-mouse CD11b:allophycocyanin (M1/70; 1:50; Becton Dickinson), hamster anti-mouse CD11c:allophycocyanin-Cy7 (HL3; 1:50; Becton Dickinson), and rat anti-mouse CD117:PE-Cy5 (2B8; 1:50; Becton Dickinson). After 20 min, cells were pelleted and resuspended in FACS Lysing solution (Becton Dickinson) to lyse RBCs, washed twice in FACS buffer, counted, and analyzed on an LSRII flow cytometer (Becton Dickinson). BD CompBead (Becton Dickinson) standard operating procedures were used to calculate color compensation parameters. Data were acquired using the WEASEL software (The Walter and Eliza Hall Institute of Medical Research, Melbourne, Australia) and evaluated applying the gating scheme outlined in Supplemental Fig. 1 to determine the relative proportion of respective cell types.

Values are given as mean ± SD or mean ± SEM. For statistical comparisons, hypotheses were tested using unpaired standard t tests. Asterisks indicate statistical significance (*p < 0.05, **p < 0.005, ***p < 0.0005).

Even though IκBζ (Nfkbiz) was shown to be expressed in the monocyte cell lineage upon proinflammatory stimulation, its physiological relevance remains largely unclear. To elucidate novel aspects of IκBζ function, we focused on its impact on Mϕ physiology. Initially, PMΦ and BMMΦ from wt and Nfkbiz^ko/ko mice were assayed for basic cell-type features, such as phagocytosis, oxidative burst, or chemotactic responsiveness, which did not exhibit any phenotypical differences between both genotypes (data not shown). Next, global gene expression patterns of unstimulated and LPS-challenged nonactivated PMΦ were compared with respect to Nfkbiz-specific differences using cDNA microarray analysis (GEO accession no. GSE43075). In accordance with the function of IκBζ as a transcriptional coactivator (23), numerous LPS-inducible genes were expressed to a lower extent in Nfkbiz^ko/ko PMΦ, including Lcn2, Edn2, Il6, Il12b, Csf2, and Csf3 (Fig. 1A). At the same time, inducibility of NF-κB target genes supposedly not regulated by IκBζ, including Cxcl1 (27), Cxcl2 (27), and Tnfa (23), did not significantly differ between both genotypes. Accordingly, the approach seemed to allow the identification of several new putative target genes of IκBζ (e.g., Ccl17, Cxcl5, and Lif) exhibiting lower expression in stimulated knockout PMΦ. Our special interest hereafter focused on the Ccl2 locus, because CCL2 secretion is essential for monocyte recruitment during the onset of inflammation (4) and is, therefore, directly linked to the guardian function of Mϕ in peripheral tissues.

Previous analyses on Nfkbiz-dependent gene expression had been performed in classically (IFN-γ) activated PMΦ (23). To confirm Ccl2 as an IκBζ target gene, we therefore determined the kinetics of induction of Ccl2 mRNA levels and CCL2 cytokine levels in IFN-γ–activated PMΦ in response to LPS by quantitative RT-PCR (qRT-PCR; Fig. 1B) and cytometric bead assay assays (Fig. 1C), respectively. TNF-α, which is supposedly regulated independently of IκBζ, was expressed at comparable levels in wt and Nfkbiz^ko/ko PMΦ. In contrast, regardless of LPS stimulation, the mRNA levels of Ccl2 were significantly lower in IκBζ-deficient PMΦ and, moreover, exhibited altered induction kinetics with mRNA levels peaking earlier compared with wt cells. Similar characteristics were observed for Il6 that served as control for IκBζ-targeted gene regulation. Consistently, CCL2 levels in supernatants of Nfkbiz^wt/wt PMΦ were ∼6-fold higher than in knockout supernatants, whereas amounts of TNF-α did not differ (Fig. 1C).

To determine whether this dependency of Ccl2 expression on the Nfkbiz^ko/ko-genotype is conserved among Mϕ, we included the Mϕ-like Raw264.7 cell line and BMMΦ into our analysis. Raw264.7 cells were genetically modified to allow the doxycycline-inducible expression of the canonical isoform of IκBζ (IκBζ_(L); Fig. 2A, upper panel). In Raw264.7/TetOn-IκBζ cells, doxycycline treatment induced Nfkbiz mRNA levels tightly correlating with increased mRNA levels of Ccl2 (maximum: ∼30-fold) and Il6, whereas the IκBζ-independent NF-κB target genes Elam1 and Nfkbia were not expressed to a higher degree (Fig. 2A, lower panel). This regulation of Ccl2 was also reflected at the level of CCL2 secretion, which was ∼5-fold induced upon doxycycline-triggered IκBζ expression (Fig. 2A, lower panel).

Similar to PMΦ, BMMΦ from wt and Nfkbiz^ko/ko mice were stimulated with LPS and compared with respect to mRNA expression of Ccl2, Il6, and Tnfa (Fig. 2B). The inducibility of Ccl2 and Il6 was massively impaired in Nfkbiz^ko/ko BMMΦ, whereas Tnfa expression did not significantly differ in both genotypes. Finally, MEFs isolated from Nfkbiz^ko/ko mice revealed a similar pattern of gene regulation and showed significantly lower levels of Ccl2 mRNA after LPS stimulation as compared with Nfkbiz^wt/wt cells (Fig. 2C).

Because our data clearly indicated an influence of IκBζ on Ccl2 transcription, we next investigated a direct interaction of IκBζ with the Ccl2 promoter. Two distinct promoter portions have been shown to be relevant in Ccl2 gene regulation: a proximal promoter region and a distal enhancer portion, both of which harbor NF-κB–binding sites that may serve as anchor points for the recruitment of IκBζ (Fig. 3A).

We first performed qPCR-coupled ChIP (qPCR-ChIP) analysis in LPS-stimulated and untreated Raw264.7 cells (Fig. 3B). The relative occupancy of the Ccl2 locus was not increased using an IκBζ-specific Ab compared with negative control loci (Gapdh, Actb) in unstimulated cells exhibiting no or minimal IκBζ expression. In contrast, the occupancy of the proximal promoter region was ∼2.5-fold higher compared with the ones of control loci in the presence of LPS-induced IκBζ. Importantly, no enrichment was detected at the distal enhancer region in these samples, indicating that IκBζ was specifically recruited to the proximal Ccl2 promoter region.

The trimethylation of histone H3 at lysine-4 (H3K4m3), a histone signature of open chromatin, is linked to promoter regions with active transcription. We therefore investigated H3K4m3 occupancy of the proximal Ccl2 promoter region. qPCR-ChIP analysis revealed increased H3K4 trimethylation after LPS stimulation of BMMΦ from Nfkbiz^wt/wt mice, which was absent in nonstimulated cells (Fig. 3C). Interestingly, almost no binding of H3K4m3 was found after LPS treatment of Nfkbiz^ko/ko BMMΦ, indicating that the transcription-enhancing H3K4 trimethylation at the Ccl2 promoter depends on IκBζ.

Expression of IκBζ is not restricted to LPS stimulation but can be induced in response to several activators of NF-κB, and especially of TLR/IL-1 signaling (23, 24). To investigate whether expression of Ccl2 and CCL2 secretion were generally impaired in IκBζ-deficient cells, we analyzed the kinetics of Nfkbiz and Ccl2 induction in IFN-γ–activated PMΦ in response to several proinflammatory stimuli, including PGN, Zym, IL-1β, and TNF-α. PGN (Fig. 4A) and Zym (Fig. 4B) strongly induced Nfkbiz, as well as Il6 and Ccl2. The mRNA induction of both cytokines, but not of Tnfa, was significantly reduced in Nfkbiz^ko/ko cells when compared with wt PMΦ. Induction of Nfkbiz and the cytokines was significantly lower in response to TNF-α (Fig. 4C) and IL-1β (Fig. 4D), but even under these conditions an impaired Ccl2 expression was observed in IκBζ-deficient cells. Consistent results were obtained at the level of cytokine secretion, when culture supernatants of cells were assayed after 24 h of stimulation (Fig. 4E).

The previous experiments used PMϕ that were classically activated with IFN-γ, which induces Mϕ M1 polarization. Naive PMΦ and PMΦ activated with IL-4 to induce alternative (M2) polarization were assayed for cytokine secretion after 24 h of stimulation with LPS, PGN, IL-1β, and TNF-α, respectively, to investigate whether gene-regulatory effects of IκBζ are dependent on Mϕ polarization (Fig. 5A). The same experiments were also performed with differentially activated BMMΦ (Fig. 5B). Moreover, stimulation of PMΦ with Zym was performed, as that trigger was later used for in vivo experiments. Interestingly, whereas IL-6 production strictly depended on IκBζ in classically (IFN-γ) activated PMΦ (Fig. 4E), in naive and IL-4–primed PMΦ, IL-6 secretion occurred largely independently of IκBζ (Fig. 5A). Unlike PMΦ, BMMΦ revealed a strict IκBζ dependence of IL-6 secretion after stimulation with LPS and PGN, regardless of their polarization status (Fig. 5B). With respect to CCL2, cytokine secretion was strongly impaired in Nfkbiz^ko/ko PMΦ, as well as in BMMΦ, regardless of whether experiments were conducted in naive, classically activated, or alternatively activated cells. Secretion of TNF-α, which is not directly regulated by IκBζ, was not impaired but generally even slightly increased in the absence of IκBζ. Altogether, these results demonstrate that CCL2 secretion, if triggered by a proinflammatory stimulus, is strictly dependent on IκBζ in both PMΦ and BMMΦ in each state of activation. In contrast, for other cytokines such as IL-6, the requirement of IκBζ is influenced by the Mϕ type and their activation status.

CCL2 has originally been identified to be necessary for the inflammatory recruitment of monocytes in several models of experimental peritonitis (4–6). Thus, to investigate whether IκBζ is also required for the regulation of Ccl2/CCL2, we used an experimental peritonitis model in IκBζ-proficient and -deficient mice (Fig. 6A). This seemed reasonable, because the induction of endogenous CCL2 in this system is highly Mϕ dependent (6). To induce systemic inflammation, we injected mice with Zym into the peritoneal cavity, and 4 h later, levels of TNF-α, IL-6, and CCL2, as well as quantities of monocyte infiltration, were determined. FACS phenotyping procedures applied for that purpose (33) are outlined in Supplemental Fig. 1. As expected, Zym-treated Nfkbiz^ko/ko mice exhibited lower levels of CCL2 (∼50%) compared with equally treated wt animals (Fig. 6B). Similar results were observed for IL-6 (∼30% of Nfkbiz^wt/wt mice), whereas levels of TNF-α were ∼5-fold higher in IκBζ-deficient animals. With respect to the physiology of inflammation, the number of infiltrating monocytes was interestingly ∼3-fold higher in wt animals (Fig. 6C).

Growing evidence suggests that the induction of NF-κB–regulated genes is not solely defined by the nuclear translocation of NF-κB, but that each NF-κB target gene has an individual expression profile regarding kinetics, cell type, stimulus specificity, or requirement of cofactors, which is thought to ensure the selectivity of an immune response. A comparison of global gene expression patterns of LPS-stimulated Nfkbiz^wt/wt and Nfkbiz^ko/ko PMΦ identified, among other genes, Ccl2 as a strongly inducible target of IκBζ-mediated transcription. To verify this finding, we challenged PMΦ and BMMΦ with various proinflammatory activators for the analysis of Ccl2 gene expression and CCL2 secretion. Our experiments consistently revealed that Nfkbiz^ko/ko Mϕ exhibited profoundly impaired Ccl2 expression and CCL2 secretion, in particular, when stimulated with TLR ligands. At the level of gene transcription, similar observations were made in untreated cells or after stimulation with IL-1β or TNF-α, which elicit only weak inflammatory responses. These findings were supported by the observation that the inducible expression of ectopic IκBζ in Raw264.7 cells tightly correlated with increasing Ccl2 mRNA levels. Thus, our results in knockout and overexpression models suggest that transcriptional regulation of Ccl2 directly depends on IκBζ.

IκBζ has been shown to exert its gene-activating effects largely on so-called secondary response genes. In contrast with primary NF-κB target genes that are immediately induced in the absence of new protein synthesis, secondary response genes require the prior synthesis and coactivator function of IκBζ, which itself is a primary NF-κB target gene (21, 23, 27). Consequently, these genes exhibit a delayed kinetics of gene induction. Consistent with this model, the kinetics of gene induction obtained in LPS-stimulated Raw264.7 cells and PMΦ could clearly discriminate between Nfkbiz and Tnfa as IκBζ-independent primary response genes on one side, and Il6 and Ccl2 as IκBζ-dependent secondary response genes on the other hand. Maximal expression of Il6 and Ccl2 was delayed but prolonged in Nfkbiz^wt/wt cells. In contrast, in the absence of IκBζ, Ccl2 and Il6 expression were considerably impaired and not sustained, but rather resembled Tnfa and Nfkbiz in the kinetics of gene induction (Fig. 1B). In addition, our ChIP analyses revealed that IκBζ was directly recruited to the proximal region of the Ccl2 promoter, which harbors a single NF-κB binding site. The recruitment of IκBζ was associated with histone H3K4 trimethylation of the proximal promoter region as a marker of active transcription. Interestingly, in the absence of IκBζ, H3K4 trimethylation did not occur, which, together with other lines of evidence, suggests that chromatin remodeling events are essential for IκBζ action.

CCL2 expression is essentially required for monocyte recruitment to sites of inflammation and has therefore been implicated in (auto)inflammatory diseases, as well as in the course of immune defense against pathogens. This ambivalence of beneficial and pathogenic effects of CCL2 underscores the requirement of a tight regulation of proinflammatory factors for an appropriate immune response. To investigate the role of IκBζ-mediated CCL2 secretion in vivo, we chose a model of experimental peritonitis in which CCL2 secretion is predominantly driven by PMΦ (6). Our results demonstrate that knockout animals indeed exhibit a decreased capacity to secrete CCL2, as well as IL-6, and thus provide further evidence for the tight control of both cytokines by IκBζ. Interestingly, whereas IL-6 and CCL2 expression were strongly reduced in IκBζ-deficient mice, TNF-α levels were not decreased but even increased in comparison with wt animals. Similar findings were observed in PGN-stimulated BMMΦ and PMΦ from Nfkbiz^ko/ko animals and might be caused by the fact that IκBζ can additionally function as a specific inhibitor of particular target genes (23).

Although cells of the monocyte/Mϕ lineage are the major source of CCL2, several other cell types, including fibroblasts and epithelial and endothelial cells, can express CCL2 (16). Further investigations are necessary to verify whether the strict control of CCL2 by IκBζ is also relevant in these cell types. Our exemplary investigation of wt and Nfkbiz^ko/ko MEFs indicates that the described mechanism is not restricted to Mϕ. Interestingly, however, we found that, unlike CCL2, the regulation of IL-6 by IκBζ might be also dependent on cell type-specific differentiation. For instance, whereas Il6 expression was strongly impaired in IκBζ-deficient BMMΦ, regardless of their polarization status, only M1 (IFN-γ)-polarized but not naive or M2 (IL-4)-polarized PMΦ revealed a profound inhibition of Il6 expression in the absence of IκBζ. Thus, this observation indicates that the assignment of a particular gene as an IκBζ-dependent secondary response gene might be also dependent on the cell type and specific chromatin context. Because our study did not aim to investigate the involvement of IκBζ in Mϕ polarization, further questions concerning this issue have not been addressed in this article. However, our data may not only reflect the fact that Mϕ polarization influences IκBζ-dependent gene regulation, but that vice versa IκBζ could have an impact on Mϕ activation.

In future studies, it will be interesting to investigate the physiological impact of IκBζ-dependent Ccl2 regulation in different (patho)physiological contexts. Our results obtained in the experimental peritonitis model indicate that IκBζ deficiency causes impaired monocyte recruitment because of reduced CCL2 secretion. Nevertheless, this dependency may not exclusively be caused by defective CCL2 production. Given the fact that IκBζ influences the expression of several inflammatory key players, including IL-6, GM-CSF, G-CSF, IL-12p40, and IL-17, additional effects may influence homeostasis and trafficking of monocytes. IL-6, for example, has been shown to orchestrate the switch from neutrophil to monocyte recruitment during acute inflammation. In that context, the observation that TNF-α levels in peritoneal cavity fluids were dramatically higher in case of Nfkbiz^ko/ko mice upon i.p. injection of Zym is interesting, because Nfkbiz^wt/wt and Nfkbiz^ko/ko PMΦ did not exhibit such differences in vitro with respect to TNF-α secretion. Similar, but so far unexplained, effects have been observed previously (23) and might be secondary because of the downregulation of IκBζ-dependent genes regulating the expression of TNF-α in certain cell types. Differences in gene expression of additional cell type contributing to the outcome of the systemic peritonitis model might therefore explain this phenomenon. Thus, the ability of IκBζ to concertedly regulate various cytokines and physiological functions might distinguish IκBζ as a key transcriptional regulator of certain inflammatory processes (34). Moreover, other biological processes, such as T_H2 polarization (35), bone morphogenesis (16), or tumor progression (9, 16, 36) have been shown to be regulated by CCL2 and could be controlled by IκBζ.

In summary, we have uncovered an essential regulatory mechanism of Ccl2 gene regulation in Mϕ. We demonstrate that IκBζ, presumably by the p50 NF-κB–mediated recruitment to the proximal Ccl2 promoter and subsequent histone modification, enables transcription of the Ccl2 locus, thus allowing production of high levels of CCL2 in the course of proinflammatory processes. As indicated by observations in an in vivo model of systemic inflammation, this IκBζ-dependent mechanism may be crucial for Mϕ physiology, for example, in the course of Mϕ-mediated effector cell recruitment.

The authors have no financial conflicts of interest.