Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2018 Mar 29;131(13):1442-1455.
doi: 10.1182/blood-2017-12-820852. Epub 2018 Jan 11.

Interleukin-18 Diagnostically Distinguishes and Pathogenically Promotes Human and Murine Macrophage Activation Syndrome

Affiliations
Free PMC article

Interleukin-18 Diagnostically Distinguishes and Pathogenically Promotes Human and Murine Macrophage Activation Syndrome

Eric S Weiss et al. Blood. .
Free PMC article

Abstract

Hemophagocytic lymphohistiocytosis (HLH) and macrophage activation syndrome (MAS) are life-threatening hyperferritinemic systemic inflammatory disorders. Although profound cytotoxic impairment causes familial HLH (fHLH), the mechanisms driving non-fHLH and MAS are largely unknown. MAS occurs in patients with suspected rheumatic disease, but the mechanistic basis for its distinction is unclear. Recently, a syndrome of recurrent MAS with infantile enterocolitis caused by NLRC4 inflammasome hyperactivity highlighted the potential importance of interleukin-18 (IL-18). We tested this association in hyperferritinemic and autoinflammatory patients and found a dramatic correlation of MAS risk with chronic (sometimes lifelong) elevation of mature IL-18, particularly with IL-18 unbound by IL-18 binding protein, or free IL-18. In a mouse engineered to carry a disease-causing germ line NLRC4T337S mutation, we observed inflammasome-dependent, chronic IL-18 elevation. Surprisingly, this NLRC4T337S-induced systemic IL-18 elevation derived entirely from intestinal epithelia. NLRC4T337S intestines were histologically normal but showed increased epithelial turnover and upregulation of interferon-γ-induced genes. Assessing cellular and tissue expression, classical inflammasome components such as Il1b, Nlrp3, and Mefv predominated in neutrophils, whereas Nlrc4 and Il18 were distinctly epithelial. Demonstrating the importance of free IL-18, Il18 transgenic mice exhibited free IL-18 elevation and more severe experimental MAS. NLRC4T337S mice, whose free IL-18 levels were normal, did not. Thus, we describe a unique connection between MAS risk and chronic IL-18, identify epithelial inflammasome hyperactivity as a potential source, and demonstrate the pathogenicity of free IL-18. These data suggest an IL-18-driven pathway, complementary to the cytotoxic impairment of fHLH, with potential as a distinguishing biomarker and therapeutic target in MAS.

Conflict of interest statement

Conflict-of-interest disclosure: AB2Bio, Ltd., holds exclusive rights to the human free IL-18 assay. E.J.S. is an employee of AB2Bio, Ltd. S.W.C. and C.G. are consultants for AB2Bio, Ltd. The remaining authors declare no competing financial interests.

Figures

None
Figure 1.
Figure 1.
IL-18 distinguishes MAS from fHLH during active hyperferritinemia. Serum samples from a cohort of healthy controls (HCs) and patients with active sJIA, MAS, IA-HLH, and fHLH were probed for (A) total IL-18, (B) IL-18 binding protein, (C) free IL-18, and (D) CXCL9. See Data supplement A for clinical details. (E) A receiver operating characteristic (ROC) for total IL-18 distinguishing MAS from fHLH was calculated. (F) Total IL-18/CXCL9 ratios were computed and (G) a ROC for this ratio distinguishing MAS from fHLH was calculated. ★Optimal cutoff of 2.3. All concentrations are in picogram/milliliter. *P < .05, **P < .01, ***P < .001, ****P < .0001 by Kruskal-Wallis test with Dunn multiple comparisons posttest, except for ROC analysis. All differences meeting corrected threshold of P < .05 between individual groups are shown except comparisons made to HC for total IL-18 and CXCL9. The open circle in MAS represents systemic lupus erythematosus–associated MAS; all others are sJIA-MAS. Open circles in fHLH represent plasma samples. AUC, area under curve.
Figure 2.
Figure 2.
IL-18 is uniquely elevated in autoinflammatory patients referred for MAS. (A) IL-18, (B) IL-18BP, (C) CXCL9, and (D) free IL-18 were measured in serum samples from a broad referral cohort of patients with autoinflammatory disease. Acronyms, genetic associations, and statistical testing can be found in supplemental Table 1 and Data supplement B. Patients per group are indicated in parentheses, with multiple samples gathered longitudinally from some patients (eg, Figure 3; supplemental Figure 3). Gray bars, range of values measured in HC samples. CANDLE, chronic atypical neutrophilic dermatosis, lipodystrophy, elevated temperature; CNO, chronic nonbacterial osteomyelitis; DADA2, deficiency of ADA2; DIRA, deficiency of IL-1 receptor antagonist; FMF, familial Mediterranean fever; PAAND, pyrin-associated autoinflammation with neutrophilic dermatosis; SAVI, STING-associated vasculopathy of infancy; TRAPS, TNF receptor–associated periodic syndrome; undiff, undifferentiated; XIAP-def, XIAP-deficiency.
Figure 3.
Figure 3.
Different patterns of chronic IL-18 in MAS-prone patients. Time course of cytokine and disease activity measurements in a patient with (A) NLRC4-associated MAS, (B) idiopathic MAS, and (C) adult-onset Still disease presenting with MAS. Dashed lines crossing the y-axes indicate upper limits of normal . In panel A, influenza vaccination is indicated by asterisks along the x-axis. CRP in milligram/liter, ferritin in nanogram/milliliter, all others in picogram/milliliter.
Figure 4.
Figure 4.
Mice bearing Nlrc4T337Smutation have chronic, inflammasome-dependent IL-18 elevation. (A) Serum IL-18 from age- and sex-matched NLRC4T337S/T337S, NLRC4T337S/WT, and WT mice. (B) NLRC4T337S/WT mice or littermate controls were assessed for serum IL-18 upon weaning at 21 days of age. (C) Serum was obtained from NLRC4T337S/T337S and WT mice housed separately until 7 weeks of age, then cohoused for 4 weeks, and then cohoused and administered broad-spectrum antibiotics for an additional 4 weeks. (D) Serum IL-18 was assessed in mice of the indicated genotypes. Each graph is representative of at least 2 separate experiments. ***P < .001, ***P < .0001, 1-way analysis of variance (ANOVA) with Tukey posttest.
Figure 5.
Figure 5.
Expression of Il18 and Nlrc4 converge in barrier epithelial tissues. (A) RNA-seq expression values of Il1b and Il18 (left) and inflammasomopathy-associated genes (right) in murine immune cells derived from analysis of data from ImmGen project. (B) The gcrma microarray expression values of Il18 and Il1b in diverse murine tissues from BioGPS dataset GNF1M. Expression values of ex vivo LPS-stimulated macrophages are included as an internal high Il1b control. Microarray expression of (C) human immune cells or (D) tissues with consistent expression (Z > 5 in both duplicates) of IL1B or IL18. Data were derived from BioGPS dataset “Barcode on Normal Tissues.” An Nlrc4 probeset was not present in multiple large murine tissue datasets, precluding a comparison of inflammasomes in human tissues. All graphs depict mean and standard error of the mean of all available replicates. Data for panels A-B were obtained from http://www.immgen.org and BioProject PRJNA2B1360. Data for panels B-D were obtained via http://www.biogps.org. DC, dendritic cell; GI, gastrointestinal; NKT, NK T cell; resp, respiratory; Treg, T regulatory cell.
Figure 6.
Figure 6.
IL-18 overproduction in NLRC4T337S/T337Smice derives from IECs and induces proliferation and major histocompatibility complex II upregulation. (A) Serum IL-18 concentration in bone marrow chimeras as depicted. Values at 0 time point represent serum concentration in donor mice. *P < .05 for all comparisons of top lines to bottom lines; repeated measures ANOVA with Tukey test for multiple comparisons. Representative of 2 experiments with at least 4 mice per group. (B) Serum IL-18 in Il18flox/flox mice of the indicated genotypes. (C) Mice were injected with EdU and histologically evaluated 4 hours later for the proportion of EdU positive (green) cells among all villous cells (4′,6-diamidino-2-phenylindole). Original magnification ×200. Analysis by Student t test of pooled data from 3 separate experiments. Each point represents a distinct organ. (D) Genes most differentially upregulated (red) or downregulated (blue) in NLRC4T337S/T337S duodenal epithelium by RNA-seq; see also Data supplements F and G.
Figure 7.
Figure 7.
Free IL-18 determines susceptibility to more severe TLR9-induced MAS. Mice of the indicated genotypes were repeatedly injected with phosphate-buffered saline or the TLR9-stimulus CpG and were assessed for (A) serum total and free IL-18; (B) relative spleen size, aspartate aminotransferase, and platelet count; and (C) serum IL-10, CXCL9, and IFN-γ. *Adjusted P < .01, **adjusted P < .001, 2-way ANOVA with Tukey posttest. Significance is only shown for comparisons where adjusted P < .05. Results are representative of at least 2 independent experiments.

Comment in

  • Fire behind the fury: IL-18 and MAS.
    McClain KL, Allen CE. McClain KL, et al. Blood. 2018 Mar 29;131(13):1393-1394. doi: 10.1182/blood-2018-02-828186. Blood. 2018. PMID: 29599143 No abstract available.

Similar articles

See all similar articles

Cited by 30 articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback