Excessive IL-10 and IL-18 trigger hemophagocytic lymphohistiocytosis-like hyperinflammation and enhanced myelopoiesis

Abstract

Background: Hyperinflammation is a life-threatening condition associated with various clinical disorders characterized by excessive immune activation and tissue damage. Multiple cytokines promote the development of hyperinflammation; however, the contribution of IL-10 remains unclear despite emerging speculations for a pathological role. Clinical observations from hemophagocytic lymphohistiocytosis (HLH), a prototypical hyperinflammatory disease, suggest that IL-18 and IL-10 may collectively promote the onset of a hyperinflammatory state.

Objective: We aimed to investigate the collaborative roles of IL-10 and IL-18 in hyperinflammation.

Methods: A comprehensive plasma cytokine profile for 87 secondary HLH patients was first depicted and analyzed. We then investigated the systemic and cellular effects of coelevated IL-10 and IL-18 in a transgenic mouse model and cultured macrophages. Single-cell RNA sequencing was performed on the monocytes/macrophages isolated from secondary HLH patients to explore the clinical relevance of IL-10/IL-18-mediated cellular signatures. The therapeutic efficacy of IL-10 blockade was tested in HLH mouse models.

Results: Excessive circulating IL-10 and IL-18 triggered a lethal hyperinflammatory disease recapitulating HLH-like phenotypes in mice, driving peripheral lymphopenia and a striking shift toward enhanced myelopoiesis in the bone marrow. IL-10 and IL-18 polarized cultured macrophages to a distinct proinflammatory state with pronounced expression of myeloid cell-recruiting chemokines. Transcriptional characterization suggested the IL-10/IL-18-mediated cellular features were clinically relevant with HLH, showing enhanced granzyme expression and proteasome activation in macrophages. IL-10 blockade protected against the lethal disease in HLH mouse models.

Conclusion: Coelevated IL-10 and IL-18 are sufficient to drive HLH-like hyperinflammatory syndrome, and blocking IL-10 is protective in HLH models.

Keywords: HLH; IL-10; IL-18; cytokine storm; hyperinflammation; macrophage polarization; myelopoiesis.

Figure 1.. Pronounced co-elevation of IL-10 and IL-18 in sHLH patients.

(A-B) Relative levels of the indicated cytokines/chemokines in the plasma samples of MAS (A) and EBV-LPD-HLH (B). (C) Relative levels of the indicated cytokines/chemokines in the plasma samples of EBV-LPD-HLH (left) and EBV-LPD-non-HLH (right) patients. (D) Summary of the ratio of median concentrations for the indicated cytokine/chemokine factor detected in EBV-LPD-HLH samples to those in EBV-LPD-non-HLH samples. (E) ROC curve evaluation of the performance of IL-18, IL-2Ra, IL-10, IFN-γ, IP-10, and GM-CSF in distinguishing EBV-LPD-HLH from EBV-LPD-non-HLH. (F) Relative levels of IL-10, IL-18, and IL-2Ra in the plasma samples of EBV-LPD patients in HLH status or HLH remission (N = 13). Results were calculated as fold changes relative to the levels detected in healthy donors. (G) Relative plasma levels of IL-10 and IL-18 in EBV-LPD-HLH (green) and EBV-LPD-non-HLH patients (blue) were plotted. The dotted lines represent cut-off values of IL-18 and IL-10 calculated according to Youden’s index, which maximized the sensitivity and specificity to differentiate EBV-LPD-HLH samples from EBV-LPD-non-HLH samples. For panels A and B, results were shown as mean ± s.d. For panels C and F, Mann-Whitney tests were performed. **, P < 0.01; ***, P < 0.001.

Figure 2.. Co-elevated IL-10 and IL-18 drive HLH-like syndrome in a pathogen- and mutation-free mouse model.

(A) Scheme for the inducible Il10- and Il18- transgene in the Dual-tg model. (B) Scheme for transgene induction and experimental design. (C) Serum IL-10 and IL-18 concentrations determined by ELISA assays (n > 10 per group). (D) AST and ALT activities in serum samples (n = 6–7 per group). (E) Survival curves and the result of log-rank test (n > 6 per group). (F) CBC results of the indicated mouse groups (n = 7 per group). (G) Representative pictures and weights of spleens (n = 6 per group). (H) Serum ferritin and triglyceride concentrations measured by ELISA assays (n = 4–5 per group). (I) Representative flow plot and summary of CD107a degranulation by splenic NK cells (n = 3 per group). (J) Representative pictures of hemophagocytes in cytospin preparations of spleen or BM cells. Scale bars, 20 μm. (K) The frequencies of hemophagocytes observed in cytospin preparations of spleen or BM cells (n = 3 per group). (L) Representative HE-staining pictures indicating hemophagocytes (black arrows) found in the liver and spleen. Scale bars, 50 μm. For panels C, D, F-I, K, Student’s t-tests or Welch’s t-tests were performed. Data are shown as mean ± s.d. unless otherwise noted. Data represent at least two independent experiments. P < 0.05 is considered as statistically significant. ns, not statistically significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 3.. Dual-tg mice present lymphopenia and enhanced myelopoiesis.

(A) Flow cytometric analysis of B cell, NK cell, and T cell fractions in spleens (n = 5 per group). (B) Flow cytometric analysis of CD4⁺ and CD8⁺ T cell fractions in the splenic T cells (n = 5 per group). (C-D) Flow cytometric analysis of neutrophil (C) and monocyte (D) fractions in the PB (n = 5 per group). (E) Flow cytometric analysis of monocyte-derived dendritic cell (moDC) fractions in the spleens (n = 3 per group). (F) Representative pictures of immunohistochemical staining of CD68 in liver, spleen, and BM (left). Semi-quantitative immunoreactivity histological scores (H scores) of CD68 (right). Scale bars, 50 μm. (G) Gating strategy for flow cytometric analysis of HSPC compartment. Lineage markers include CD3, CD4, CD8, B220, CD11b, Gr1, and Ter119. (H) BM cellularity (femurs) of the indicated mouse groups (n = 5 per group). (I-K) Percentages of LT-HSC, ST-HSC (I), MMP2, MMP3, MMP4 (J), CLP, CMP, MEP, and GMP (K) populations in the BM nucleated cells (n = 5 per group). Student’s t-tests or Welch’s t-tests were performed. Data are shown as mean ± s.d. Data represent at least two independent experiments. P < 0.05 is considered as statistically significant. ns, not statistically significant, *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Figure 4.. IL-10 in combination with IL-18 polarizes macrophages to a distinct state.

(A) Representative flow plots of CD80 and CD206 expression on macrophages (Gr1⁻CD115⁻F4/80⁺SSC^Int) of spleen and BM. Data are representative of at least five independent experiments. (B) Experimental design for RNA-seq analysis. (C) Numbers of DEGs (FDR < 0.05, FC > 2) in pairwise comparisons of in vitro cultured mouse macrophages. (D) GSEA screening for KEGG pathway enrichment (FDR < 0.05) in the upregulated genes (FDR < 0.05, FC > 2) of M1018 compared to Naïve macrophages. Red arrows indicate highlighted pathways that were also enriched in M10 or M18 macrophages. Blue arrows indicate highlighted pathways that were exclusively enriched in M1018 macrophages. (E) Heatmap showing expression of genes encoding cytokines and chemokines. (F) MCP-1(CCL2), MCP-3(CCL7), MIP-1α(CCL3), MIP-1β (CCL4), IL-1β, and IFN-γ in the supernatants of in vitro polarized mouse macrophages (Figure 4B) determined by Luminex assay (mean ± s.d., triplicates). (G) Venn diagram plotted with the differentially upregulated genes (FDR < 0.05, FC > 1.5) in M1018, M10, and M18 compared to Naïve macrophages. The top 20 genes upregulated in the M1018 but not upregulated in M10 or M18 macrophages were listed. (H) Representative flow plot (top) and summary (bottom) showing granzyme B expression by the BM macrophages (Gr1⁻CD115⁻F4/80⁺SSC^Int) of Dual-tg and Ctrl mice (mean ± s.d., n = 4 per group). ***P<0.001 (Student’s t-test). For panels A, F, and H, data are representative of two independent experiments. (I) KEGG pathway enrichment analysis (FDR < 0.1) of the 631 genes shown in Figure 4G. (J) GSEA analysis of proteasome pathway. NES, normalized enrichment score; FDR, false-discovery rate.

Figure 5.. scRNA-seq analysis of the monocyte/macrophage populations isolated from patients with sHLH.

(A) Scheme of experimental design. (B) The t-distributed stochastic neighbor embedding (t-SNE) plot demonstrates the main cell types identified in the samples of sHLH patients and healthy donors collectively. (C) The summary shows the cell proportions of the disease group within the indicated monocyte/macrophage clusters. (D) KEGG pathway enrichment analysis of the upregulated genes in cluster 12 compared to other monocyte/macrophage clusters (excluding cluster 11) identified in the samples analyzed collectively. (E) Expression levels of GZMA and GZMB in different monocyte/macrophage clusters identified in the samples analyzed collectively. (F) Monocle prediction of the developmental trajectory of monocyte/macrophage clusters with pseudotime (up) and cluster (11 and 12) information (down) mapped on. (G) Heat map shows expression levels of marker genes of cMoP and activated macrophage ordered by pseudotime. cMoP, common monocyte progenitor; Mφ, macrophage.

Figure 6.. Blocking IL-10 is protective in HLH disease models.

(A) Scheme of neutralizing antibody treatment in Dual-tg/Xiap KO mice. (B) CBC results of Dual-tg/Xiap KO mice treated with the indicated neutralizing antibodies after Poly(I:C) induction (mean ± s.e.m., n > 3 per group). (C) Survival curves of Dual-tg/Xiap KO mice treated with the indicated neutralizing antibodies (n > 3 per group). *P < 0.05 (log-rank test). (D) Representative flow plots showing CD80 and CD206 expression in spleen and BM macrophages from Dual-tg/Xiap KO mice treated with the indicated neutralizing antibodies. Cells were gated as Gr1⁻CD115⁻F4/80⁺SSC^int cells. (E) Representative flow plots showing the HSPC fractions in Dual-tg/Xiap KO mice treated with the indicated neutralizing antibodies. Cells were gated as Lin⁻(CD3, CD4, CD8, B220, CD11b, Gr1, and Ter119) cells. (F) Scheme of neutralizing antibody treatment in LCMV-infected Prf1^−/− mice. (G) Survival curves of LCMV-infected Prf1^−/− mice with the indicated neutralizing antibody treatment (n = 3–6 per group). Data are representative of two independent experiments. **P<0.01 (log-rank test). Data represent at least two independent experiments.