Production of IL-18 Binding Protein by Radiosensitive and Radioresistant

Production of IL-18 Binding Protein by Radiosensitive and Radioresistant Cells in CpG-Induced Macrophage Activation Syndrome This information is current as of October 4, 2021. Mathilde Harel, Charlotte Girard-Guyonvarc'h, Emiliana Rodriguez, Gaby Palmer and Cem Gabay J Immunol published online 10 July 2020 http://www.jimmunol.org/content/early/2020/07/09/jimmun

ol.2000168 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2020/07/10/jimmunol.200016 Material 8.DCSupplemental http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on October 4, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2020 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Published July 10, 2020, doi:10.4049/jimmunol.2000168 The Journal of Immunology

Production of IL-18 Binding Protein by Radiosensitive and Radioresistant Cells in CpG-Induced Macrophage Activation Syndrome

Mathilde Harel,*,† Charlotte Girard-Guyonvarc’h,† Emiliana Rodriguez,*,† Gaby Palmer,*,†,1 and Cem Gabay*,†,1

IL-18 binding protein (IL-18BP) acts as a naturally occurring IL-18 decoy receptor. If the balance between IL-18 and IL-18BP is dysregulated, abnormal levels of free bioactive IL-18 are detected, such as in the sera of Il-18bp knockout (KO) mice with CpG- induced macrophage activation syndrome. To determine the cellular sources of Il-18bp in vivo, we selectively depleted Il-18bp expression in either radiosensitive or radioresistant cells using bone marrow transfer between wild-type (WT) and Il-18bp KO mice. Following repeated CpG injections, Il-18bp KO (donor)→ Il-18bp KO (recipient) chimeric mice exhibited more severe disease, with an enhanced Ifn-g signature and circulating free Il-18 levels, in comparison with WT→WT chimeras. Interestingly, Downloaded from the phenotype of KO→WT and WT→KO mice did not differ from that of WT→WT mice. Consistent with this ﬁnding, serum Il- 18bp levels were similar in these three groups of mice. The contribution of radioresistant and radiosensitive cells to Il-18bp production varied markedly according to the organ examined, with a major contribution of radiosensitive cells in the spleen as opposed to a major contribution of radioresistant cells in the lung. Finally, Ifn-g blockade abrogated the CpG-induced but not the constitutive Il-18bp production. Our results demonstrate that circulating Il-18bp is induced in response to Ifn-g during CpG- induced macrophage activation syndrome and is present at high levels in the circulation to prevent the deleterious systemic effects http://www.jimmunol.org/ of Il-18. The Journal of Immunology, 2020, 205: 000–000.

nterleukin-18 is a proinflammatory cytokine that belongs to when considering the pathogenic role of IL-18 in inflammatory the IL-1 family and has initially been described as an IFN-g– conditions (7–12). I inducing factor (1–6). IL-18 activity is regulated by IL-18 Hemophagocytic lymphohistiocytosis (HLH) or hemophago- binding protein (IL-18BP), a naturally occurring soluble inhibitor cytic syndromes correspond to rare, life-threatening, hyper- that makes high-affinity complexes with IL-18, thus preventing its inflammatory conditions with several causes, symptoms, and binding to its cell surface receptors. IL-18BP is present in large outcomes. Traditionally, HLH is divided into primary and sec- by guest on October 4, 2021 concentrations in the circulation. As a consequence, circulating ondary forms that share various clinical and laboratory features. In levels of unbound, free IL-18 are undetectable in healthy indi- the primary forms of HLH or familial HLH (fHLH), an impairment viduals. Furthermore, rather than total IL-18 levels, the IL-18/IL- in cytolytic activity provoked by autosomal recessive disorders 18BP balance and the concentrations of free IL-18 are relevant associated with viral infections causes the syndrome. In contrast to fHLH, the secondary form, also termed macrophage activation syndrome (MAS), is a complex disease with probable influences of multiple genes and environmental factors (13–16). MAS can occur *Department of Pathology and Immunology, School of Medicine, University of as a life-threatening complication of inflammatory diseases, such † Geneva, CH-1211 Geneva 4, Switzerland; and Division of Rheumatology, Depart- as systemic juvenile idiopathic arthritis (sJIA), as well as fol- ment of Medicine, University Hospitals, CH-1211 Geneva 14, Switzerland lowing viral infections and in patients with cancer (17). The 1G.P. and C.G. contributed equally to this work. mechanisms responsible for the pathogenesis of MAS are unclear, ORCIDs: 0000-0002-4901-2059 (C.G.-G.); 0000-0001-7567-1798 (G.P.); 0000- 0001-6853-3063 (C.G.). but recently, Weiss et al. (18) showed an association between MAS and elevated serum levels of free IL-18. Received for publication February 19, 2020. Accepted for publication June 15, 2020. High levels of free IL-18 have been measured in patients with MAS This work was supported by Swiss National Science Foundation Grant 310030_712674, the Institute of Rheumatology Research, and the Rheumasearch secondary to sJIA (18). More importantly, a child with sJIA and Foundation. recurrent episodes of severe MAS was successfully treated by the Address correspondence and reprint requests to Prof. Cem Gabay, Division administration of recombinant human IL-18BP (tadekinig alfa) (19). of Rheumatology, Department of Medicine, University Hospitals of Geneva, Using a murine model of MAS induced by repeated CpG in- 26 Avenue de Beau-Se´jour, CH-1211 Geneva 14, Switzerland. E-mail address: [email protected] jections (20, 21), we recently showed that Il-18bp–deficient (Il- 2/2 The online version of this article contains supplemental material. 18bp ) mice developed a more severe form of the disease (22). In addition, serum levels of free Il-18 were detectable only in Il- Abbreviations used in this article: BM, bone marrow; fHLH, familial HLH; HLH, 2/2 hemophagocytic lymphohistiocytosis; IL-18BP, IL-18 binding protein; Il-18bp2/2, 18bp but not in wild-type (WT) mice following CpG stimu- 2 2 Il-18bp–deficient; KO, knockout; KO→KO, Il-18bp / BM transplanted into Il- lation and correlated with a more pronounced Ifn-g signature and 18bp2/2 recipient mice; KO→WT, Il-18bp2/2 BM transplanted into WT recipient mice; MAS, macrophage activation syndrome; pt-Il18bp, primary Il18bp transcript; clinical and biological signs of MAS. These findings support the sJIA, systemic juvenile idiopathic arthritis; WT, wild-type; WT→KO, WT BM trans- hypothesis that an imbalance between Il-18 and Il-18bp levels is 2/2 planted into Il-18bp recipient mice; WT→WT, WT BM transplanted into WT involved in the pathogenesis of MAS. recipient mice. Previous in vitro studies showed that Il-18bp is expressed by a Copyright Ó 2020 by The American Association of Immunologists, Inc. 0022-1767/20/$37.50 variety of cell types, but little is known regarding the regulation of

www.jimmunol.org/cgi/doi/10.4049/jimmunol.2000168 2 RADIOSENSITIVE AND RADIORESISTANT CELLS PRODUCE IL-18BP

Il-18bp production in vivo. Most importantly, several in vitro studies A noncommercial rabbit anti-mouse IL-18BP Ab provided by MAB Dis- showed that IL-18BP production is enhanced in response to IFN-g, covery GmbH (Neuried, Germany) served to coat 96-well plates (dilution thus acting as a negative feedback loop to dampen IL-18 activities in PBS to 1 mg/ml). Nonspecific binding was blocked with 1% BSA (Sigma-Aldrich) in PBS. The range of Il-18bp standard (recombinant (4, 7, 23–28). In this study, we show that both radioresistant and mouse Il-18bp isoform d; Creative BioMart) was from 20 to 0.156 ng/ml. radiosensitive cells produce Il-18bp during CpG-induced MAS but A goat anti-mouse Il-18bp isoform c–biotinylated Ab (Bio-Techne) was that their relative contribution varies according to the organ exam- used as the detection Ab. The samples and standards were diluted in PBS– ined. Furthermore, Il-18bp is produced as an Ifn-g signature gene in 1% BSA. In addition, to verify that this assay detects both the unbound form of Il-18bp as well as Il-18bp bound to mature Il-18, recombinant response to CpG injection but not in naive mice. mouse Il-18bp (3ng/ml in PBS, 1% BSA) was preincubated without or with recombinant mouse mature Il-18 at five different concentrations Materials and Methods ranging from 3 pg/ml to 30 ng/ml for 15 min before assessment of Il-18bp Mice and treatments concentrations. Il-18bp quantification was not affected by addition of Il-18, indicating that this ELISA recognizes both the uncomplexed form of Il- Frozen C57BL/6N Il18bptm1.1(KOMP)Vlcg (Il-18bp+/2) mouse embryos, 18bp and Il-18bp bound to mature Il-18 (data not shown). generated by the Knockout Mouse Project Repository (University of For both mature free Il-18 and Il-18bp measurements, after washing the California, Davis, Davis, CA), were purchased and implanted into the detection Ab, peroxidase-conjugated streptavidin (Jackson Immuno- uterus of foster females to obtain Il-18bp+/2 mice, which were maintained research Europe, Newmarket, U.K.) diluted in PBS–1% BSA was added. at Charles River Laboratories (L’Arbresle, France). Il-18bp+/2 mice were The colorimetric reaction was developed with 3,39,5,59-tetramethylbenzi- then bred in the conventional area of the animal facility of the Geneva dine and stopped by the addition of 2 N H2SO4 after 20 min. Plates were University School of Medicine (Geneva, Switzerland) to obtain Il-18bp2/2 read with a Dynatech MR 5000 Microplate Reader at an OD of 450 nm, and WT-cohoused littermates. Genotyping was performed on total DNA using 570 nm as the reference wavelength. extracted from ear biopsy specimens using a four primer PCR combining a primer pair specific for the WT allele (59-TCTTTCTTGTAGCCC- RNA extraction and quantitative RT-PCR Downloaded from CACTGCTAGG-39 forward; 59-ACTAAGTGGAGCTTGGCAGGA- Spleen, liver, and lung total RNA was extracted using TRIzol (Life tech- TAGG-39 reverse, amplified product 226 bp) and a primer pair specific for 2/2 nologies, Carlsbad, CA) and further purified on RNeasy columns (QIA- the Il-18bp allele (59-GCAGCCTCTGTTCCACATACACTTCA-39 GEN). cDNA was prepared from 500 ng of total RNA using the SuperScript forward; 59TTGCCTTGGGGACAGTACATTAGGG-39 reverse, amplified II Reverse Transcriptase (Life Technologies). Relative primary transcript product 624 bp) (22). C57BL/6 congenic B6.SJL-Ptprca Pep3b/BoyJ levels of Il18bp and mRNA levels of Il18, Il18bp, Ifng, Cxcl9, and CIIta (CD45.1) mice, carrying the differential CD45.1 allele of the pan leuko- were normalized to mRNA levels of the ribosomal large 60S subunit L32 cyte marker CD45, were obtained from The Jackson Laboratory. All ex- protein (Rpl32), used as housekeeping gene, using a comparative method http://www.jimmunol.org/ periments were approved by the Geneva cantonal authority for animal (2–DCt). Rpl32, Il18, Il18bp, pt-Il18bp, Ifng, Cxcl9, and CIIta cDNA experimentation (licenses GE/151/17 and GE/66/18) and performed on levels were estimated by quantitative PCR (40 cycles, annealing temper- sex- and 10- to-12-wk-old aged-matched animals. A class B phosphor- ature of 60˚C) using the iQ SYBR Green Supermix (Bio-Rad Laboratories othioate CpG 1826 oligonucleotide, agonist of TLR9, (59-TCCAT- AG, Cressier, Switzerland). The sequences of the primers used are listed in GACGTTCCTGACGTT-39, Eurofins Genomics GmbH, Ebersberg, Table I (Eurofins, Ebersberg, Germany). Non–reverse-transcribed RNA Germany) was injected i.p. at a dose of 2.5 mg/g body weight on days 0, 2, samples and water were included as negative controls. and 4 or once only. In some experiments, rat IgG1 anti–Ifn-g monoclonal (clone XMG1.2) or isotype control Abs (BioLegend, San Diego, CA) were Flow cytometric analysis administered i.p. at, respectively, 7.5 mg/g body weight or 500 mg/mouse 2 h before CpG injection. Monitoring of mice, assessment of clinical Spleen single-cell suspensions were obtained without prior digestion of readouts, and sample collection have been previously described (22). the organ and were treated with RBC lysis buffer (155 mM ammonium by guest on October 4, 2021 chloride, 1 mM potassium bicarbonate, and 77.5 mMEDTA[Sigma- Bone marrow transplantation Aldrich, Saint-Louis, MO]). Cells were then incubated in presence of rat anti-mouse CD16/CD32 Ab (Fc block; BD Pharmingen) and immu- For bone marrow (BM) transplantation studies, mice were lethally irradiated 137 nostained with mAbs specific for extracellular markers using a single (single 800 cGy radiation; gamma radiation, [ Cs] source). On the next mix, including the following: CD45.1–allophycocyanin (clone A20; BD 3 6 day, 2 10 RBC-lysed BM cells were injected retro-orbitally under Pharmingen), CD45.2-BV480 (clone 104; BD Horizon), CD3ε-FITC isoflurane (Attane, Piramal Critical Care) anesthesia. CD45.2+ WT BM + (clone 17A2; BioLegend), B220-BUV395 (clone RA3-6B2; BD Hori- nucleated cells were infused into lethally irradiated CD45.1 WT mice zon), CD4-PerCP Cy5.5 (clone RM4-5; BD Pharmingen), CD8-BV605 → + + (WT WT), CD45.1 WT cells were infused into CD45.2 WT mice (clone 53-6.7; BD Horizon), NK1.1–Alexa Fluor 700 (clone PK136; BD → + 2/2 + (WT WT), CD45.2 Il-18bp cells were infused into CD45.1 WT Pharmingen), CD11b-BUV737 (clone M1/70; BD Horizon), CD11c- → + mice (knockout [KO] WT), CD45.1 WT cells were infused into BV786 (clone HL3; BD Horizon), Ly-6C–eF450 (clone HK1.4; eBio- + 2/2 → + 2/2 CD45.2 Il-18bp mice (WT KO), or CD45.2 Il-18bp cells were science), and Ly-6G–BV711(clone 1A8, BD Horizon). Cell viability was + 2/2 → infused into CD45.2 Il-18bp mice (KO KO). All mice received an assessed with propidium iodide. Data were acquired on a BD LSRFor- antibiotic treatment (Baytril 10% – enrofloxacine; Bayer, Leverkusen, tessa analyzer (BD Biosciences, USA) and analyzed using the FlowJo Germany) from 3 d before to 24 d after irradiation to avoid infectious Analysis software (BD Biosciences, San Jose, CA). The gating strategy complications. Mice were allowed to recover for 8 wk before MAS in- is illustrated in Supplemental Fig. 1. duction by repeated i.p CpG injections at a dose of 2.5 mg/g body weight on day 0, 2, and 4 according the MAS induction protocol. Statistical analysis Laboratory tests and cytokine measurements Results are represented as individual values, except for variations from baseline body weights, which are expressed as the mean 6 SD. The Blood cell counts were assessed on peripheral blood collected in EDTA- Kruskal–Wallis test with Dunn posthoc test or two-way ANOVA with coated vials using a Sysmex blood cell counter (Sysmex, Yverdon, Swit- Holm–Sidak correction for multiple comparisons tests were used as indi- zerland). Plasma levels of Cxcl9 were measured by sandwich ELISA cated in the figure legends. The p values ,0.05 were considered signifi- according to the manufacturer’s protocol (DuoSet, DY492; R&D Systems, cant. Only statistically significant differences are shown. All plots and Abingdon, U.K.). Plasma-mature free Il-18 levels were determined by a statistical analyses were performed using GraphPad Prism 8 software previously validated ELISA (22). Briefly, the 96-well plates were coated (GraphPad Software, La Jolla, CA). Adobe Illustrator (Adobe, San Jose, with recombinant human IL-18BPa (AB2 Bio, Lausanne, Switzerland) at CA) was used to reformat the graphics and figures. 5mg/ml in PBS. Aspecific binding was blocked with 1% BSA (Merck Millipore, Darmstadt, Germany) in PBS–0.1% Tween 20. The range of the Results Il-18 standard curve was 1000–1.37 pg/ml (R&D Systems, Abingdon, U.K.). Tested samples and standards were added to the plate at a 1:5 di- Depletion of Il-18bp production in either radiosensitive or lution and incubated with biotinylated rat monoclonal anti-mouse Il-18 Ab radioresistant compartments does not result in severe MAS (MEDICAL & BIOLOGICAL LABORATORIES, Nagoya, Japan) in following repeated CpG injections 0.1 M HEPES (pH7), 0.35M NaCl, and 0.4% Triton X-100. The plasma levels of Il-18bp were determined with a homemade ELISA, In this study, we wanted to dissect the relative contribution of the specificity of which was validated using the plasma of Il-18bp2/2 mice. radiosensitive and radioresistant cells to Il-18bp production in an The Journal of Immunology 3 experimental model of MAS. We thus generated four groups of BM whereas the presence of circulating free Il-18 was detected in chimeras by transplanting WT BM into WT recipient mice KO→KO mice only (Fig. 2D). (WT→WT), Il-18bp2/2 BM into WT recipient mice (KO→WT), Il-18bp is produced by both radioresistant and radiosensitive WT BM into Il-18bp2/2 recipient mice (WT→KO), and Il-18bp2/2 cells in response to CpG BM into Il-18bp2/2 recipient mice (KO→KO) (Supplemental Fig. 2A). The use of WT and Il-18bp2/2 mice carrying differen- The relative contribution of radiosensitive and radioresistant cells tial CD45.1 or CD45.2 alleles of the CD45 pan leukocyte marker to Il-18bp production following CpG stimulation, as assessed by allowed for the distinction of leukocytes of donor and recipient measuring Il18bp mRNA levels in different organs, varied largely origin after the graft. Eight weeks after BM transfer, mice of all according to the organ considered. Indeed, Il18bp mRNA levels groups received CpG injections on days 0, 2, and 4 and were were markedly reduced in the spleen of KO→WT mice, consistent euthanized on day 7. Reconstitution efficiency was evaluated in with a major role of radiosensitive cells (80%) in Il-18bp pro- the spleen of WT→WT, KO→WT, and WT→KO mice at the end duction (Fig. 3A). In contrast, as shown in WT→KO mice, radi- of the experiment by evaluating the proportion of CD45.1+ or oresistant cells contributed to a large extent to the production of CD45.2+ donor cells among total CD45+ cells. When considering IL-18bp in the liver (65%) and lung (90%) (Fig. 3B, 3C). the total CD45+ cell population in the spleen, the efficiency of Most importantly, although Il18bp mRNA levels varied largely reconstitution was around 90% regardless of the allelic variance in the different organs when Il-18bp production was selectively and the genotype (Supplemental Fig. 2B). When analyzing the depleted in radiosensitive or radioresistant cells, circulating levels CD45+ cellular subtypes, we observed a lower efficiency of re- of Il-18bp did not differ between WT→KO or KO→WT and constitution for T cells: 75,4% compared with 98.8% for B cells WT→WT mice (Fig. 3D). Downloaded from and 96.2% for myeloid cells (Supplemental Fig. 2C). Induction of Il-18bp expression is Ifn-g dependent Similar to previously reported findings in Il-18bp2/2 mice (22), KO→KO chimeras exhibited severe clinical manifestations of MAS We have previously shown that Il-18bp expression is markedly after repeated CpG injections, including significant weight loss upregulated during CpG-induced MAS in the liver and weakly (Fig. 1A) and reduced RBC as compared with WT→WT mice, upregulated in the spleen (22). Because Ifn-g is known to stimulate whereas platelet counts tended to decrease (Fig. 1B). In contrast, the expression of Il-18bp in vitro (4, 7, 26–28), we decided to body weights and blood cell counts in WT→KO and KO→WT did determine whether Ifn-g is also involved in the increased Il-18bp http://www.jimmunol.org/ not significantly differ from WT→WT mice. The depletion of Il- production in response to CpG injection in vivo. The adminis- 18bp in both compartments also resulted in severe splenomegaly in tration of a monoclonal anti–Ifn-g Ab abrogated CpG-induced KO→KO mice (Fig. 1C). In contrast, and in agreement with pre- CIIta and Cxcl9 mRNA expression (Fig. 4A, 4B), thus confirm- vious observations in Il-18bp2/2 compared with WT mice (22), the ing the effectiveness of Ifn-g neutralization. Of note, baseline leukocyte counts (Fig. 1B) and liver weight (Fig. 1C) were com- spleen CIIta and Cxcl9 mRNA levels were also decreased by Ifn-g parable in all four groups of chimeric mice after CpG injections. blockade. Most importantly, CpG-stimulated Il-18bp levels in the Significant changes in spleen immune cell populations were ob- liver, lung, and serum were completely abrogated by the admin- → istration of the anti–Ifn-g Ab. In the spleen, no induction of Il18bp

served only in KO KO mice (Supplemental Fig. 3). Indeed, by guest on October 4, 2021 KO→KO mice exhibited an increased proportion of B cells as expression was observed following CpG stimulation. Consistently, compared with WT→WT mice, whereas the proportion of Ifn-g blockade was devoid of any effect on spleen Il18bp mRNA T lymphocytes decreased. Among T cells, the proportion of CD8+ levels. Finally, the injection of the anti–Ifn-g Ab had no effect on T cells increased, whereas the percentage of CD4+ T cells tended to baseline Il18bp mRNA expression in liver and lung or on Il-18bp decrease. There was also a significant decrease in NK cells and in serum levels (Fig. 4C, 4E). undefined CD11b+ cell population in KO→KO as compared with Increased steady-state levels of Il18bp mRNA might reflect WT→WT mice. enhanced transcription, higher stability of the transcript, or both. To examine whether Il18bp is transcriptionally regulated by Ifn-g, Increased Ifn-g production and Ifn-g molecular signature in we also measured nonspliced/primary Il18bp transcript (pt-Il18bp) → KO KO mice upon CpG injections levels by quantitative RT-PCR, using forward and reverse primers Ifn-g has been shown to play a critical role in the development of annealing, respectively, to the first exon and the first intron of the most of the pathological manifestations of CpG-induced MAS (21, Il18bp gene (Table I). Consistent with Il18bp mRNA levels, the 22, 26, 29). Therefore, we determined the transcript levels of Ifn-g levels of pt-Il18bp did not vary at baseline or following CpG in- and of two of its response genes, CIIta and Cxcl9, in the spleen, jection with or without Ifn-g blockade in the spleen. In contrast, in liver, and lung. KO→KO mice exhibited increased Ifn-g mRNA liver and lung, pt-Il18bp levels were upregulated following CpG levels in all three organs as compared with the three other groups stimulation and decreased upon injection of the anti–Ifn-g Ab of mice (Fig. 2A). Consistent with this result, CIIta mRNA levels (Fig. 4D). Without excluding potential additional posttranscrip- were significantly enhanced in KO→KO mice as compared with tional regulation, these observations indicate that Il-18bp is up- the other groups (Fig. 2B). Cxcl9 mRNA levels were also higher in regulated at the transcriptional level by Ifn-g in liver and lung but the spleen, liver, and lung of KO→KO mice than in the three other not in the spleen. groups of mice (Fig. 2C). However, in comparison with CIIta, the increase of Cxcl9 mRNA expression was globally less marked. Discussion Indeed, in the spleen, the increase in Cxcl9 mRNA levels reached Using BM transplantation experiments, we observed that depletion statistical significance only when KO→KO were compared with of Il-18bp from both the radioresistant and the radiosensitive WT→KO mice. Liver Cxcl9 mRNA levels were significantly in- compartments in KO→KO chimeras led to a more severe form of creased in KO→KO compared with WT→WT and KO→WT mice, MAS upon repeated CpG injections as compared with WT→WT but not in comparison with WT→KO mice (Fig. 2C). Of note, mice. This result is consistent with the increased disease severity, circulating levels of Cxcl9 were significantly higher in KO→KO enhanced Ifn-g signature, and elevated free Il-18 serum levels as compared with all other groups of mice. Il18 mRNA levels previously observed in Il-18bp KO mice (22). In contrast, selective in the different organs were identical in all the mouse groups, Il-18bp deficiency in either radiosensitive cells or in radioresistant 4 RADIOSENSITIVE AND RADIORESISTANT CELLS PRODUCE IL-18BP

FIGURE 1. Only KO→KO mice develop a severe CpG-induced MAS. Four groups of mice were →

created following BM transfer: WT WT, Downloaded from KO→WT, WT→KO, and KO→KO. Eight weeks after BM transfer, mice of all groups received CpG injections on days 0, 2, and 4 and were euthanized on day 7. All the results are represented as mean 6 SD for WT→WT (n = 15), KO→WT (n = 10), WT→KO (n = 10), and KO→KO (n = 9) groups.

Variations from baseline body weights are shown http://www.jimmunol.org/ (A). Blood was sampled before the experiments and at sacriﬁce, and RBC, platelet, and WBC numbers are shown. & , 0.01 versus before CpG for each group (B). Spleen and liver weights, normalized to total body weight, are shown (C). Each individual symbol represents one mouse. Statistical analysis was performed using a Krus- kal–Wallis test with a Dunn posthoc test (A and C) or a two-way ANOVAwith Holm–Sidak correction for multiple comparisons tests (B). by guest on October 4, 2021

cells did not result in a significant change of weight loss, spleen agreement with these observations, in the current study, Il18 enlargement, blood cell count, and Ifn-g molecular signature as mRNA levels in spleen, lung, and liver were similar in all groups compared with WT→WT mice. Consistent with these results, free of BM chimeras during CpG-induced MAS. In addition to the Il-18 levels remained undetectable in WT→WT, WT→KO, and regulation of its mRNA expression, IL-18 activity is controlled at KO→WT mice. Taken together, these findings indicate that both the protein level. Indeed, IL-18 is produced as a biologically inert radiosensitive and radioresistant cells contribute to the production of propeptide and requires inflammasome activation and proteolytic Il-18bp during CpG-induced MAS and play a critical role in the processing by caspase-1 to be converted into a bioactive mature regulation of excessive Il-18 signaling. form (30–33). Alternatively, inflammasome-independent process- We have previously shown that Il18 mRNA expression in spleen ing mechanisms for IL-18 cleavage have also been described. In and liver increases following repeated CpG injections but without particular, upon release from dying cells, the inactive form of IL- significant difference between WT and Il-18bp2/2 mice (22). In 18 becomes a target for various extracellular enzymes, such as The Journal of Immunology 5 Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 2. Ifn-g signature and free Il-18 levels are enhanced in KO→KO mice. Following BM transfer, mice received CpG injections on days 0, 2, and 4 and were euthanized on day 7. Spleen (left panels), liver (middle panels) and lung (right panels) transcript levels for Ifn-g (A), CIIta (B), Cxcl9 (C), and Il18 (D) were normalized to the housekeeping gene Rpl32. Plasma levels of Cxcl9 (E, left panel) and free Il-18 (E, right panel) were assessed by ELISA. Each dot represents one mouse. Horizontal and vertical bars represent mean 6 SD for WT→WT (n = 15), KO→WT (n = 10), WT→KO (n = 10), and KO→KO (n = 9) mice. Statistical analysis was performed using a Kruskal–Wallis test with a Dunn posthoc test. 6 RADIOSENSITIVE AND RADIORESISTANT CELLS PRODUCE IL-18BP

FIGURE 3. Both radiosensitive and radioresistant cells contribute to Il-18bp production during CpG-induced MAS. Following BM transfer, mice received CpG injections on days 0, 2, and 4 and were euthanized on day 7. Spleen (A), liver (B), and lung (C) Il18bp mRNA levels were normalized to the housekeeping gene Rpl32. Il-18bp plasma levels were determined by ELISA (D). Each dot represents one mouse. Horizontal and vertical bars represent mean 6 SD for WT→WT (n = 15), KO→WT (n = 10), WT→KO (n = 10), and KO→KO (n = 9) mice. Statistical analysis was Downloaded from performed using a Kruskal–Wallis test with a Dunn posthoc test. http://www.jimmunol.org/

chymase, meprin-b, and granzyme B, which are able to generate possible compensatory mechanisms explaining stable circulating biologically active forms of the cytokine (34–36). CpG does not levels of Il-18bp in chimeric mice. We did, however, not find any by guest on October 4, 2021 directly activate the inflammasome/caspase-1 pathway, and it is experimental evidence for this. Indeed, Ifn-g mRNA levels were still not clear by which mechanism(s) maturation of Il-18 is similar in WT→WT, KO→WT, and WT→WT chimeras. Fur- achieved during CpG-induced MAS. However, even assuming thermore, mRNA levels of Ifn-g response genes, such as CIIta and similar Il-18 maturation in all groups of mice, the buffering ca- Cxcl9 as well as Cxcl9 serum concentrations, were not different in pacity of Il-18bp is specifically lost in KO→KO chimeras, leading these three groups of chimeric mice. Based on these observations, to excessive Il-18 signaling and explaining the more severe phe- homeostasis of Il-18bp levels is unlikely to result from changes in notype observed in this group. Ifn-g signaling. A better characterization of Il-18bp–producing Using a newly developed ELISA to measure Il-18bp, we ob- cells will certainly improve our understanding of the direct or served that Il-18bp levels were already high in naive conditions indirect mechanisms and mediators governing Il-18bp production. (∼460 ng/ml) and further increased up to 2.2 mg/ml after a single Radiosensitive cells are usually associated to the hematopoietic CpG stimulation. These high Il-18bp concentrations are consistent compartment. Indeed, radiation induces lethal mutations primarily in with the absence of free Il-18 in the serum of both naive and CpG- mitotically active cells (37), hence cells of hematopoietic origin. In stimulated WT mice and supports the hypothesis that Il-18bp is contrast, stromal cells display a low rate of proliferation and are thus critical to prevent the development of systemic manifestations considered as radioresistant cells. However, this statement has some related to the presence of unbound Il-18 in the circulation. exceptions. Indeed, some stromal cells are partially radiosensitive, The relative contribution of radioresistant and radiosensitive such as keratinocytes and epithelial intestinal cells, which leads, re- cells to Il-18bp production in response to CpG stimulation varied spectively, to fur discoloration and diarrhea in irradiated mice. Also, markedly according to the organ examined. Indeed, most of the Il- some cells from the hematopoietic lineage are partially radioresistant, 18bp production in the spleen derived from hematopoietic cells, such as resident tissue macrophages and dendritic cells. Corbaz et al. whereas the situation was completely opposite in the lung. Sur- (24) showed that intestinal endothelial cells and macrophages pro- prisingly, despite major variations in Il18bp mRNA levels in duce IL-18BP in patients with Crohn’s disease. Based on this pre- KO→WT and WT→KO mice in the three organs investigated and vious publication, the general principle of radiosensitivity, and our thus major organ differences regarding the contribution of radio- results, we hypothesize that endothelial cells as well as liver and lung sensitive versus radioresistant cells to Il-18bp production, the se- resident macrophages, such as, respectively, Kupffer cells and alve- rum levels of Il-18bp were not significantly different in mice olar macrophages, may represent important cellular sources of Il- deficient for Il-18bp in either the radiosensitive or radioresistant 18bp in these organs. This is also consistent with the fact that these compartments as compared with mice with WT cells in both cell types are either completely or partly radioresistant (38–41). In compartments. These findings suggest a tight regulation of the addition, hematopoietic cells, which represent the major source of Il- global Il-18bp production to maintain constant Il-18bp levels in 18bp in the spleen, including nonresident macrophages, lymphocytes, the circulation. Changes in Ifn-g levels might participate to or granulocytes, could also contribute to Il-18bp production. The Journal of Immunology 7 Downloaded from http://www.jimmunol.org/ by guest on October 4, 2021

FIGURE 4. Ifn-g blockade with a monoclonal anti–Ifn-g Ab suppresses the induction of Il-18bp expression. WT mice were treated with an isotype control (n = 5) or a monoclonal anti–Ifn-g (n = 5) Ab and then received a single PBS injection (baseline) or a single CpG injection. All mice were euthanized at 24 h. Spleen (left panels), liver (middle panels), and lung (right panels) mRNA levels of CIIta (A), Cxcl9 (B), Il18bp (C), and pt-Il18bp (D) were normalized to the housekeeping gene Rpl32. Il-18bp plasma levels were determined by ELISA (E). Each dot represents one mouse. Horizontal and vertical bars represent mean 6 SD. Statistical analysis was performed using a two-way ANOVA with Holm–Sidak correction for multiple comparisons tests. 8 RADIOSENSITIVE AND RADIORESISTANT CELLS PRODUCE IL-18BP

Table I. Primers used for quantitative RT-PCR

GenBank Accession Sequence Forward Sequence Reverse Product Length Gene Number Primer Primer (pb) L32 NM_172,086.2 59-CACCAGTCAGACCGATATGTGAAAA-39 59-TGTTGTCAATGCCTCTGGGTTT-39 64 Il18 NM_008360.2 59-CAGGCCTGACATCTTCTG-39 59-CTGACATGGCAGCCATT-39 104 Il18bp NM_010531.2 59-ACATCTGCACCTCAGACAACT-39 59-TGGGAGGTGCTCAATGAAGGAACCA-39 234 pt-Il18bp GeneID: 16068 59-CCAAGAAGCTATTCGGGGCT-39 59-TCACCCCAAGGTCAGGTGTA-39 126 Ifng NM_008337.4 59-CTTCTTCAGCAACAGCAAGGCG-39 59-ATGCTTGGCGCTGGACCTGTG-39 89 Cxcl9 NM_008599.4 59-CGAGGCACGATCCACTACAA-39 59-AGGCAGGTTTGATCTCCGTT-39 113 CIIta NM_001243761.2 59-CAGCACTCAGAAGCACGGG-39 59-ATCCATGGTGGCACACAGACT-39 62

As HLH are highly heterogenous, it is likely that the respective with the predominant role of cells of hematopoietic origin as a role/importance of different cytokines varies depending on the cellular source of Il-18bp in the spleen, human neutrophils con- subtype of HLH studied. Indeed, a recent study demonstrates that stitutively produce IL-18BP, the levels of which are not further most manifestations of a murine model of fHLH could appear in an enhanced by various stimulations in vitro (47). Therefore, our Ifn-g–independent manner (42). By opposition, IFN-g has been results led us to conclude that the different Il-18bp–producing shown to play a central role in human MAS as well as in mouse cells are subject to different types of regulation. Further experi- models of MAS (20–22, 26, 29, 43). In the CpG-induced MAS ments on the characterization of Il-18bp–producing cells are model, it was previously shown that the different MAS manifes- needed to better elucidate the specific mechanisms involved in the Downloaded from tations were predominantly dependent on Ifn-g with only minor regulation of constitutive and inducible Il-18bp production in contributions of TNF-a and type I IFNs when Il-10 signaling is particular cell types. fully competent (20, 21, 26, 29). In addition, we recently dem- In conclusion, our results indicate that several cell types con- onstrated a protective role of Il-18bp in CpG-induced MAS by tribute to the production of Il-18bp in a model of MAS and that Il- using Il-18bp2/2 mice. The use of Abs blocking either Il-18 or 18bp behaves as an Ifn-g signature gene during CpG-induced Ifn-g markedly attenuated the severity of MAS manifestations in MAS. Taken together, our findings support the critical role of Il- http://www.jimmunol.org/ Il-18bp–deficient mice (22). Of note, free Il-18 levels were ele- 18bp to avoid excessive systemic effects of IL-18. vated in the circulation of these mice as well as in patients with MAS in comparison with healthy controls and patients with fHLH Acknowledgments (18), suggesting that the IL-18/IFN-g axis contributes to the We thank Sabina Troccaz (University of Geneva, School of Medicine, pathogenesis of MAS. Geneva, Switzerland) for help with the in vivo experiments. We are also IL-18 is a major inducer of IFN-g, and IFN-g, in turn, has been grateful to Praxedis Martin-Rodriguez for helpful scientific advice. described as an inducer of IL-18BP production. This stimulatory effect of IFN-g has been described in vitro in various non- by guest on October 4, 2021 hematopoietic cell types, such as epithelial cells, keratinocytes, Disclosures and fibroblasts, as well as in PBMCs (7, 23–25, 27). In addition, The authors have no financial conflicts of interest. circulating levels of IL-18BP correlated with the expression of other IFN-g–responsive genes in patients with sJIA and MAS (18, References 43). However, the role of Ifn-g in the regulation of Il-18bp pro- 1. Nakamura, K., H. Okamura, M. Wada, K. Nagata, and T. Tamura. 1989. duction had not been examined in vivo. Using a neutralizing anti– Endotoxin-induced serum factor that stimulates gamma interferon production. Ifn-g Ab, we observed that Il-18bp production in response to a Infect. Immun. 57: 590–595. 2. Boraschi, D., and C. A. Dinarello. 2006. IL-18 in autoimmunity: review. Eur. single CpG stimulation is dependent on Ifn-g signaling. Il18bp Cytokine Netw. 17: 224–252. therefore represents a bona fide Ifn-g molecular signature gene. 3. Gracie, J. A., S. E. Robertson, and I. B. McInnes. 2003. Interleukin-18. J. These findings also indicate that Il-18bp acts in a negative feed- Leukoc. Biol. 73: 213–224. 4. Novick, D., S. Kim, G. Kaplanski, and C. A. Dinarello. 2013. Interleukin-18, back loop to dampen excessive Il-18 activity during MAS. more than a Th1 cytokine. Semin. Immunol. 25: 439–448. Several studies demonstrate that IFN-g stimulation increases 5. Kaplanski, G. 2018. Interleukin-18: biological properties and role in disease pathogenesis. Immunol. Rev. 281: 138–153. Il18bp expression at the mRNA level and de novo IL-18BP pro- 6. Palomo, J., D. Dietrich, P. Martin, G. Palmer, and C. Gabay. 2015. The inter- tein synthesis in various human cell lines such as HaCat, DLD-1, leukin (IL)-1 cytokine family--Balance between agonists and antagonists in in- THP-1 and PBMC (23, 27, 44–46). Also, two well-designed flammatory diseases. Cytokine 76: 25–37. 7. Nold-Petry, C. A., T. Lehrnbecher, A. Jarisch, D. Schwabe, J. M. Pfeilschifter, studies showed the crucial importance of a proximal gamma H. Muhl, and M. F. Nold. 2010. Failure of interferon gamma to induce the anti- IFN activation site element within the human IL18BP promoter inflammatory interleukin 18 binding protein in familial hemophagocytosis. PLoS for its upregulation by IFN-g (45, 46). These authors also dem- One 5: e8663. 8. Dinarello, C. A., D. Novick, S. Kim, and G. Kaplanski. 2013. Interleukin-18 and onstrated different epigenetic regulation in epithelial and mono- IL-18 binding protein. Front. Immunol. 4: 289. cytic cells at this proximal gamma IFN activation site. To our 9. Novick, D., S. H. Kim, G. Fantuzzi, L. L. Reznikov, C. A. Dinarello, and M. Rubinstein. 1999. Interleukin-18 binding protein: a novel modulator of the knowledge, few studies examined specifically the regulation of Il- Th1 cytokine response. Immunity 10: 127–136. 18bp in mouse cell lines or mice. In this study, we demonstrated 10. Dinarello, C. A. 2001. Novel targets for interleukin 18 binding protein. Ann. that the effect of Ifn-g blockade was particularly efficient on the Rheum. Dis. 60(Suppl. 3): iii18–iii24. 11. Mazodier, K., V. Marin, D. Novick, C. Farnarier, S. Robitail, N. Schleinitz, CpG-induced production of Il-18bp in the liver and lung. In ad- V. Veit, P. Paul, M. Rubinstein, C. A. Dinarello, et al. 2005. Severe imbalance of dition, we showed that Il-18 is transcriptionally regulated by Ifn-g IL-18/IL-18BP in patients with secondary hemophagocytic syndrome. Blood in these two organs. Conversely, in the spleen, Il-18bp was neither 106: 3483–3489. 12. Girard, C., J. Rech, M. Brown, D. Allali, P. Roux-Lombard, F. Spertini, induced by CpG nor altered by the administration of the anti–Ifn-g E. J. Schiffrin, G. Schett, B. Manger, S. Bas, et al. 2016. Elevated serum levels of Ab. These findings are consistent with the fact that different cel- free interleukin-18 in adult-onset Still’s disease. Rheumatology (Oxford) 55: 2237–2247. lular sources are involved in Il-18bp production in these different 13. Zhang, K., J. Biroschak, D. N. Glass, S. D. Thompson, T. Finkel, M. H. Passo, organs, some of which appear to be responsive to Ifn-g. In line B. A. Binstadt, A. Filipovich, and A. A. Grom. 2008. Macrophage activation The Journal of Immunology 9

syndrome in patients with systemic juvenile idiopathic arthritis is associated with 30. Ghayur, T., S. Banerjee, M. Hugunin, D. Butler, L. Herzog, A. Carter, L. Quintal, MUNC13-4 polymorphisms. Arthritis Rheum. 58: 2892–2896. L. Sekut, R. Talanian, M. Paskind, et al. 1997. Caspase-1 processes IFN-gamma- 14. Vastert, S. J., R. van Wijk, L. E. D’Urbano, K. M. de Vooght, W. de Jager, inducing factor and regulates LPS-induced IFN-gamma production. Nature 386: A. Ravelli, S. Magni-Manzoni, A. Insalaco, E. Cortis, W. W. van Solinge, et al. 619–623. 2010. Mutations in the perforin gene can be linked to macrophage activation 31. Gu, Y., K. Kuida, H. Tsutsui, G. Ku, K. Hsiao, M. A. Fleming, N. Hayashi, syndrome in patients with systemic onset juvenile idiopathic arthritis. Rheu- K. Higashino, H. Okamura, K. Nakanishi, et al. 1997. Activation of interferon- matology (Oxford) 49: 441–449. gamma inducing factor mediated by interleukin-1beta converting enzyme. Sci- 15. Zhang, M., C. Bracaglia, G. Prencipe, C. J. Bemrich-Stolz, T. Beukelman, ence 275: 206–209. R. A. Dimmitt, W. W. Chatham, K. Zhang, H. Li, M. R. Walter, et al. 2016. A 32. Akita, K., T. Ohtsuki, Y. Nukada, T. Tanimoto, M. Namba, T. Okura, heterozygous RAB27A mutation associated with delayed cytolytic granule po- R. Takakura-Yamamoto, K. Torigoe, Y. Gu, M. S. Su, et al. 1997. Involvement of larization and hemophagocytic lymphohistiocytosis. J. Immunol. 196: 2492– caspase-1 and caspase-3 in the production and processing of mature human 2503. interleukin 18 in monocytic THP.1 cells. J. Biol. Chem. 272: 26595–26603. 16. Schulert, G. S., M. Zhang, A. Husami, N. Fall, H. Brunner, K. Zhang, 33. Fantuzzi, G., and C. A. Dinarello. 1999. Interleukin-18 and interleukin-1 beta: R. Q. Cron, and A. A. Grom. 2018. Brief report: novel UNC13D intronic variant two cytokine substrates for ICE (caspase-1). J. Clin. Immunol. 19: 1–11. disrupting an NF-kB enhancer in a patient with recurrent macrophage activation 34. Omoto, Y., K. Tokime, K. Yamanaka, K. Habe, T. Morioka, I. Kurokawa, syndrome and systemic juvenile idiopathic arthritis. Arthritis Rheumatol. 70: H. Tsutsui, K. Yamanishi, K. Nakanishi, and H. Mizutani. 2006. Human mast 963–970. cell chymase cleaves pro-IL-18 and generates a novel and biologically active IL- 17. Ramos-Casals, M., P. Brito-Zero´n, A. Lo´pez-Guillermo, M. A. Khamashta, and 18 fragment. J. Immunol. 177: 8315–8319. X. Bosch. 2014. Adult haemophagocytic syndrome. Lancet 383: 1503–1516. 35. Banerjee, S., and J. S. Bond. 2008. Prointerleukin-18 is activated by meprin beta 18. Weiss, E. S., C. Girard-Guyonvarc’h, D. Holzinger, A. A. de Jesus, Z. Tariq, in vitro and in vivo in intestinal inﬂammation. J. Biol. Chem. 283: 31371–31377. J. Picarsic, E. J. Schiffrin, D. Foell, A. A. Grom, S. Ammann, et al. 2018. 36. Omoto, Y., K. Yamanaka, K. Tokime, S. Kitano, M. Kakeda, T. Akeda, Interleukin-18 diagnostically distinguishes and pathogenically promotes human I. Kurokawa, E. C. Gabazza, H. Tsutsui, N. Katayama, et al. 2010. Granzyme B and murine macrophage activation syndrome. Blood 131: 1442–1455. is a novel interleukin-18 converting enzyme. J. Dermatol. Sci. 59: 129–135. 19. Kiltz, U., D. Kiefer, J. Braun, E. J. Schiffrin, C. Girard-Guyonvarc’h, and 37. Duran-Struuck, R., and R. C. Dysko. 2009. Principles of bone marrow trans- C. Gabay. 2020. Prolonged treatment with Tadekinig alfa in adult-onset Still’s plantation (BMT): providing optimal veterinary and husbandry care to irradiated disease. Ann. Rheum. Dis. 79: e10. mice in BMT studies. J. Am. Assoc. Lab. Anim. Sci. 48: 11–22. 20. Weaver, L. K., N. Chu, and E. M. Behrens. 2019. Brief report: interferon- 38. Klein, I., J. C. Cornejo, N. K. Polakos, B. John, S. A. Wuensch, D. J. Topham, Downloaded from g-mediated immunopathology potentiated by toll-like receptor 9 activation in a R. H. Pierce, and I. N. Crispe. 2007. Kupffer cell heterogeneity: functional murine model of macrophage activation syndrome. Arthritis Rheumatol. 71: properties of bone marrow derived and sessile hepatic macrophages. Blood 110: 161–168. 4077–4085. 21. Behrens, E. M., S. W. Canna, K. Slade, S. Rao, P. A. Kreiger, M. Paessler, 39. Beattie, L., A. Sawtell, J. Mann, T. C. M. Frame, B. Teal, F. de Labastida Rivera, T. Kambayashi, and G. A. Koretzky. 2011. Repeated TLR9 stimulation results in N. Brown, K. Walwyn-Brown, J. W. J. Moore, S. MacDonald, et al. 2016. Bone macrophage activation syndrome-like disease in mice. J. Clin. Invest. 121: 2264– marrow-derived and resident liver macrophages display unique transcriptomic 2277. signatures but similar biological functions. J. Hepatol. 65: 758–768.

22. Girard-Guyonvarc’h, C., J. Palomo, P. Martin, E. Rodriguez, S. Troccaz, 40. Soysa, R., S. Lampert, S. Yuen, A. N. Douglass, W. Li, K. Pfeffer, and http://www.jimmunol.org/ G. Palmer, and C. Gabay. 2018. Unopposed IL-18 signaling leads to severe I. N. Crispe. 2019. Fetal origin confers radioresistance on liver macrophages via TLR9-induced macrophage activation syndrome in mice. Blood 131: 1430– p21cip1/WAF1. J. Hepatol. 71: 553–562. 1441. 41. Schyns, J., Q. Bai, C. Ruscitti, C. Radermecker, S. De Schepper, S. Chakarov, 23. Paulukat, J., M. Bosmann, M. Nold, S. Garkisch, H. Ka¨mpfer, S. Frank, F. Farnir, D. Pirottin, F. Ginhoux, G. Boeckxstaens, et al. 2019. Non-classical J. Raedle, S. Zeuzem, J. Pfeilschifter, and H. Mu¨hl. 2001. Expression and release tissue monocytes and two functionally distinct populations of interstitial mac- of IL-18 binding protein in response to IFN-gamma. J. Immunol. 167: 7038– rophages populate the mouse lung. Nat. Commun. 10: 3964. 7043. 42. Burn, T. N., L. Weaver, J. E. Rood, N. Chu, A. Bodansky, P. A. Kreiger, and 24. Corbaz, A., T. ten Hove, S. Herren, P. Graber, B. Schwartsburd, I. Belzer, E. M. Behrens. 2020. Genetic deﬁciency of Interferon-g reveals interferon- J. Harrison, T. Plitz, M. H. Kosco-Vilbois, S. H. Kim, et al. 2002. IL-18-binding g-independent manifestations of murine hemophagocytic lymphohistiocytosis. protein expression by endothelial cells and macrophages is up-regulated during Arthritis Rheumatol. 72: 335–347. active Crohn’s disease. J. Immunol. 168: 3608–3616. 43. Put, K., A. Avau, E. Brisse, T. Mitera, S. Put, P. Proost, B. Bader-Meunier,

25. Wittmann, M., R. Doble, M. Bachmann, J. Pfeilschifter, T. Werfel, and H. Mu¨hl. R. Westhovens, B. J. Van den Eynde, C. Orabona, et al. 2015. Cytokines in by guest on October 4, 2021 2012. IL-27 Regulates IL-18 binding protein in skin resident cells. PLoS One 7: systemic juvenile idiopathic arthritis and haemophagocytic lymphohistiocytosis: e38751. tipping the balance between interleukin-18 and interferon-g. Rheumatology 26. Canna, S. W., J. Wrobel, N. Chu, P. A. Kreiger, M. Paessler, and E. M. Behrens. (Oxford) 54: 1507–1517. 2013. Interferon-g mediates anemia but is dispensable for fulminant toll-like 44. Mo¨ller, B., J. Paulukat, M. Nold, M. Behrens, N. Kukoc-Zivojnov, receptor 9-induced macrophage activation syndrome and hemophagocytosis in J. P. Kaltwasser, J. Pfeilschifter, and H. Mu¨hl. 2003. Interferon-gamma induces mice. Arthritis Rheum. 65: 1764–1775. expression of interleukin-18 binding protein in fibroblast-like synoviocytes. 27. Mu¨hl, H., H. Ka¨mpfer, M. Bosmann, S. Frank, H. Radeke, and J. Pfeilschifter. Rheumatology (Oxford) 42: 442–445. 2000. Interferon-gamma mediates gene expression of IL-18 binding protein in 45. Bachmann, M., J. Paulukat, J. Pfeilschifter, and H. Mu¨hl. 2009. Molecular nonleukocytic cells. Biochem. Biophys. Res. Commun. 267: 960–963. mechanisms of IL-18BP regulation in DLD-1 cells: pivotal direct action of the 28. Harms, R. Z., A. J. Creer, K. M. Lorenzo-Arteaga, K. R. Ostlund, and STAT1/GAS axis on the promoter level. J. Cell. Mol. Med. 13(8B): 1987–1994. N. E. Sarvetnick. 2017. Interleukin (IL)-18 binding protein deficiency disrupts 46. Bachmann, M., J. Pfeilschifter, and H. Mu¨hl. 2018. Epigenetic regulation by natural killer cell maturation and diminishes circulating IL-18. Front. Immunol. CpG methylation splits strong from retarded IFNg-induced IL-18BP in epithelial 8: 1020. versus monocytic cells. Biochim. Biophys. Acta. Gene Regul. Mech. 1861: 191– 29. Heikenwalder, M., M. Polymenidou, T. Junt, C. Sigurdson, H. Wagner, S. Akira, 199. R. Zinkernagel, and A. Aguzzi. 2004. Lymphoid follicle destruction and im- 47. Fortin, C. F., T. Ear, and P. P. McDonald. 2009. Autocrine role of endogenous munosuppression after repeated CpG oligodeoxynucleotide administration. Nat. interleukin-18 on inflammatory cytokine generation by human neutrophils. Med. 10: 187–192. FASEB J. 23: 194–203. Figure S1

Figure S1. Gating strategy for flow cytometry analysis. Debris (SSC-A vs FSC-A), doublets (SSC-H vs SSC-W, then FSC- H vs FSC-W) and dead cells (Propidium Iodide (PI)+ cells) were excluded. For the study, only CD45+ cells (both CD45.1+ and CD45.2+ pooled using the “Make Or Gate” boolean tool) have been analyzed. T cells were defined as CD3+ cells. CD4+ and CD8+ cells were selected among T cells. B cells were defined as B220+ cells. cDC cells were selected as CD3- B220- CD11c+ cells. Then, remaining CD11b+ cells were analyzed. Among them, NK cells correspond to NK1.1+ cells and neutrophils to Ly6C+ Ly6G+cells. The remaining “undefined CD11b+ cells” likely correspond to monocytes, macrophages, and granulocytes. Figure S2

Figure S2. The radiosensitive compartment was efficiently reconstituted. Schematic representation of the CpG-induced MAS protocol used following bone marrow (BM) transfer. Four groups of mice were created following BM transfer, namely WT recipient mice transplanted with WT BM (WT→WT), WT recipient mice transplanted with Il-18bp-/- BM (KO→WT), Il-18bp-/- recipient mice transplanted with WT BM (WT→KO), and Il-18bp-/- recipient mice transplanted with Il-18bp-/- BM (KO→KO). Eight weeks after the BM transfer, the four groups of chimeric mice were injected with CpG on days 0, 2 and 4, and euthanized on day 7 (A). The reconstitution efficiency for each group, with the exception of the KO→KO group, was determined on day 7 in CpG-treated mice based on allelic differences between donor and recipient mice (CD45.1+ or CD45.2+). The proportion of CD45.1+ or CD45.2+ donor cells among all CD45+ spleen cells are represented (CD45.2+ WT → CD45.1+ WT: n=10 / CD45.1+ WT → CD45.2+ WT: n=5 / CD45.2+ KO → CD45.1+ WT: n=10 / CD45.1+ WT → CD45.2+ KO: n=5) (B). By using the same approach, but regardless of mouse groups, the efficiency of reconstitution is shown for the different immune cell types (n=44) (C). Each dot represents one mouse. Horizontal and vertical bars represent mean ± SD. Statistical analysis was performed using a Kruskal-Wallis test with Dunn’s post-hoc test. The P-value corresponds to the comparison of the different cell subsets to CD4+ and CD8+ T cells. Figure S3

Figure S3. Variations of spleen CD45+ subpopulations in response to CpG stimulation. BM transplanted mice received CpG injections on days 0, 2 and 4 and were euthanized on day 7. At sacrifice, the effects of CpG injections on CD45+ subpopulations in the spleen were studied by flow cytometry. Each dot represents one mouse. Horizontal and vertical bars represent mean ± SD: WT→WT (n=15), KO→WT (n=10), WT→KO (n=10) and KO→KO (n=9). Statistical analysis was performed using a Kruskal-Wallis test with Dunn’s post-hoc test.