All work described herein was approved by the Institutional Animal Care and Use Committee at University of Nebraska Medical Center (UNMC). Il18bp^{tm1(KOMP)Vlcg} (IL-18BPKO, KOMP repository), Il18^tm1Aki (IL-18KO, Jackson), C57BL/6J [(IL-18KO controls) Jackson], and C57BL/6Tac [(IL-18BPKO controls) Taconic] mice used in these studies were derived from breeding colonies at theUNMC. IL-18BPKO mice were generated by a deletion of 1,573 bp starting at position 102,017,311 and ending at position 102,015,739 on chromosome 7. This deletion would effectively knockout known mouse IL-18BP isoforms c and d (17). Further information on the generation of the IL-18BPKO can be found here: http://velocigene.com/komp/detail/12770.

Single cell suspensions from whole spleens were created by cutting splenic tissue into small pieces and passing through 70 µm nylon screens in RPMI 1640 (HyClone) with 10% FBS (HyClone). Red blood cells were then lysed using ammonium chloride lysis buffer. Bone marrow was isolated from femurs by cleaving bone ends and using a 22 g needle with syringe to flush RPMI 1640 with 10% FBS through the bone. Marrow was disassociated by pipetting and red blood cells were lysed using ammonium chloride lysis buffer. Isolated cells were counted using a hemocytometer and resuspended in PBS, then labeled with Live/Dead Fixable Blue reagent (ThermoFisher) according manufacturer’s recommendation. Cells were then washed with PBS followed by flow cytometry staining buffer (FCSB: PBS with 0.75% BSA, 1 mM EDTA, and 0.05%NaN₃). Fc receptors were blocked using mixed rat and mouse IgG (ChromePure, Jackson Immunoresearch) followed by incubation with cocktails for antibodies targeting surface proteins for 30 min at 4°C. Cells were washed with FCSB, and when, appropriate, incubated with fluorochrome-conjugated streptavidin for 20 min at 4°C. This was followed by additional washes with FCSB, and cells were then fixed with 3% PFA for 30 min at room temperature. Following fixation, we washed cells twice before resuspension in FSCB prior to analysis. Cells were analyzed using a BD LSR II within 24 h of labeling and fixation.

For intracellular flow cytometry experiments, isolated splenocytes were cultured with RPMI 1640 (Hyclone) with 10% FBS (Hyclone), 2.05 mM l-glutamine (HyClone), 50 I.U./mL penicillin, 50 µg/mL streptomycin (Corning), and 50 µM β mercaptoethanol (Sigma) and stimulated with either 50 ng/mL IL-18 (MBL), 2 ng/mL IL-2 (Cell Sciences), both IL-18 and IL-2, 500 ng/mL LPS Escherichia coli 026:B6 (Sigma Aldrich), or nothing. Following overnight culture, splenocytes were treated with Cell Stimulation Cocktail (a mixture of phorbol 12-myristate 13-acetate (PMA), ionomycin, brefeldin A, and monensin, eBioscience) or brefeldin A and monensin (BioLegend) as control for 5 h. Splenocytes were surface labeled as described above. Following fixation and wash, cells were washed twice with permeabilization buffer (BioLegend) and then resuspended in permeabilization buffer and Fc receptors were blocked again as above. Permeabilized cells were then incubated with cocktails of antibodies for 30 min at 4°C, washed with permeabilization buffer, then washed with FCSB before final resuspension prior to analysis as described above. Flow cytometry data were analyzed using FlowJo (Treestar).

The following monoclonal antibodies and reagents were used: CD27 (LG.3A10, BV421 Biolegend); CD19 (6D5, Biotin BioLegend); Ly-6G (1A8, Biotin BioLegend); CD3 (17 A2, Biotin and FITC BioLegend); CD45 (30-F11, BUV395 BD Biosciences); TER-119 (TER-119, Biotin BioLegend); CD11b (M1/70, PerCP-Cy5.5 Biolegend and BV605 BD Biosciences); KLRG1 (2F1, PE BioLegend); NK1.1 (PK136, PE-Cy7 BioLegend); NKp46 (29A1.4, PE and APC BioLegend); CD122 (TM-β1, PE BioLegend); NKG2A/C/E (20d5 FITC eBioscience); Ly-49D (4E5 FITC BioLegend); Ly-49C/I (5E6, FITC BioLegend); Ly-49H (3D10, FITC BioLegend); Ly-49G2 (eBio4D11, FITC eBioscience); CD244.2/2B4 (eBio244F4, PE eBioscience); CD314/NKG2D (CX5, PE eBioscience); CD49b (DX5, PE BioLegend); TNF-α (MP6-XT22, PE BioLegend); IFN-γ (XMG1.2, eFluor450 eBioscience); IL-18Rα (112614, APC R&D Systems); streptavidin [BUV395 (BD Biosciences); and APC-Cy7 (BioLegend)]. Although not used to generate the data shown on IL-18Rα expression among NK cells, we observed extremely similar results during panel optimization when we compared IL-18Rα clone 112614 (R&D) with IL-18Rα clone BG/IL18RA from BioLegend.

Single cell suspensions from whole spleens from seven IL-18BPKO and eight C57BL/6Tac female mice age 16–20 weeks were created as described above. To generate cytosolic lysates of the splenocyte preparations, we used M-PER Mammalian Protein Extraction Reagent (ThermoScientific) with the addition of PMSF, Protease Inhibitor Cocktail III, and Phosphatase Inhibitor II (Research Products International). Following lysis, samples were spun at 21,000 × g for 10 min to pellet debris and protein concentration was calculated using Protein Assay Dye Reagent (BioRad) and then diluted into Laemmli buffer. Samples were heated to 95°C for 5 min, then 20 µg of lysate per lane of was loaded onto 15% polyacrylamide gels, separated, and then transferred onto nitrocellulose. Murine recombinant IL-18 (MBL) was used at 1 ng per lane as a positive control. Blots were blocked with 5% non-fat milk, then probed with IL-18 (1:1,000, Abcam, rabbit polyclonal ab71495). HRP-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch) was used at 1:20,000 and Clarity Western ECL Substrate (BioRad) was used prior to exposure. Blots were then stripped and reprobed with β tubulin antibodies (SantaCruz, rabbit polyclonal, H-235). Secondary antibody and ECL were used as for IL-18. IL-18 antibody specificity was confirmed using IL-18KO splenic lysates. While non-specific bands were present at ~32 kD, ~70 kD, and ~125 kD, a specific band of ~23 kD fitting with pro-form IL-18 was observed in wild-type (WT) splenic lysates but not among IL-18KO lysates (Figure S1 in Supplementary Material). Surprisingly, an alternate antibody for IL-18 (Santa Cruz, goat polyclonal sc-6179) produced non-specific bands in WT and IL-18KO splenic lysates at ~22 kD, which could easily be misconstrued for pro-form IL-18 (data not shown).

Whole spleens from IL-18BPKO, IL-18KO, and C57BL/6Tac mice were fixed in 10% neutral buffered formalin for 48 h prior to embedding with paraffin. All sections were cut to 4 µm thickness. Sections were then deparaffinized with xylene and rehydrated with gradated ethanol. For antigen retrieval, we used the 2100 Antigen Retriever (Aptum Biologics, Ltd.) with R-universal epitope recovery buffer (Electron Microscopy Sciences) according to manufacturer’s recommendation. During antibody validation and titration, we observed that for both IL-18 (Abcam, rabbit polyclonal ab71495) and IL-18BP (Santa Cruz, goat polyclonal S-19) detection, R buffer was superior to 10 mM sodium citrate for antigen retrieval (data not shown). Titration results demonstrated that ideal staining for IL-18 was at 2 µg/mL and IL-18BP was at 6 µg/mL (data not shown). Interestingly, four other polyclonal antibodies [IL-18BPd (AF122) and IL-18BPc (AF129, both R&D); IL-18BP (141012, US Biological Life Sciences); IL-18BP (Santa Cruz, H-61)] were tested for IL-18BP specificity but none produced specific staining histologically (data not shown).

Following antigen retrieval, we blocked endogenous peroxidases and phosphatases using Bloxall (Vector Labs), and blocked potential endogenous biotin with the Avidin/Biotin Blocking System (BioLegend). Sections were then blocked with Section Block (Electron Microscopy Sciences) to reduce non-specific binding of antibody.

For IHC, primary antibodies were diluted into 5% normal donkey serum (Equitech-Bio, Inc.) at the concentrations given above and applied to sections for 1 h at room temperature. After washing, we applied biotinylated donkey anti-goat IgG and biotinylated donkey anti-rabbit IgG (Jackson Immunoresearch). We then utilized the ABC Elite kit (Vector Labs) and visualized antibody localization using DAB Quanto (Thermo Scientific), both according to manufacturers’ recommendations. Finally, sections were counterstained with haemotoxylin, dehydrated and mounted with Cytoseal XYL (Thermo Scientific). Whole slides were scanned with a Coreo Au Slide Scanner (Ventana) and images from scanned slides were taken using the Ventana image viewer.

For IF, TrueBlack Lipofuscin Autofluorescence Quencher (Biotium) was used according to manufacturer’s recommendation to reduce tissue autofluorescence prior to the application of Section Block. Primary antibodies were applied in a 2-step process. IL-18BP was applied first as above. To amplify IL-18BP signal, we targeted the IL-18BP with chicken anti-goat IgG (Abcam), followed by AF488-conjugated donkey anti-chicken IgY (Jackson ImmunoResearch). Next, IL-18 (described above), Mac-3 (BD Biosciences, rat monoclonal M3/84, 1:100), and CD3 (Abcam, rabbit polyclonal ab5690, 1:100) antibodies were applied for 1 h at room temperature. Biotinylated donkey anti-rat IgG and anti-rabbit IgG (Jackson ImmunoResearch) and streptavidin-conjugated CF555 (Biotium) at 8 µg/mL were then used for detection. Sections were mounted with ProLong Diamond antifade mountant with DAPI (Molecular Probes). Photos were taken using a Zeiss AxioCam 503 mono camera attached to Zeiss Axio Imager.A2 using Zeiss Zen software. We merged and prepared images with ImageJ.

Mice were administered an i.p. injection of 35 µg LPS from E. coli 026:B6 (Sigma Aldrich) diluted into 500 µL saline. This quantity of LPS is in the range of the protective amount [see Ref. (37)]. After 8 h, submandibular blood and tissues were collected from each animal. Blood was collected using Goldenrod lancets (Medipoint, Inc.) in K₂ EDTA microtainers (BD Biosciences) and centrifuged at 10,000 × g for 5 min. Plasma was removed, aliquoted, and stored at −80°C until use. ELISAs for IL-18 (eBioscience) and IFN-γ (BioLegend) were performed according to manufacturers’ recommendations.

Female C57BL/6Tac and IL-18BPKO mice were injected i.p. with 1.5 mL 3% thioglycollate (Becton Dickinson). Four days post injection, animals were euthanized and their peritoneal cavities were flushed with 10 mL PBS twice. Recovered peritoneal exudate cells were pooled from three mice per genotype per experiment. Peritoneal exudate cells were washed two times with PBS, resuspended in complete RPMI 1640 (HyClone) containing 10% FBS (HyClone), 2.05 mM l-glutamine (HyClone), 50 I.U./mL penicillin, 50 µg/mL streptomycin (Corning), and 25 mM HEPES (HyClone), and then plated at 2 × 10⁶ per well in 12 well plates. Three hour post-plating, plates were washed two times with complete RPMI to remove non-adherent cells, thereby enriching for macrophage. Adherent peritoneal macrophage was then cultured with complete RPMI (control) or complete RPMI with 1 µg/mL LPS from E. coli 026:B6 (Sigma Aldrich). After 24 h, the supernatants were removed, centrifuged, aliquoted, and stored at −80°C until use. Plates were washed two times with complete RPMI to remove any residual cytokines, and then treated with 1 mL complete RPMI (control) or 1 mL complete RPMI with nigericin (Tocris) at varying molarity for 90 min. These supernatents were then removed, centrifuged, aliquoted, and stored at −80°C until analysis. ELISAs for IL-18 (eBioscience), IL-1β, and IL-6 (BioLegend) were performed according to manufacturers’ recommendations.

We performed statistical analysis and constructed figures with GraphPad Prism v6.03 (GraphPad Software, Inc.). Significance was determined using the following non-parametric tests: Mann–Whitney U test for comparisons of two groups, and Kruskal–Wallis test for comparison of three groups. p-Values less than 0.05 are considered to be statistically significant. Further information, including sample sizes and number of replicates is provided in the legends accompanying each figure. For Table 1, fold change was calculated using mean proportion following stimulation as dividend and mean proportion of freshly isolated NK cells as divisor. The resulting quotient was subtracted by 1. Thus, for example, a fold change of 0.8 represents an 80% increase from baseline.

Assuming that NK cells without the inhibition of IL-18BP could be abnormally polarized, we investigated NK cells from IL-18BPKO mice for maturational and functional differences. We identified NK cells using the gating strategy shown in Figure 1A. In brief, NK cells were defined as NK1.1⁺, lineage (CD3, CD19, and Ly-6G), living, singlet splenocytes. Cross-confirmation using anti-NKp46 antibodies revealed this approach successfully identified NK cells. Although splenocyte numbers were similar among IL-18BPKO and WT mice (Figure 1B), we observed both a decrease in proportion and number of NK cells among IL-18BPKO mice compared to WT (Figures 1C,D). Thus, it immediately became apparent that IL-18BP influences NK cell homeostasis.

Natural killer cell development can be delimited using the relative surface expression of CD27, CD11b, and killer cell lectin such as receptor G1 (KLRG1). The TNF receptor family member CD27 is expressed by immature NK cells, which have been shown to highly proliferative and be less cytotoxic than those lacking CD27 (38). CD11b (or integrin alpha-M) functions as a cellular adhesion molecule capable of binding ICAM or as a complement receptor for iC3b [reviewed in Ref. (39)]. CD11b expression increases with NK cell maturation and has been demonstrated to denote enhanced effector functions (40, 41). Finally, KLRG1 is an inhibitory receptor with cadherin binding capability [see Ref. (42)]. Expression of KLRG1 by NK cells denotes lower proliferation and turnover than among KLRG1⁻ NK cells (43).

Following this model of NK cell differentiation, we divided NK cells according to CD27, CD11b, and KLRG1 expression to determine if IL-18BP influences NK cell differentiation. We observed striking differences in NK cell maturation between the two genotypes (Figures 2A–C). Immature CD27⁺, CD11b⁻ NK cells were expanded in proportion and number among IL-18BPKO mice compared to controls (Figure 2B), while CD27⁺, CD11b⁺ NK cells were relatively similar between the IL-18BPKO mice (data not shown). Interestingly, CD27⁻, CD11b⁺ NK cells were significantly reduced in proportion and number among IL-18BPKO mice (Figure 2C). There was no difference observed among CD27⁻, CD11b⁻ NK cells (data not shown). Accordingly, KLRG1-expressing NK cells [being chiefly CD27⁻, CD11b⁺ among both WT and IL-18BPKO (Figure 2D)] were also reduced in overall number among IL-18BPKO mice (Figure 2E). These results demonstrate that NK cells in IL-18BPKO mice display disrupted maturation. The population of less mature, highly proliferative CD27⁺, CD11b⁻ NK cells is expanded, while mature CD27⁻, CD11b⁺ NK cells, including KLRG1-expressers, are substantially reduced.

At the same time as the splenic analysis, we investigated bone marrow for NK cell differences. Gating and identification of bone marrow NK cells are shown in Figure S2A in Supplementary Material. Interestingly, unlike in the spleen, overall bone marrow NK cell proportion was slightly elevated among IL-18BPKO mice and yet numbers of total NK cells were not significantly different (Figures S2B–D in Supplementary Material). However, bone marrow NK cells from IL-18BPKO mice were similarly disrupted as we observed in the spleen (Figures 3A–C). CD27⁺, CD11b⁻ cells were increased in proportion and number while CD27⁻, CD11b⁺ NK cells were reduced (Figures 3B,C). Thus, NK cells in the bone marrow are also impacted by IL-18BP deficiency. In combination, our splenic and bone marrow data reveal that IL-18BP has a profound effect on NK cells. In its absence, immature NK cells are expanded in lymphoid tissue, while mature NK cells are reduced. Furthermore, total splenic NK cell abundance is reduced with presumably unmitigated IL-18 signaling and it may be that the slight elevation of total NK cells that we observed in the bone marrow is to compensate for peripheral reductions.

In an environment with freely accessible IL-18, it is likely that IFN-γ responses could be amplified among NK cells. However, our phenotypic data revealed that normal NK cell maturation is impacted in the absence of IL-18BP. Thus, deranged cytokine output could be expected. To investigate this, we isolated splenocytes from IL-18BPKO and WT mice and stimulated them with IL-18, IL-18 with the prosurvival cytokine IL-2, or indirectly via LPS. Following overnight incubation with the stimulants described above, we then treated NK cells with PMA and ionomycin to elicit robust cytokine production. Splenocyte cultures were then labeled for IFN-γ and TNF-α expression and analyzed by flow cytometry.

Surprisingly, we observed that NK cells from IL-18BPKO were polarized to greater TNF-α production rather than IFN-γ production (Figures 4A–E). Indeed, the proportion of total IFN-γ⁺, TNF-α⁻ IL-18BPKO NK cells was reduced following all stimulations as well as among unstimulated splenocytes (Figure 4B). Alternatively, IFN-γ⁺, TNF-α⁺, and IFN-γ⁻, TNF-α⁺ NK cells were increased among NK cells from IL-18BPKO mice (Figures 4C,D). Interestingly, among unstimulated and indirectly stimulated (LPS) NK cells, the proportion of IFN-γ−, TNF-α⁻ NK cells was also increased among IL-18BPKO mice (Figure 4E), suggesting an expanded, undifferentiated compartment.

The outcomes of these stimulations demonstrate that not only does IL-18BP influence NK cell maturation and abundance, it also directs NK cell effector status. TNF-α, in addition to being a potent agent of cell death, is a stimulatory factor capable of inducing NFκB signaling (44). Thus, the absence of IL-18BP generates NK cells that are uniquely polarized and capable of enhanced TNF-α release that could lead to heightened cytotoxicity or inflammation.

In an environment of mismanaged IL-18 signaling brought about by the removal of IL-18BP, immature NK cells are increased while mature NK cells are decreased. This could indicate that IL-18 responsiveness among NK cell subsets is differentially regulated, perhaps by IL-18R expression. The IL-18R is a heterodimer comprised of alpha and beta chains, with IL-18Rα augmenting ligand binding within the dimer, while IL-18Rβ functions in signaling (4–7). Previous reports of murine NK cell surface expression of IL-18R are not in agreement and none analyzed NK cell subsets (11–13). It is plausible that NK cell subsets possess varying levels of IL-18R, which would suggest enhanced or diminished responsiveness to IL-18. Such receptor variance could help elucidate some of the maturational differences we have observed.

To investigate if IL-18R expression varies with NK cell maturation, we first examined NK cells from WT and IL-18BPKO mice for expression of IL-18Rα (four experiments, six mice per genotype). Our results confirmed that NK cells are mostly (>95% in WT) IL-18Rα⁺ (Figure 5A), in line with one previous report (11). However, we observed the presence of an IL-18Rα⁻ CD27⁺, CD11b⁻, NK cell subset among WT and IL-18BPKO mice, and these IL-18Rα⁻ NK cells were also proportionally expanded in IL-18BPKO mice (Figure 5A). Importantly, IL-18Rα⁻ NK cells were also observed in the bone marrow as well as among IL-18KO mice and their respective WT controls (data not shown). Once upregulated, IL-18Rα-expression appears greatest among NK cells with the highest CD27 expression and declines slightly with NK cell maturity (Figure 5A). Since the contraction of NK cells observed in the absence of IL-18BP is chiefly among CD27⁻, CD11b⁺ NK cells (where we observe slightly lower IL-18Rα expression), the expression of patterns of IL-18Rα alone cannot fully explain the observed difference in IL-18BPKO mice.

IL-18Rα⁻ NK cells reside within the CD27⁺, CD11b⁻ compartment, which has been shown to harbor immature NK cells that give rise to terminally differentiated subsets (38, 40). We hypothesized that the expression of IL-18Rα occurs during maturation and that the IL-18Rα⁻ NK cell subset is actually antecedent to IL-18Rα⁺ NK cells. A paradigm for NK cell maturation has been established using CD27 and CD11b (40). Therefore, we incorporated IL-18Rα expression into this model. We designated NK cells as IL-18Rα⁻, CD27⁺, CD11b⁻ (NK1); IL-18Rα⁺, CD27⁺, CD11b⁻ (NK2); IL-18Rα⁺, CD27⁺, CD11b⁺ (NK3); and IL-18Rα⁺, CD27⁻, CD11b⁺ (NK4) in a maturational continuum (Figure 5B). We then analyzed these subsets for expression of relevant NK cell surface proteins (three experiments, one IL-18BPKO, and one WT per experiment).

All NK cell subsets were positive for NKp46, 2B4, NKG2D, IL-2Rβ, and CD49b (see Figure S3A in Supplementary Material), confirming an overall NK cell phenotype for this novel IL-18Rα⁻ compartment. Furthermore, our results indicate that the hypothesized NK1 → NK4 maturation follows similar trends among WT and IL-18BPKO mice (Figure 5C) NKp46, NKG2D, and 2B4 had highest expression among NK1 sloping downward with maturation. Alternatively, IL-2Rβ and CD49b were lowest among the NK1 subset before plateauing and decreasing with maturation. Interestingly, high CD49b expression has previously been shown to be late in the maturation of NK cell precursors and following acquisition of NK1.1 (45).

We also examined NK cell subsets for expression of activating and inhibitory receptors of the Ly49 family and NKG2A/C/E (46, 47). For representative gating, see Figure S3B in Supplementary Material. Again, similar maturation trends were observed among WT and IL-18BPKO mice for Ly49C/I, Ly49H, Ly49G2, and Ly49D (Figure 5D). However, we observed that the NK1 subset among IL-18BPKO mice had lower proportions of NKG2A/C/E expressers than WT mice, although what drives this deficiency is unclear. Nevertheless, the combined phenotypic data characterize a previously undescribed NK cell subset lacking IL-18Rα expression. These seemingly immature IL-18Rα⁻ NK cells contribute to (but do not fully account for) the expansion of CD27⁺, CD11b⁻ cells observed in IL-18BPKO mice.

The absence of IL-18BP results in a highly modified NK cell compartment. Presumably, without the inhibition provided by IL-18BP, IL-18 signaling is unabated. It is previously been demonstrated that IL-18 signaling promotes the production of IFN-γ which in turns drives IL-18BP production, thereby providing a negative feedback loop during times of enhanced mature IL-18 release. It is unknown how steady state circulating IL-18 levels may be impacted with the removal of a key component of this signaling dynamic. To answer this question, we measured IL-18 levels among male and female IL-18BPKO (N = 23) and WT (N = 21) animals. Median IL-18 concentration in the WT cohort was 177.4 pg/mL, but we were unable to detect IL-18 in the plasma of IL-18BPKO mice (Figure 6A). This surprising finding could be explained by uncontrolled IL-18 uptake, reduced IL-18 production, and/or compromised IL-18 release among IL-18BPKO mice. The source(s) of circulating steady-state IL-18 have not been established. However, since we had already observed splenic NK cell differences, we targeted the spleen for our inquiry into IL-18 production and expression. Being home to DCs, resident macrophage, and a ready pool of monocytes, we believed we would be able to detect IL-18 perturbations, if they were apparent.

We first examined IL-18 distribution among splenic sections from WT (N = 10) and IL-18BPKO (N = 10) mice. Antibody specificity was determined using IL-18KO tissue (Figure 6B). Histological results revealed similar patterns of IL-18 distribution among splenic tissue (Figure 6B). IL-18 was richly associated with red pulp macrophage, yet also localized to the marginal zone as well as the interior of the white pulp among both IL-18BPKO and WT mice. These results implied that IL-18 is produced and distributed normally in IL-18BPKO mice.

Although our histological data provided evidence that IL-18 is present in equal abundance in tissue, we also examined splenic lysates. It is previously been shown that murine splenic lysates possess detectable IL-18 in the pro-form (48). We similarly observed pro-form IL-18 (~23 kD) in equivalent abundance from IL-18BPKO and WT mice (Figure 6C). We observed no cleaved mature IL-18 among our lysates from either genotype of mouse. These data provided additional evidence that IL-18 production is comparable between WT and IL-18BPKO mice and that it exists chiefly as the pro-form.

Release of mature IL-18, while incompletely understood, is thought to require two steps: priming and activation (49, 50). To determine if IL-18BPKO mice were capable of releasing mature IL-18, we cultured thioglycollate-elicited peritoneal macrophage (TEM) in the presence or absence of LPS (priming) and then followed with nigericin (activation). Nigericin, a pore-forming potassium ionophore, alters cellular potassium levels and activates the NLRP3 inflammasome, leading to caspase 1 activation and subsequent IL-18 maturation and release (51, 52).

With this approach, we observed no deficiency in IL-18 release among IL-18BPKO macrophage exposed to increasing concentrations of nigericin (Figure 6D). Importantly, the IL-18 ELISA shows less than 10% cross reactivity with pro-form IL-18 (manufacturer’s communication). Thus, the measured IL-18 levels are not unduly influenced by passive pro-form IL-18 release from necrotic cells. In addition, we found that IL-18BPKO and WT macrophage produced comparable amounts of IL-1β following nigericin treatment, as well as comparable IL-6 amounts following the initial LPS stimulation (Figure S4 in Supplementary Material). From these results, we conclude that IL-18BPKO TEM responded normally to established stimuli. In total, the analysis of splenic tissue and macrophage responses indicate that there is no deficiency in IL-18 production or release in IL-18BPKO mice and that primary macrophage respond relatively similarly between the two genotypes. Therefore, the severe reduction of IL-18 in the plasma may be resulting from uncontrolled IL-18 uptake or some form of IL-18 mismanagement in the absence of the IL-18BP.

The absence of IL-18 has been associated with obesity in mice that is observable between 20 and 24 weeks of age (53, 54). Since we were unable to observe circulating IL-18 in IL-18BPKO mice, we reasoned that IL-18BPKO mice may be prone to obesity. This could indicate that IL-18 signaling is obstructed without IL-18BP. We followed WT (N = 5) and IL-18BPKO (N = 6) female mice until at least 24 weeks of age in order to observe any change is weight gain (Figure 7A). While we cannot rule out spontaneous obesity at much later time points, we did not observe differences between WT and IL-18BPKO mice that were comparable to published reports on the absence of IL-18. This suggests that the absence of circulating IL-18 observed with IL-18BP deficiency does not induce metabolic differences similar to those seen in IL-18KO mice, implying functional IL-18 signaling remains intact.

LPS challenge leads to increased circulating IL-18 as well as increased IL-18-dependent IFN-γ production (34, 37). However, in an inflammatory environment lacking IL-18BP, amplified and uncontrolled IL-18 signaling could result in heightened IFN-γ responses. To test this hypothesis, we administered LPS to WT (N = 9), IL-18BPKO (N = 10), and IL-18KO (N = 9) female mice and recovered plasma 8 h following injection. We then tested for circulating IL-18 and IFN-γ levels. Among WT mice, circulating IL-18 levels were substantially increased from the steady-state conditions shown above and were significantly higher than among IL-18BPKO mice (Figure 7B). Indeed, in 9 out of 10 IL-18BPKO mice, circulating IL-18 was again not detectable. By contrast, we observed grossly elevated IFN-γ levels among IL-18BPKO mice compared to WT (Figure 7C). Interestingly, the IFN-γ level was highest (174.3 ng/mL) from the single IL-18BPKO mouse (denoted by black arrow in Figures 7B,C) in which we detected circulating IL-18 (1235.8 pg/mL). By contrast, IL-18KO mice produced significantly lower IFN-γ than either WT or IL-18BPKO animals following LPS challenge. These data indicate that IFN-γ responses are magnified in the absence of IL-18BP, further indication of functional, yet uncontrolled, IL-18 signaling.

Mice lacking IL-18BP have severely reduced quantities of circulating IL-18, as well as aberrant proportions of NK cell subsets. To gauge how the absence of IL-18 may influence NK cell maturation according to the CD27 and CD11b paradigm, we analyzed splenic and bone marrow NK cells from female IL-18KO (N = 4) and WT mice (N = 4). We observed no difference in proportion or number of NK cells or NK cell subsets among IL-18KO mice (Figure 7D and data not shown). These data demonstrate that IL-18 is not required for normal NK cell maturation. Thus, NK cell differences in IL-18BPKO mice cannot be explained by insufficient or absent IL-18 signaling.

The absence of IL-18BP creates skewed abundance and functional differences among NK cells, including an expanded subset of IL-18Rα⁻ NK cells. We also observed a severe reduction in circulating IL-18 among IL-18BP-deficient animals, but we found no defect in IL-18 expression or release. Furthermore, we found that animals lacking IL-18 possess phenotypically normal NK cells. In the absence of IL-18BP, it can be assumed that available IL-18R heterodimers present on the plasma membrane of leukocytes (e.g., neutrophils, basophil, NK cells, T cells, and DCs) would be freely available for unimpeded IL-18 signaling. Such signaling could be responsible for NK cell alterations we observed. IL-18 has been reported to both enhance or be detrimental to NK cell survival (11, 55) and it is currently unknown how IL-18 signaling may impact proportions of NK cell subsets. To gauge if/how IL-18 signaling alters abundance of CD27 and CD11b subsets, we cultured splenocytes from WT and IL-18BPKO mice for 24 h in the presence of IL-18.

The data from the stimulations are shown in Figure 8 with a fold change summary in Table 1. Compared to proportions immediately after isolation (fresh), short-term culture with IL-18 resulted in increased proportions of CD27⁺, CD11b⁻ NK cells with reduced proportions of CD27⁻, CD11b⁺ NK cells in both WT and IL-18BPKO cultures. Interestingly, while proportions of CD27⁺, CD11b⁺ NK cells were elevated in the presence of IL-18 among WT cultures, they were somewhat reduced among IL-18BPKO cultures. From these results, it appears that abundant free IL-18 induces and produces proportional changes similar to what we observed with IL-18BPKO mice, that is, elevated CD27⁺, CD11b⁻ proportions with concomitant reductions in CD27⁻, CD11b⁺ proportions.

Intriguingly, results differed in presence of IL-2. While IL-2 alone did not alter CD27⁺, CD11b⁻ proportions compared to fresh, IL-2 with IL-18 reduced the proportions of CD27⁺, CD11b⁻ NK cells in both WT and IL-18BPKO cultures, opposite to what was observed with IL-18 alone. Among WT cultures, IL-2 stimulation with or without IL-18 resulted in increased CD27⁺, CD11b⁺ and reduced CD27⁻, CD11b⁺ proportions, similar to IL-18 alone. Finally, while IL-18BPKO cultures responded similarly for CD27⁺, CD11b⁺ proportions, the CD27⁻, CD11b⁺ proportions remained relatively unchanged. The combined results indicate that IL-18 as well as IL-2 are capable of shifting proportions of NK cells subsets. Furthermore, there are unique responses to identical stimuli among CD11b⁻ expressing NK cells subsets from WT and IL-18BPKO mice.

Interleukin-18 binding protein plays a profound role in NK cell development and function. However, the identity and localization of IL-18BP-expressing cells in lymphoid tissue is unclear. IL-18BP transcripts have been reported in the murine spleen (18). Based on these findings and the results from our exploration of IL-18 distribution, we presumed that a similar pattern of distribution would be seen with IL-18BP. To test this, we compared IL-18BP expression between WT and IL-18BPKO mice. Surprisingly, IL-18BP localized to the splenic white pulp among cells having a dendritic morphology (Figure 9A), but generally lacking from the splenic red pulp where we had observed substantial IL-18 expression. These data suggest that IL-18 and IL-18BP are not equivalently expressed among splenocytes and points to unique IL-18 management specialization by a DC or macrophage.

To further investigate splenic IL-18BP in situ, we performed IF experiments on wild-type mice to establish localization, lineage, and co-expression of IL-18. As suggested from the IHC data, IL-18BP-expressing splenocytes were generally localized to the T cell rich, CD3-expressing zones of the splenic white pulp (Figure 9B). Furthermore, IL-18BP-expressing cells were also found to express Mac-3 (lysosome-associated membrane glycoprotein 2, CD107b), which denotes a DC or macrophage (Figure 9C). Finally, we observed that IL-18BP-expressing cells in splenic white pulp also expressed IL-18 (Figure 9D). Combined with data above on IL-18 distribution within the spleen (Figure 6B), these results generate further questions as to the source(s) of IL-18BP and IL-18 that appears to be crucial in normal NK cell differentiation within the spleen. Either the IL-18BP-expressing cells identified have the unique role of attenuating IL-18 signal throughout the spleen, or this is governed by an extrinsic cellular source and is conveyed to the spleen via the blood stream.