Interleukin-18 drives ILC3 proliferation and promotes interleukin-22 production via NF-κB

Abstract

Group 3 innate lymphoid cells (ILC3s) are important regulators of the immune system, maintaining homeostasis in the presence of commensal bacteria, but activating immune defenses in response to microbial pathogens. ILC3s are a robust source of IL-22, a cytokine critical for stimulating the anti-microbial response. We sought to identify cytokines that can promote proliferation and induce or maintain IL-22 production by ILC3s and determine a molecular mechanism for this process. We identified IL-18 as a cytokine that cooperates with an ILC3 survival factor, IL-15, to induce proliferation of human ILC3s, as well as induce and maintain IL-22 production. To determine a mechanism of action, we examined the NF-κB pathway, which is activated by IL-18 signaling. We found that the NF-κB complex signaling component, p65, binds to the proximal region of the IL22 promoter and promotes transcriptional activity. Finally, we observed that CD11c⁺ dendritic cells expressing IL-18 are found in close proximity to ILC3s in human tonsils in situ. Therefore, we identify a new mechanism by which human ILC3s proliferate and produce IL-22, and identify NF-κB as a potential therapeutic target to be considered in pathologic states characterized by overproduction of IL-18 and/or IL-22.

Introduction

Innate lymphoid cells (ILCs) play many roles in protective immunity and disease (1). There are several types of ILCs, each of which produces a characteristic cytokine profile and is regulated by unique transcription factors (2). Group 3 ILCs are composed of ILC3s and lymphoid tissue-inducer cells (LTi). They are dependent on the transcription factors RAR-related orphan receptor gamma isoform b (RORγt) and aryl hydrocarbon receptor (AHR), and produce the cytokines IL-17A and IL-22 (3–7). ILC3s are enriched within human secondary lymphoid tissues (SLTs), such as the tonsils and lymph nodes (8–10).

ILC3s are critical regulators of homeostasis and immunity (9–12). ILC3s are the primary steady state source of IL-22, a cytokine critical for tissue regeneration and for maintenance of barrier function in the gut, skin, oral mucosa, and lung (3, 13). IL-22 signaling in epithelial cells drives genes involved in proliferation and wound healing (14, 15). In addition to bolstering the physical barrier of the epithelium, IL-22 stimulates epithelial cells to produce anti-microbial peptides necessary for barrier maintenance and prevention of infection by commensal bacteria (16, 17). ILC3s are also necessary for the formation of cryptopatches and isolated lymphoid follicles in the intestinal lamina propria (4) as well as repair of lymph nodes following infection (18). IL-22 and IL-18 have recently been shown to cooperatively contribute to murine intestinal immunity to multiple infectious agents. IL-18 induces IL-22 during T. gondii infection in the murine ileum, while IL-22 induces IL-18 during C. rodentium infection (17). A combination of IL-18 and IL-22 was shown to be critical for clearance of rotavirus in infected mice (19).

In this study, we identified IL-18 as a cytokine that can induce proliferation of and sustain IL-22 production by human ILC3s. IL-18 signals through the IL-18 receptor to activate NF-κB signaling. In the tonsil, ILC3s reside in close proximity to dendritic cells (DCs), a source of IL-18. Together, these data support the hypothesis that DC-derived IL-18 stimulates ILC3 function by maintaining the population through proliferation and by sustaining production of IL-22 through an NF-κB-dependent mechanism. Our study further clarifies the role of DCs in ILC3 function and identifies NF-κB as a potential target for future therapies against IL-22-mediated diseases.

Materials and Methods

Isolation of human ILC3s and developmental precursors

All procedures were performed with approval of the Ohio State University Institutional Review Board. Normal human pediatric tonsils were obtained following routine tonsillectomy from Nationwide Children’s Hospital (Columbus, OH). ILC3s and developmental precursors were isolated as previously described (20). Briefly, total mononuclear cells were depleted of CD19⁺ and/or CD3⁺ cells via magnetic negative selection (Miltenyi Biotec). For some experiments, B and/or T cell depleted mononuclear cells were used immediately for flow cytometric analysis. Alternatively, ILC3s were sorted directly from the depleted fraction by gating on CD3⁻CD14⁻ CD19⁻CD20⁻CD34⁻CD16⁻CD94⁻CD117⁺ events on a FACSAria II cell sorter (BD Biosciences). Purity analysis routinely revealed that sorted populations were ≥97% pure.

Cell culture

ILC3s purified by fluorescence activated cell sorting (FACS) were cultured in a round-bottom 96-well plate (Costar) at a starting density of 2.5 × 10⁴ cells/mL in α-MEM medium containing 10% fetal bovine serum, penicillin G (100 μg/mL), and streptomycin (100 μg/mL) (Invitrogen). Cells were cultured with the indicated recombinant human cytokines, including IL-15 (1 nM, Amgen), IL-18 (100 ng/mL, MBL), IL-1β (10 ng/mL, PeproTech), IL-12 (10 ng/mL, Genetics Institute), IL-6 (20 ng/mL, R&D), IL-27 (10 ng/mL, R&D), IL-21 (100 ng/mL, R&D), IL-23 (20 ng/mL, Miltenyi), IL-7 (10 ng/mL, Miltenyi), IL-10 (10 ng/mL, Shering), IL-25 (100 ng/mL, R&D), IFN-α (20 U/mL, Shering), IFN-γ (2 U/mL, Peprotech), TRAIL (10 ng/mL, R&D), and TGF-β (20 ng/mL, R&D). Pam2CSK4 (TLR2/6 agonist) was obtained from InvivoGen (1 μg/mL). All cell culture was performed in the presence of IL-15 because it serves as a survival factor for ILC3s (21). ILC3s were cultured for 14 days in proliferation assays, unless otherwise indicated. Cells were enumerated by a Trypan blue exclusion assay.

Flow cytometry

Antibodies for human CD34, CD3, CD16, and CD117 were purchased from BD Biosciences; those for CD94, IL-18Rα, IL-18Rβ, IL-6Rα, TGF-βRII, IL-27Rα, IL-10Rα, IL-17RB, TRAIL R1, and IL-22 were purchased from R&D Systems; those for CD14, CD19, CD20, IL-7R (CD127), IL-12R, IFNAR2, IL-10Rβ, and IL-21R were purchased from Miltenyi Biotec. Unless otherwise indicated, antibodies were used according to manufacturers’ instructions. For survival studies, annexin V protein (BD Biosciences, conjugated to BV421 fluorochrome) was added at specific time points to cultured ILC3s resuspended in annexin V binding buffer (BD Biosciences, 10× solution diluted in distilled water for 1× working concentration). Following annexin V binding, cells were stained with propidium iodide (BD Biosciences) immediately prior to analysis by flow cytometry. For proliferation studies, 5 μM CellTrace Violet Stain (tracer dye) from Life Technologies was added to cells prior to culture and analyzed by flow cytometry after 7 or 14 days. For 5-ethynyl-2′-deoxyuridine (EdU) incorporation, cells were pulsed with 10 μM EdU for 30 minutes and cultured for 1 hour in EdU-free medium before cytometric analysis. For intracellular IL-22 staining, ILC3s were cultured for 14 days with either IL-15 alone or IL-15+IL-18. Staining for intracellular IL-22 was then performed following a 4-hour incubation in 2 μM GolgiPlug (BD Biosciences), and using the Cytofix/Cytoperm Plus Fixation/Permeabilization Kit (BD Biosciences) and α-IL-22 PE (R&D Systems). Flow cytometry was performed on an LSR II flow cytometer (BD Biosciences) and analysis was performed using FlowJo Software (Treestar, Inc.).

Cell lysis and immunoblotting

Protein lysate preparation and immunoblotting were performed as described previously (22). Primary antibodies used were: rabbit polyclonal anti-human GAPDH antibody (Santa Cruz), rabbit polyclonal anti-human p65 antibody (Rockland), and rabbit monoclonal anti-human phospho-p65^ser536 (Cell Signaling).

Enyzme-linked immunosorbent assay (ELISA)

ELISA was performed with the eBioscience Human IL-22 ELISA Ready-SET-Go! Kit (ThermoFisher) and IL-22 standard (ThermoFisher) according to manufacturer’s protocols. Supernatant was collected from sorted ILC3s cultured with IL-15+DMSO, IL-15+IL-18+DMSO, or IL-15+IL-18+TPCK. TPCK was used at a concentration of 10 μM (Sigma-Aldrich).

EMSA and antibody-supershift assays

Nuclear extracts were isolated using a Nuclear Extraction Kit (Active Motif). Complementary oligonucleotides—probe 1 (−248 bp to −216 bp, referred to as transcription start site) and probe 2 (−191 bp to −163 bp)—containing a putative NF-κB binding site from the human IL22 promoter were synthesized. Electrophoretic mobility shift assays (EMSA) were performed as described previously (23). In brief, probes 1 and 2 were ³²P-labeled and incubated with the nuclear extracts (2 μg) in the presence of poly[dI-dC] (1 μg). For the antibody gel supershift assays, nuclear extracts were incubated with antibodies for p65 (Rockland) or p50 (Millipore) overnight at 4°C before the addition of the IL22 promoter probes.

Transient transfection and luciferase assay

293T cells were maintained in DMEM supplemented with 10% fetal bovine serum, penicillin G (100 μg/mL), and streptomycin (100 μg/mL) (Invitrogen). The cell line was obtained from ATCC in 2008; it has not been authenticated since receipt. 293T cells were seeded into 24-well plates at a density of 2.5 × 10⁴ cells per well and grown overnight. Co-transfection of the IL22-Luc construct with a vector expressing p50, or a vector expressing p65, or a combination of both expression vectors, or an empty vector was performed using Lipofectamine 2000 with Plus Reagent (Invitrogen) per manufacturer’s protocol. The pGL3 basic reporter vector was used as a control for basal promoter activity. A Renilla luciferase vector, pRL-TK (Promega), was co-transfected to serve as an internal control for transfection efficiency. Cells were harvested after 48 hours and assessed for luciferase activity as previously described (22). The experiment was performed with the wild-type IL22 promoter sequence and the sequence carrying the mutated putative NF-κB binding sites. In mutation 1, the sequence corresponding to probe 1 (5′-GAAAATTTCTGGGATTTGTC-3′) was changed to 5′-GCACATCTCCGAGCTTCGAC-3′. In mutation 2, the sequence corresponding to probe 2 (5′-GGGAAACACT-3′) was changed to TTGACACTCT. In the double mutation, both sequences were changed accordingly.

Immunohistochemistry (IHC), imaging, and multispectral analysis of dual color IHC

Paraffin-embedded tonsillar tissue sections (0.5 μm in thickness) were stained for IHC using a Bond Rx autostainer (Leica). Briefly, slides were baked at 65°C for 15 minutes and automated software performed dewaxing, rehydration, antigen retrieval, blocking, primary antibody incubation, post primary antibody incubation, detection (DAB or RED), and counterstaining using Bond reagents (Leica). Samples were then removed from the machine, dehydrated through ethanol series and xylenes, mounted, and coverslipped. Antibodies for the following markers were diluted in Antibody Diluent (Leica): rabbit antibodies-CD3 (1:100, Dako), IL-18 (1:1000, Sigma); mouse antibodies-CD11c (1:100, Leica), IL-1β (1:100, CST), Mast Cell Tryptase (1:300, Leica), RORγt (1:800, Millipore); goat antibody-CD117 (1:50, R&D Systems). Images were digitally captured using an Axiocam 105 color camera, an Observer.Z1 microscope, and Plan-NEOFLUAR objectives (Zeiss). Broad field high resolution images were collected as 5×5 grids and digitally stitched together using ZEN 2 (blue edition) software (Carl Zeiss Microscopy GmbH). Some dual stained samples were imaged using the PerkinElmer’s Vectra® multispectral slide analysis system. For visualization of the component images, the multispectral images were spectrally unmixed using Inform software. Briefly, red (CD117), brown (MCT), and blue (DAPI) channels were defined using Nuance spectral libraries. Individual component images were used to identify single- and dual-positive cells.

Results

IL-18 stimulation promotes ILC3 proliferation

To explore the role of IL-18 and other cytokines in the expansion of ILC3s, CD3⁻CD34⁻CD94⁻CD117⁺ ILC3s were purified by FACS from fresh human tonsils (Supplemental Fig. 1A). ILC3s were then stimulated with cytokines as indicated for 14 days (Fig. 1A, Supplemental Fig. 1B). All ILC3 cultures were performed in the presence of IL-15, which served as a survival factor (21). ILC3s stimulated with IL-6, TGF-β, IL-21, IL-23, IL-27, IL-10, IL-7, TRAIL, IL-25, IL-12, or TGF-β plus IL-6 demonstrated little to no increased expansion—and in one case (IFN-α) demonstrated decreased expansion—compared to controls cultured in IL-15 alone. However, when ILC3s were cultured in IL-15 plus IL-18 they demonstrated significant expansion compared to ILC3s cultured in IL-15 alone, as measured by both fold change and absolute numbers (Fig. 1A, Supplemental Fig. 1B). We also compared cultures of ILC3s with IL-18 to its combination with IL-12 or TGF-β to determine if IL-12 or TGF-β enhanced or impaired ILC3 proliferation induced by IL-18. The addition of IL-12 or TGF-β did not significantly alter the proliferation of ILC3s compared to stimulation with IL-18 (Supplemental Fig. 2A). Additionally, treatment of ILC3s with IFN-γ led to a decrease in cell number compared to controls (Supplemental Fig. 2B).

Figure 1. ILC3 proliferation in response to treatment with cytokines in the presence of IL-15.

(A) Relative fold change was calculated and averaged from 6 donors as the absolute number of cells enumerated (by Trypan blue exclusion) after 14 days of culture in IL-15 plus cytokine divided by the absolute number of cells enumerated after 14 days of culture in IL-15 alone (** denotes p < 0.01, * denotes p < 0.05. IL-15+IL-18 vs other treatments; n=6). Data are expressed as mean ± SD. (B) ILC3s were sorted, labeled with tracer dye, and cultured in IL-15 alone or IL-15 plus IL-18 for 14 days. Cells were analyzed by flow cytometry at 7 and 14 days for retention of tracer dye. Histograms of a representative experiment are shown, n=3.

To determine if the increase in cell number in the presence of IL-15 plus IL-18 compared to IL-15 alone could be attributed to proliferation, we performed a fluorescent dye proliferation assay. ILC3s were labeled with tracer dye prior to culture, and after 7 and 14 days cultured cells were harvested, counted, and examined for retained fluorescence intensity by flow cytometry. While ILC3s cultured in IL-15 alone as well as those cultured in IL-15 plus IL-18 both proliferated over 7 days and 14 days, those cultured in IL-15 plus IL-18 showed significantly increased proliferation compared to those cultured in IL-15 alone (at 14 days: IL-15 alone MFI 5746 vs. IL-15+IL-18 MFI 992, p=0.04) (Fig. 1B). These data were supported by a 5-ethynyl-2′-deoxyuridine (EdU) incorporation assay, in which greater EdU incorporation was observed in ILC3s cultured in IL-15 plus IL-18 compared to those cultured in IL-15 alone (Supplemental Fig. 2C). We assessed survival between the two conditions by annexin/PI staining. We found no difference of the percentage of live (annexin⁻/PI⁻) or apoptotic cells (annexin⁺/PI⁻) on treatment days 1, 3, 7, and 14 (data not shown). These data suggest that the quantitative expansion of ILC3s following culture in IL-15 plus IL-18 is due to an increase in proliferation rather than enhanced survival. We also observed that treatment of ILC3s with the TLR2/6 agonist Pam2CSK4 did not augment ILC3 proliferation in the presence of IL-15 alone or in the presence of IL-15 plus IL-18 (data not shown).

Expression of IL-18Rα and IL-18Rβ on the cell surface of ILC3s

Human IL-18 signals through a heterodimeric receptor composed of the IL-18Rα and IL-18Rβ subunits (24). To validate that IL-18 can signal through its receptor on ILC3s, we examined the expression of IL-18Rα and IL-18Rβ on the surface of ILC3s as well as in early and late hematopoietic progenitor cells (HPCs) by flow cytometry. Whereas surface IL-18Rβ was expressed by the majority of early (CD34⁺CD117⁻) HPCs, late (CD34⁺CD117⁺) HPCs, and mature ILC3s, robust surface expression of IL-18Rα was restricted to mature ILC3s (Fig. 2). Of note, mature natural killer (NK) cells derived from peripheral blood also express both IL-18R subunits (Supplemental Fig. 3A). The constitutive expression of the heterodimeric receptor on the mature ILC3s is consistent with the above proliferation data and suggests that IL-18 signaling may have a unique role in regulating differentiation or homeostasis of ILC3s. We also assessed for cytokine receptor expression for the other cytokines used in the proliferation assay (Supplemental Fig. 3B). Freshly isolated ILC3s showed expression of the receptor for IL-7, while receptors for IL-6, TGF-β, IL-21, IL-27, IFN-α, IL-10, TRAIL, IL-17, IL-25, and IL-12 were either absent or negligibly expressed following flow cytometric staining.

Figure 2. Expression of IL-18Rα and IL-18Rβ on HPC progenitors and ILC3s freshly isolated from human tonsil.

Histograms show a representative donor stained ex vivo with isotype (unfilled) or antibody specific for the indicated antigen (filled). CD34+CD117− HPCs were more specifically defined as CD3⁻CD19⁻CD34⁺CD117⁻CD94⁻. CD34⁺CD117⁺ HPCs were more specifically defined as CD3⁻CD19⁻CD34⁺CD117⁺CD94⁻. ILC3s were defined as CD3⁻CD19⁻CD34⁻CD117⁺CD94⁻.

IL-18 sustains IL-22 protein expression in ILC3s

Previous studies have demonstrated that freshly isolated ILC3s constitutively produce high levels of IL-22 (8–10, 25–30). In order to determine whether IL-18 plays a role in regulating IL-22 production, ILC3s were isolated from human tonsils by FACS and cultured with IL-15 and the presence or absence of IL-18 for 14 days. A minor fraction of cultured cells expressed CD94 at day 14 (which identifies mature NK cells) under both culture conditions. We then sought to characterize the in vitro-derived cells generated in response to each treatment. Following culture, CD94 expression and IL-22 production were mutually exclusive, demonstrating that only ILC3s, and not in vitro-derived mature NK cells, were capable of producing IL-22 in response to cultures containing IL-18 (Fig. 3A). While ILC3s cultured with IL-18 acquired CD94 surface expression at a rate similar to that of ILC3s cultured in the presence of IL-15 alone, the inclusion of IL-18 in the cultures led to a significantly greater fraction of cells expressing IL-22 (1.1% vs 12.3%) (Fig. 3B). Thus, IL-18 expands ILC3s and maintains their production of IL-22.

Figure 3. ILC3s treated with IL-18 maintain expression of IL-22.

(A) Representative staining for surface CD94 and intracellular IL-22 protein at day 14 of culture. (B) Percentage of ILC3s staining positively for IL-22 was assessed via flow cytometry after a 4 hr culture with either IL-15 alone or IL-15+IL-18 (p < 0.05; n=4). Data are expressed as mean ± SD.

Inhibition of NF-κB blocks IL-18-induced IL-22 production

To determine a potential mechanism by which IL-18 may be regulating IL-22 production in ILC3s, we considered the downstream intermediates through which IL-18 signals (31, 32), i.e. NF-κB (33) and MAP kinases (34, 35). To test whether NF-κB signaling mediates IL-22 production in response to IL-18, we employed the NF-κB antagonist N-Tosyl-L-phenylalanine chloromethyl ketone (TPCK). ILC3s were treated with IL-15+DMSO, IL-15+IL-18+DMSO, or IL-15+IL-18+TPCK. We confirmed that the addition of IL-18 stimulated IL-22 production compared to IL-15+DMSO (Fig. 4). Furthermore, we found that treatment of ILC3s with TPCK in the presence of IL-15+IL-18 almost completely blocked this response. This suggests that the stimulatory effect of IL-18 on IL-22 production by ILC3s is mediated mainly by NF-κB signaling.

Figure 4. The NF-κB antagonist TPCK inhibits IL-22 production by ILC3s in the presence of IL-18.

ILC3s were cultured for 14 days in the presence of IL-15+DMSO, IL-15+IL-18+DMSO, or IL-15+IL-18+TPCK. At day 14, the supernatants were harvested for ELISA to measure production of IL-22. Data are expressed at mean ± SEM (n=8). For statistical analysis, data were transformed by log base 2 in order to stabilize the variance. Linear mixed model was performed to account for the covariance structure due to repeated measures on the same donor. P values were adjusted for multiple comparisons by Holm’s procedure. A P value of 0.05 or less was considered statistically significant. ** denotes p < 0.01.

IL-18 activates NF-κB via phosphorylation of p65 and induces NF-κB binding to the IL22 promoter

To further support that NF-κB mediates IL-18 signaling in ILC3s, we identified three NF-κB binding sites in the proximal promoter region of the human IL22 gene, two of which are immediately adjacent to one another (Fig. 5A). NF-κB dimers bind target gene regulatory regions through binding sites that generally match the consensus sequence 5′-GGGRNTY(C/T)C-3′ (R = A or G, Y = C or T, N = any nucleotide) (36). The promoter sequence corresponding to probe 1 contains the two adjacent putative NF-κB binding sites, and the promoter sequence corresponding to probe 2 contains the third binding site. The probe 2 site has high homology with the consensus binding sequence, while the probe 1 site has more mismatched bases. The addition of IL-18 to sorted ILC3s cultured in IL-15 revealed increased phosphorylation of NF-κB p65 at serine 536 in ILC3s (Fig. 5B). To test whether nuclear transcription factors in ILC3s bind to these sites, we synthesized two DNA probes containing the NF-κB binding sites in the IL22 promoter. In electrophoretic mobility shift assays, more DNA-protein complexes were observed in nuclear lysates prepared from ILC3s treated with IL-18 in the presence of the survival cytokine IL-15 than from ILC3s treated with IL-15 alone (Fig. 5C). Furthermore, an antibody supershift experiment showed the presence of NF-κB subunits p65 and p50 in the DNA-protein complexes (Fig. 5D). Taken together, these data demonstrate that phosphorylated NF-κB is capable of specific yet faint binding to the IL22 promoter in ILC3s stimulated with IL-15 alone, in contrast to specific and avid binding to the IL22 promoter in ILC3s stimulated with IL-15 plus IL-18. Thus, IL-18 phosphorylates NF-κB, and activated NF-κB binds to the IL22 promoter, suggesting it participates in the regulation of IL-22 expression.

Figure 5. NF-κB is activated by IL-18 and binds to the IL22 promoter.

(A) The IL22 promoter sequence. Boxes indicate predicted NF-κB binding sites. Oligonucleotide probes were synthesized using the indicated ranges of the IL22 promoter sequence. (B) ILC3s were treated with either IL-15 alone or IL-15+IL-18 and subjected to immunoblotting with the indicated antibodies. (C) EMSA was performed with ³²P-labeled predicted NF-κB binding sites (probe 1 or probe 2) and nuclear extract (NE) prepared from ILC3s treated with IL-15 alone or IL-15 plus IL-18 or no NE (−). The DNA-protein complexes are indicated with arrows. (D) Antibody gel supershift assay. Probe 1 was incubated with NE from ILC3s treated with IL-15+IL-18 and pre-incubated with no antibody, p65 antibody, p50 antibody, or without NE. The DNA-protein complexes and antibody gel supershifted complexes are indicated with arrows.

NF-κB subunit p65 positively regulates transcription at the IL22 promoter

Since we established that NF-κB can bind avidly to the IL22 promoter in an IL-18-dependent manner, we next set out to determine whether NF-κB is also capable of regulating IL-22 transcription. To this end, we performed a luciferase reporter gene assay in which the IL22 promoter was cloned immediately upstream of the luciferase gene in the pGL3 firefly luciferase construct. The IL22-Luc promoter construct was co-transfected with an empty vector or an expression vector containing p65 or p50 or with both expression vectors. Transfections with the p65 vector (p65 and p65/p50) demonstrated strong increases in luciferase activity compared to empty vector control (p<0.01) (Fig. 6A). Transfection with p65/p50 showed a partially attenuated effect on luciferase activity compared to p65. Of note, a similar mechanism has been reported for other target genes (37). The p50 subunit is not thought to be activing, and as expected, did not stimulate luciferase activity. To further characterize the NF-κB binding site, additional IL22 promoter constructs were generated, which contained mutations at the putative NF-κB binding sites around which probes 1 and 2 were designed. Mutation 1 (mut1) corresponds to the putative binding sites found in probe 1 and mutation 2 (mut2) corresponds to the putative binding site found in probe 2. Loss of the putative binding site due to mut2, but not mut1, resulted in decreased luciferase activity in response to p65 (p<0.01) (Fig. 6B). Double mutation did not impair luciferase activity beyond the effects of mut2 alone. These data demonstrate that the NF-κB binding site from probe 2 mediated IL22 expression through interaction with the NF-κB subunit p65. The promoter activity of the NF-κB binding site from probe 2 showed a greater contribution than those from probe 1, which corresponds with their relative sequence homologies to the NF-κB consensus binding site. These findings in combination with above EMSA showing occupancy of both NF-κB sites show that while NF-κB signaling is involved, it likely works in concert with other yet undetermined cofactors. Collectively, these data further support the hypothesis that IL-18 positively regulates IL-22 production by ILC3s through an NF-κB-mediated mechanism.

Figure 6. NF-κB subunit p65 positively regulates transcription through the IL22 promoter.

(A) 293T cells were transfected with the IL22-Luc reporter construct with empty vector, (NC), p50, p65, or co-transfected with both p50 and p65 vectors (p65/50) (n=4). (B) 293T cells were transfected with the IL22-Luc reporter construct or IL22-Luc mutation 1, IL22-Luc mutation 2 and IL22-Luc double mutation together with p65 vector (n=3). Linear mixed model was used to compare the four mutation groups by taking into account the block factor experiment. P value was adjusted for multiple comparisons by Holm’s procedure. A P value of 0.05 or less was considered significant. Ratio of each protein versus “nc” was log2 transformed before entering the linear mixed model. * denotes p < 0.05, ** denotes p < 0.01. Data are expressed as mean ± SEM.

RORγt⁺ ILC3s reside near IL-18 producing dendritic cells (DCs) in human SLT

ILC3s have previously been defined by IHC procedures as lymphoid-shaped CD117⁺ cells residing within the interfollicular and lamina propria regions of the human tonsil (10, 25, 38). To refine our identification of ILC3s by IHC, we defined ILC3s as RORγt⁺CD3⁻ cells with lymphoid morphology. We hypothesized that CD11c⁺ DCs within the human tonsil might be a physiological source of IL-18 for ILC3s. We observed that CD11c⁺IL-18⁺ cells with DC morphology were present in the interfollicular areas proximal to clusters of RORγt⁺ cells with lymphoid morphology (Fig. 7A, 7B). Most of the latter lymphoid cells co-expressed CD3, consistent with Th17 cells; however, occasional RORγt⁺CD3⁻ cells were also present, consistent with ILC3s (Fig. 7C). Serial sections were used in Fig. 7 to demonstrate the proximity of RORγt⁺CD3⁻ cells with IL-18 expressing cells. This shows that CD11c⁺IL-18⁺ DCs exist in close proximity to ILC3s in the interfollicular areas and may serve as a source of IL-18 in situ.

Figure 7. ILC3s reside in proximity to IL-18⁺CD11c⁺ DCs within the human tonsil.

Immunohistochemical staining depicted in sections of human tonsil. IF = interfollicular region, GC = germinal center. Original magnifications are 100× (top row) and 400× (middle and bottom rows). Sections B and C are consecutive, or serial, sections. (A) Brown indicates CD11c staining, red indicates IL-18 staining. White arrows indicate examples of co-stained cells. (B) Brown indicates RORγt staining and red indicates IL-18 staining. White arrows indicate examples of RORγt+ cells associating with IL-18+ cells. (C) Brown indicates RORγt staining and red indicates CD3 staining. White arrows indicate examples of RORγt+CD3- cells.

ILC3s can also be phenotypically defined by expression of CD117 and the absence of mast cell tryptase (MCT). Therefore, we also sought to use CD117⁺/MCT⁻ cells of lymphoid morphology as another way to identify ILC3s by IHC. Most of the CD117⁺ cells co-stained with MCT, but we did identify several CD117⁺MCT⁻ cells (Supplemental Fig. 4A–D). The majority of these CD117⁺MCT⁻ cells stained weakly for CD117 compared to double positive cells. This is consistent with the expectation that CD117 expression on ILC3s is lower than on mast cells. Furthermore, weakly staining CD117⁺ cells were smaller and more consistent with lymphoid morphology than strongly staining CD117⁺ cells, which were predominantly larger in size. We similarly observed that IL-18⁺ cells were found proximally to CD117⁺ cells in the interfollicular region (Supplemental Fig. 4E). IL-18⁺ cells were found in close proximity to RORγt⁺, CD3⁺, and MCT⁺ cells, in part due to the abundance of IL-18⁺ cells (Fig. 7B, Supplemental Fig. 4F and 4G). However, RORγt⁺ and MCT⁺ cells were mutually exclusive (Supplemental Fig. 4H).

These data support the notion that ILC3s co-localize with IL-18⁺ DCs in the interfollicular regions and that CD11c⁺ DCs are a primary source of IL-18 protein in human SLTs. Together, these findings suggest that CD11c⁺ DCs are capable of stimulating RORγt⁺ ILC3s resulting in quantitative expansion and production of IL-22.

Discussion

Human ILC3s play critical roles in antimicrobial immunity and homeostasis. As such, it is important to define the mechanisms by which these cells proliferate and regulate cytokine production. Of the thirteen cytokine conditions tested, only one (IL-18) resulted in significant expansion of ILC3s when cultured in combination with IL-15, a survival factor for ILC3s (25). Interestingly, IL-18 was able to override the anti-proliferative effects of TGF-β.

IL-18 is an important regulator of ILC function. Group 1 ILCs—ILC1s and NK cells—produce interferon-gamma (IFN-γ) in response to IL-18 stimulation (39–41). Furthermore, IL-12 stimulation induces the IL-18 receptor on NK cells while IL-18 reciprocally induces the IL-12 receptor (39, 42–45). This is hypothesized to be the mechanism by which the combination of IL-12 and IL-18 stimulation synergizes to activate NK cells. Recent work in a murine rotavirus model showed that IL-18, IL-22, and ILCs are involved in clearing infection following flagellin challenge (19).

In this study we show that human ILC3s are responsive to IL-18 stimulation. In addition to driving ILC3 proliferation, we found that IL-18 can stimulate and sustain production of IL-22. IL-23 has been reported to stimulate IL-22 production in human (9) and murine (46) ILC3s in the short term (<5 days), however, this effect does not appear to be sustained in human ILC3s (after 14 days), nor does IL-23 drive ILC3 proliferation (25). Our earlier work showed that IL-1β induced ILC3 homeostasis (25), which is consistent with the data presented here, as both cytokines belong to the same cytokine superfamily and are mediated by similar signaling pathways. However, here we also provide a molecular mechanism (the NF-κB signaling pathway) by which IL-18 induces the production of IL-22, and speculate that a comparable pathway exists for ILC3 activation by IL-1β.

NF-κB mediates downstream processes induced by IL-18 in other immune cell types, such as B cells (47). Furthermore, B cells reside in secondary lymphoid tissues, and therefore may be able to receive IL-18 from the same source as ILC3s. In our study, we found that IL-18⁺ DCs reside in close proximity to ILC3s in the tonsil. This suggests that DCs acting on ILC3s may regulate the ILC3 response to various pathogens. High DC production of IL-18 in human SLTs may expand the pool of ILC3s ready to help combat microbial infection. Together, our findings support a wealth of literature that suggests that dendritic cells have important and complex interactions with ILCs (19, 25, 42, 48, 49).

Excess IL-22 production, which can be exacerbated by the presence of IL-18, can contribute to immune-mediated disease and cancer, likely through its tissue regenerative effects (13, 50). IL-22 is associated with a variety of cancers, including malignancies of the skin, thyroid, lung, breast, stomach, pancreas, liver, cervix, and colon (13). ILC3s have been shown to accumulate in gastrointestinal cancers and promote tumor progression through IL-22 production (50, 51). IL-18 is an important inflammatory mediator and has also been associated with gastric cancer (52–54). In late stages of gastric cancer in particular, IL-18 can drive angiogenesis (52, 53), tumor proliferation (54), and tumor migration (53, 54). The role of IL-18 in inflammation is a double-edged sword. On one hand it can play important roles in activating immune cells to detect and kill pathogens (55, 56) and cancer cells (57). On the other hand, IL-18 can lead to immune dysfunction and contribute to a variety of diseases (58) including autoimmune disease (59–62), inflammatory bowel disease (63–67), cardiovascular disease (68–71), and cancer (72–74). Future studies may answer the question of whether these IL-18-associated diseases are mediated in part through the action of IL-18 on ILC3s and IL-22.

By identifying NF-κB as a positive transcriptional mediator of IL-22 production in ILC3s, we can now begin to explore the possibility of targeted therapy. Our findings suggest that NF-κB antagonists may be used in anti-tumor therapy not just to attack the tumor cells directly, but also to target inflammation by inhibiting the production of IL-22, a cytokine that can drive tumor growth (75, 76). Alternatively, IL-18 may be neutralized to prevent expansion of the ILC3 population and remove one of the drivers of IL-22 production. Furthermore, IL-22 production may be targeted more directly by a neutralizing antibody.

In summary, IL-18 cooperates with IL-15 to promote ILC3 proliferation and IL-22 production. We describe an IL-18-induced, NF-κB-mediated mechanism that regulates IL-22 in ILC3s. At steady state, IL-18 produced by DCs mediates IL-22 production by ILC3s to help maintain normal tissue integrity. In some disease states such as autoimmune diseases, IL-18 production may be induced in order to commensurately increase the ILC3 population able to produce IL-22, thereby accommodating the increased demands for tissue repair and host defense.