Sterile liver injury induces a protective tissue-resident cDC1-ILC1 circuit through cDC1-intrinsic cGAS-STING-dependent IL-12 production

Abstract

Tissue-resident immune cells are critical to the initiation and potentiation of inflammation. However, the tissue-protective cellular communication networks initiated by resident immunity during sterile inflammation are not well understood. Using single-cell transcriptomic analysis, we show the liver-resident cell connectome and signalome during acute liver injury. These analyses identify Il12b as a central regulator of liver injury-associated changes in gene expression. Interleukin (IL)-12 produced by conventional type 1 dendritic cells (cDC1s) is required for protection during acute injury through activation of interferon (IFN)-γ production by liver-resident type 1 innate lymphoid cells (ILC1s). Using a targeted in vivo CRISPR-Cas9 screen of innate immune sensing pathways, we find that cDC1-intrinsic cGAS-STING signaling acts upstream of IL-12 production to initiate early protective immune responses. Our study identifies the core communication hubs initiated by tissue-resident innate immune cells during sterile inflammation in vivo and implicates cDC1-derived IL-12 as an important regulator of this process.

Keywords: CP: Immunology; IL-12; ILC1; cDC1; liver injury; single-cell RNA-seq; sterile inflammation.

Figure 1.. scRNA-seq analysis of APAP-induced liver injury identifies early activation states of liver-resident immune cells

(A) Uniform manifold approximation and projection (UMAP) plot of 2,558 mouse liver cells harvested from healthy or 24 h post-acetaminophen (APAP) treatment. Cells are colored by their annotations derived from cluster-specific analysis. (B) UMAP indicating cell source: WT (blue) or APAP treated (red). (C) The frequencies of APAP-enriched myeloid subsets from the total WT or APAP-treated cells (n = 2 mice per cohort). (D–F) The numbers and frequencies of (D and E) macrophage, monocyte, and (F) dendritic cell subsets from total liver CD45⁺ cells isolated from saline-treated control (WT) or APAP-treated mice 24 h after treatment (n = 7 mice per cohort). Data are representative of two independent experiments (D), and samples were compared using two-tailed Student’s t test with Welch’s correction, assuming unequal SD. Data are presented as individual points with the mean ± SEM (*p < 0.05, **p < 0.01). See also Figure S1 and Table S1.

Figure 2.. scRNA-seq receptor-ligand analysis identifies upstream regulators of liver injury-associated differentially expressed genes

(A–D) CellChat receptor-ligand analysis of the inferred intercellular communication networks from APAP-enriched cell states. Different cell types are represented by different colors within the circle plot. Arrows are proportional to the interaction strength between two cell types, while node size is relative to the number of cells within that population. (A) Weighted interactions stemming from activated cDC1s. (B) Weighted interactions from activated KCs. (C) Weighted interactions from LAMs. (D) Weighted interactions from monocytes. (E) NicheNet ligand activity prediction analysis. Results are shown for the top 20 ranked key regulator ligands and their target genes of interest. The top 25% of the interaction scores between each of these key regulators and their top 50 most strongly predicted gene targets were plotted within the regulatory potential matrix. Colored dots above the key regulators denote which cell types following APAP treatment express the ligand. (F) NicheNet circos plot visualization of the top active ligand-target links between Il12b and its top-predicted target genes. Width of the target block and degree of arrow transparency are proportional to the ligand-target regulatory potential value; larger/darker indicates a higher regulatory potential. Color of the target block denotes which signal is upstream of the target: IL-23 (green), IL-12 (blue), IL-23 and IL-12 (purple), or unknown (gray). Colored dots above the predicted targets denote which cell populations within the APAP-treated dataset express the gene. See also Figure S2 and Tables S2, S3, S4, S5, S6, and S7.

Figure 3.. IL-12 signaling in hematopoietic-derived cells protects against acute liver injury

(A) Plasma ALT concentrations from WT or IL12b^−/− mice before and 24, 48, and 72 h after APAP injection (n = 8 mice per group). (B) Plasma ALT concentrations from WT or IL12b^−/− mice before and 48 h after CCl₄ injection (n = 8 mice per group). (C) Histology of the liver left lateral lobes (hematoxylin and eosin staining) of WT or IL12b^−/− mice before and 48 h after CCl₄ injection (n = 6–8 mice per group). Scale bars represent 500 μm. (D) Quantification of centrilobular hepatocyte necrosis and damaged areas around the central veins in the livers of WT or IL12b^−/− mice before and 48 h after CCl₄ injection (n = 6–8 mice per group). (E) Plasma ALT concentrations of mice (that had been injected with a neutralizing anti-Il12p40 mAb or isotype Ig 12 h before and 12 h after CCl₄ injection) 48 h after CCl₄ injection (n = 8 mice per cohort). (F) Bone marrow chimeric mice were generated utilizing WT hosts reconstituted with either WT, Il12rb2^−/−, or Stat4^−/− bone marrow. Plasma ALT concentrations of WT:WT, WT:Il12rb2^−/−, and WT:Stat4^−/− mice before and 48 h after CCl₄ injection (n = 8 mice per group). Data are representative of two independent experiments (A–F), and samples were compared using two-tailed Student’s t test with Welch’s correction, assuming unequal SD. Data are presented as individual points with the mean ± SEM (*p < 0.05, **p < 0.01, ***p < 0.001).

Figure 4.. cDC1-derived IL-12 protects against acute liver injury

(A and B) Il12b^YFP reporter mice were either untreated or injected with CCl₄ and livers were harvested 12, 24, 36, and 48 h later to assess YFP⁺ cells by flow cytometry. (A) Representative flow plots of YFP⁺ percentage within liver CD45⁺Lin⁻CD88⁺ macrophage (left), CD45⁺Lin⁻CD88⁻MHCII⁺CD11c⁺CD11b⁺XCR1⁻ cDC2 (middle), or CD45⁺Lin⁻CD88⁻MHCII⁺CD11c⁺CD11b⁻XCR1⁺ cDC1 (right) populations 24 h after CCl₄ injection. (B) Kinetics of the percentages of YFP⁺ cells in macrophage, cDC2, and cDC1 populations in the liver after CCl₄ injection (n = 4 mice per group). (C) Schematic of adoptive transfer model using Xcr1^DTR mice. Briefly, Xcr1^DTR mice were either untreated or treated with PBS, DT, or DT followed by adoptive transfer of 1 × 10⁷ D9 bone marrow-derived WT cDCPs or IL12b^−/− cDCPs 1 day prior to CCl₄ treatment. Mice were bled and plasma ALT concentrations were analyzed 48 h after CCl₄ injection. (D) Representative flow plots showing liver cDC2 and cDC1 populations in Xcr1^DTR mice 24 h after PBS (left), DT (middle), or DT treatment plus adoptive transfer of bone marrow-derived cDCPs (right) (n = 8 mice per cohort). (E) The frequencies of cDC1s from total liver CD45⁺ cells of PBS-treated, DT-treated, or DT-treated plus adoptively transferred (+cDCP) Xcr1^DTR mice 24 h after treatment (n = 8 mice per cohort). (F) The numbers of liver cDC1s from PBS-treated, DT-treated, or DT-treated plus adoptively transferred (+cDCP) Xcr1^DTR mice 24 h after treatment (n = 8 mice per cohort). (G) Plasma ALT concentrations from naive, PBS-treated, DT-treated, DT-treated plus WT transferred cDCPs, or DT-treated plus IL12b^−/− cDCPs Xcr1^DTR mice 48 h after CCl₄ injection (n = 8–9 mice per group). Data are representative of three (D–F) and two independent experiments (A, B, and G), and samples were compared using two-tailed Student’s t test with Welch’s correction, assuming unequal SD. Data are presented as individual points with the mean ± SEM (NS, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). See also Figure S3.

Figure 5.. cDC1-derived IL-12 is required for group 1 ILC IFN-γ production following sterile liver injury

(A–D) Xcr1^DTR mice were untreated (naive) or CCl₄, CCl₄ + DT, CCl₄ + DT + WT transferred cDCP, or CCl₄ + DT + IL12b^−/− transferred cDCP treated before group 1 ILC IFN-γ was assessed 18 h later (n = 7–10 mice per group). (A) Representative flow plots of IFN-γ⁺ percentage within liver CD45⁺Lin⁻NK1.1⁺CD200r⁺CD49b⁻IL18r⁺ ILC1 from CCl₄- (top left), CCl₄ + DT- (top right), CCl₄ + DT + WT transferred cDCP- (bottom left), or CCl₄ + DT + IL12b^−/− transferred cDCP-treated (bottom right) cohorts. (B) The percentages of IFN-γ⁺ IL18r⁺ ILC1s, IL18r⁻ ILC1s, and CD45⁺Lin⁻NK1.1⁺CD200r⁻CD49b⁺ NK cells from the indicated cohorts. (C) The numbers of IFN-γ⁺ IL18r⁺ and IL18r⁻ ILC1s. (D) The numbers of IFN-γ⁺ NK cells. (E and F) Mixed bone marrow chimeric mice were generated using congenically distinct WT (CD45.1 × 2) hosts reconstituted with a 1:1 mixture of either WT (CD45.1):Il12rb2^−/− (CD45.2) or WT (CD45.1):Stat4^−/− (CD45.2) bone marrow. The percentages of IFN-γ⁺ (E) IL18r⁺ ILC1s and (F) NK cells in the liver of WT:Il12rb2^−/− and WT:Stat4^−/− mice were derived from either WT or knockout (KO) bone marrow (n = 4–5 mice per group). Data are representative of two independent experiments (A–F), and samples were compared using two-tailed Student’s t test with Welch’s correction, assuming unequal SD. Data are presented as individual points with the mean ± SEM (NS, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). See also Figure S4.

Figure 6.. cGAS-STING signaling is critical for cDC1-derived IL-12 production and protective function during sterile liver injury

(A) Schematic of adoptive transfer model system using CRISPR cRNP-edited cells. Briefly, 1 × 10⁷ Il12b^YFP bone marrow-derived cDCPs were cRNP edited against target genes or nontargeting control (NTC) before adoptive transfer into DT-treated Xcr1^DTR mice 1 day prior to CCl₄ treatment. Livers were then harvested, and the percentages of YFP⁺ cDC1 were analyzed 24 h after CCl₄ injection. (B and C) Unedited Il12b^YFP bone marrow-derived cDC1s were transferred into Xcr1^DTR mice 1 day prior to vehicle or CCl₄ treatment. (B) Representative flow plots of adoptively transferred YFP⁺ cDC1s from vehicle- (left) or CCl₄-treated (right) mouse livers 24 h after treatment. (C) Percentages of YFP⁺ (IL-12⁺) from vehicle- or CCl₄-treated mouse livers 24 h after treatment (n = 7–8 mice per group). (D) Analysis of liver IL-12⁺ cDC1s from adoptively transferred target (Myd88, Mavs, Aim2, Nlrp3, Tmem173 [STING]) cRNP-edited cells as a percentage of NTC-cRNP-edited cells 24 h after CCl₄ treatment (n = 6 mice per group). (E) Percentages of IL-12⁺ cDC1s in the liver of Il12b^YFP or Il12b^YFP × Sting^Gt mice before and 24 h after CCl₄ injection (n = 6–8 mice per group). (F) Analysis of liver IL-12⁺ cDC1s from adoptively transferred bone marrow-derived Il12b^YFP or Il12b^YFP × Sting^Gt cells 24 h after CCl₄ treatment (n = 7 mice per group). (G) Bone marrow chimeric mice were generated utilizing WT or cGAS^−/− hosts reconstituted with Il12b^YFP bone marrow. Percentages of liver IL-12⁺ cDC1s from WT:Il12b^YFP or cGAS^−/−:Il12b^YFP mice 24 h after CCl₄ treatment (n = 7 mice per group). (H) Bone marrow chimeric mice were generated utilizing WT or cGAS^−/− hosts reconstituted with either WT or cGAS^−/− bone marrow. Plasma ALT concentrations of WT (host):WT (BM), WT:cGAS^−/−, cGAS^−/−:WT, and cGAS^−/−:cGAS^−/− mice 48 h after CCl₄ injection (n = 5 mice per group). (I) Analysis of liver IL-12⁺ cDC1s from adoptively transferred target (Mb21d1 [cGAS]) cRNP-edited cells as a percentage of NTC-cRNP-edited cells 24 h after CCl₄ treatment (n = 6 mice per group). Data are representative of three (D and I) and two (B, C, and E–H) independent experiments, and samples were compared using two-tailed Student’s t test with Welch’s correction, assuming unequal SD. Data are presented as individual points with the mean ± SEM (NS, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001). See also Figure S5.