Anti-commensal IgG Drives Intestinal Inflammation and Type 17 Immunity in Ulcerative Colitis

Abstract

Inflammatory bowel disease is a chronic, relapsing condition with two subtypes, Crohn's disease (CD) and ulcerative colitis (UC). Genome-wide association studies (GWASs) in UC implicate a FCGR2A variant that alters the binding affinity of the antibody receptor it encodes, FcγRIIA, for immunoglobulin G (IgG). Here, we aimed to understand the mechanisms whereby changes in FcγRIIA affinity would affect inflammation in an IgA-dominated organ. We found a profound induction of anti-commensal IgG and a concomitant increase in activating FcγR signaling in the colonic mucosa of UC patients. Commensal-IgG immune complexes engaged gut-resident FcγR-expressing macrophages, inducing NLRP3- and reactive-oxygen-species-dependent production of interleukin-1β (IL-1β) and neutrophil-recruiting chemokines. These responses were modulated by the FCGR2A genotype. In vivo manipulation of macrophage FcγR signal strength in a mouse model of UC determined the magnitude of intestinal inflammation and IL-1β-dependent type 17 immunity. The identification of an important contribution of IgG-FcγR-dependent inflammation to UC has therapeutic implications.

Keywords: Fcγ receptors; IL-1β; IgG; inflammatory bowel disease; type 17 immunity.

Graphical abstract

Figure 1

Anti-commensal IgG Is Associated with the Magnitude of Intestinal Inflammation (A) Analysis of human Ig heavy-chain gene transcripts in healthy and UC colonic biopsies. Data were derived from Gene Expression Omnibus (GEO) dataset GEO: GSE9452. (B) IgG- and IgA1- and IgA2-bound SYBR green^hi microbes in UC and household healthy control (HHC) stool samples were analyzed as household pairs (n = 6 per group). (C) Correlation of pooled IgG- and IgA1- and IgA2-bound bacterial levels with clinical activity index (CAI) (n = 12). (D) Murine colon luminal IgG and IgA levels following two cycles of DSS administration (cDSS) and normalized to total protein content (n = 10 per group). Medians are indicated. (E) Quantification of IgG-bound bacteria in stool after a single acute course of 6-day 2% DSS administration (aDSS) or H₂O with (red) and without (black) paired serum pre-incubation (n = 6–9 per group). Medians are indicated. (F) Correlation of IgG-bound commensals (no serum) from pooled control and colitic mice at day 21 after aDSS with markers of colonic inflammation; length and neutrophil count (n = 20) are shown. In (D)–(F), data are pooled from two independent experiments. (G) Opsonization of commensal bacterial species with day 21 aDSS serum or healthy control serum (day 0) (n = 5 per group). Medians are indicated. Data are representative of two independent experiments. (H) Confocal image of control or inflamed colons from cDSS-treated mice (red, IgG; green, Ki67; white, phalloidin). Data representative of three independent experiments. (I and J) Quantification of IgG- and IgA-expressing IgM⁻ B cell subsets in murine colons (I) and colon-draining MLN (J) at day 21 after aDSS administration versus controls (day 0) (n = 5 per group). For absolute cell-count quantification plots, medians are indicated. Data are representative of three independent experiments. p values were calculated via limma with multiple correction using the Benjamini-Hochberg (BH) procedure (A), ratio paired t test (B), linear regression analysis (C and F), or the nonparametric Mann-Whitney U test (D, E, G, I, and J). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S1 and Table S1.

Figure 2

Activating FcγR Signaling in Intestinal Inflammation (A) Analysis of the top significantly differentially expressed genes (log2 FC > 2) in human active UC mucosal biopsies compared with controls. Data are from GEO: GSE38713. (B) FCGR enrichment in inflamed active UC (n = 15) compared with healthy control (n = 13), remission (n = 8), and non-inflamed mucosal biopsies (n = 7). Data are from GEO: GSE38713. (C) FCGR expression in colonic biopsies from UC patients refractory (n = 16) or responsive (n = 8) to infliximab treatment and healthy controls prior to treatment initiation (n = 6). Data are from GEO: GSE16879. Minimum to maximum box-and-whisker plots are shown in (B) and (C). (D) Area under the receiver operator curve (AUROC) analysis for FCGR2A expression in infliximab-refractory mucosal biopsies compared with infliximab-responsive biopsies. Data were derived as in (C). (E) Gene-set enrichment analysis (GSEA) of the KEGG FcγR-mediated phagocytosis pathway in UC and healthy colonic biopsies (left) and cDSS-inflamed and healthy murine colons (bottom). Data are from GEO: GSE38713 (UC = 15, HC = 13) and GEO: GSE42768 (cDSS = 5, H₂O = 5). (F) FcγR expression by murine colonic leukocytes at day 7 after aDSS. Data are representative of 3 independent experiments. (G) qPCR of FcγR mRNA expression in flow-sorted murine colonic CX3CR1⁺Ly6C^loMHC-II^hi macrophages, as identified in (F). Data are representative of 2 independent experiments. (H) Single-cell RNA-seq of immune cell subsets in healthy human colon. Data are from GEO: GSE81861. The heatmap illustrates the top 50 cell-type-specific markers ranked by AUROC score. (I) qPCR of FcγR mRNA expression in flow-sorted human ileal CD14⁺CX3CR1⁺SSC^int macrophages. (J) Colonic “monocyte waterfall” subset quantification by flow cytometry at day 7 after aDSS versus controls (n = 3 per group). Flow plots of CD11b⁺CX3CR1⁺ waterfall subsets (left) and quantification of absolute cell counts for the indicated subsets (right) are shown. For absolute cell counts, medians are indicated. Data are representative of 3 independent experiments. (K) FcγR quantification of CX3CR1⁺Ly6C^loMHC-II^hi macrophages, as shown in (J), by flow cytometry. Means ± SEM are indicated and normalized to the A:I ratio of control-treated mice. Data are representative of two independent experiments. p values were calculated via limma with multiple correction using the BH procedure (A–C), AUROC analysis (D and H), or Student’s two-tailed t test (J and K). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S2 and Table S2.

Figure 3

Mucosal FcγR Expression Correlates with IL-1β and CXCL8 (A) Log2 fold-change comparison of differentially expressed cytokine and chemokine genes across two independent UC cohorts. Data are from GEO: GSE38713 (UC = 15, HC = 13) and GEO: GSE59071 (UC = 74, HC = 11). (B and C) Correlation summary of FCGR2A expression with UC-associated cytokine and chemokine gene transcripts (B) and correlation with IL1B and CXCL8 (C) in n = 85 mucosal biopsies. Data are from GEO: GSE59071. (D) Hierarchical clustering of cytokine and chemokine and FCGR2A expression in healthy controls and non-inflamed and inflamed UC patients. Data are derived from GEO: GSE38713. (E and F) Correlation of Fcgr3 expression with Il1b (E) and candidate-gene (F) expression in whole inflamed colonic tissue by qPCR (n = 16). Data are representative of two independent experiments. (G) IL1B and CXCL8 expression in healthy human LPMC stimulated with Ova or Ova-IC for 16 h (n = 5 per condition). Means ± SEM are indicated. Data are representative of two independent experiments. (H) IL-1β, CXCL1, and CXCL2 in supernatants of inflamed murine LPMCs stimulated with Ova or Ova-IC for 16 h (n = 3-9). Paired samples represent LPMCs from a single mouse. Data are pooled from two independent experiments. (I) Flow-cytometry profiling of pro-IL-1β-expressing cells at day 14 after aDSS (n = 6). Medians are indicated. Data are representative of three independent experiments. (J) Pro-IL-1β expression by colonic CX3CR1⁺ MNP subsets and neutrophils in Rag2^−/− mice treated with control or anti-flagellin-enriched serum IgG and 7-day aDSS or uninflamed H₂O-treated controls (n = 5 per group). Medians are indicated. (K) Day 7 weight loss for aDSS-treated mice shown in (J) (n = 5 per group). Medians are indicated. p values were calculated via limma with multiple correction using the BH procedure (A), linear regression analysis (B, C, E, and F), Student’s two-tailed t test (G), ratio paired t test (H and I), or the Kruskal-Wallis test with Dunn’s multiple-comparisons test (J). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S3.

Figure 4

IgG-Induced IL-1β Production by Human Macrophages (A) Venn diagram of significant co-expressed genes between UC and Ova-IC-stimulated colonic macrophages (left) and IC-induced UC-associated cytokines and chemokines (right), as determined by sdef. Human UC data are from GEO: GSE38713 (UC = 15, HC = 13). For macrophage stimulation, n = 4 per condition. (B) GSEA of inflammasome genes (left) and heatmap of selected core enrichment genes (right) in flow-sorted murine intestinal CX3CR1⁺Ly6C^loMHC-II^hi macrophages stimulated with Ova and Ova-IC for 4 h (n = 4 per condition). (C) IL-1β production by WT and Nlrp3^−/− BMDMs primed with Ova and Ova-IC for 4 h followed by 30 min of ATP stimulation (^∗∗∗∗, WT versus Nlrp3^−/− Ova-IC; ####, WT Ova versus Ova-IC) (n = 3 per condition). Means ± SEM are indicated. Data are representative of three independent experiments. ND, not detected. (D) IL-1β ELISA of human MDMs stimulated with plate-coated IgG (c-IgG) and intestinal commensals (Comm.) for 24 h ± anti-FcγRIIA IgG-blocking antibody (n = 3 per condition). Means ± SEM are indicated. Data are representative of two independent experiments. (E) qPCR of IL1B mRNA (left, n = 3 per condition) and western blot of IL-1β in cell lysates and cell-free supernatants (right) from human MDMs as stimulated in (D). For qPCR (left), means ± SEM are indicated. Data are representative of two independent experiments. (F) The effect of NLRP3 inhibition on IgG-induced MDM IL-1β production. Western blot (left) and IL-1β production (left, ELISA) by human MDMs stimulated as in (D) after NLRP3 “Trim-Away” or ± NLRP3 inhibitor MCC950 (right, ELISA) (n = 3 per condition). Means ± SEM are indicated. (G) IL-1β production by human MDMs stimulated as in (D) ± mitochondrial ROS inhibitor mitoTEMPO (n = 3 per condition). Means ± SEM are indicated. Data are representative of 2 independent experiments. (H) Fold change in IL1B expression in Ova-IC-stimulated MDMs, normalized to R/R fold change (n = 6 per group). Data are pooled from 3 independent experiments. Medians are indicated. (I) Schematic of FcγR A:I ratios with FCGR2A SNP (dbSNP: rs1801274) and murine transgenic models. (J) Il1b induction in WT, Fcgr2b^−/−, and M-TG BMDMs stimulated with Ova and Ova-IC for 3 h (n = 3 per condition). Means ± SEM are indicated. Data are representative of 3 independent experiments. p values were calculated via the standard DESeq2 method with multiple correction using the BH procedure (A and B), two-way ANOVA with Tukey’s multiple comparisons test (C, F, and J), one-way ANOVA with Tukey’s multiple-comparisons test (D, E, and G), or parametric Student’s t test (H). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S4.

Figure 5

MNP FcγR A:I Ratio Modulates Intestinal Inflammation (A) Weight loss in co-housed sex-matched WT (black circles, n = 8) and Fcgr2b^−/− (red circles, n = 9) mice after a single 6-day course of 2% DSS. Means ± SEM are indicated. Data are representative of 3 independent experiments. (B) Confocal microscopy of colonic CD45⁺ leukocytes in mice treated as in (A). Data are representative of 2 independent experiments. (C) Flow-cytometric quantification of colonic neutrophil frequency (top) and absolute neutrophil counts (bottom) in WT and Fcgr2b^−/− colons at day 25 after aDSS (n = 5–7 per group) or in healthy controls (day 0) (n = 3–5 per group). Medians are indicated. Data are representative of 3 independent experiments. (D) Pro-IL-1β expression by colonic CD11b⁺CX3CR1⁺ MNPs at day 25 after aDSS (n = 5–7 per group). (E) Frequency of pro-IL-1β expression within the Ly6C^hiMHC-II⁻ monocyte and Ly6C⁺MHC-II^int macrophage populations as shown in aDSS (n = 5–7 per group) versus controls (day 0) (n = 3–5 per group). Medians are indicated. Data are representative of 3 independent experiments. (F) Chemokine mRNA levels in flow-sorted Ly6C^loMHC-II^hi colonic macrophages at day 21 after aDSS (n = 4–5 per group). Data normalized to WT mRNA levels. Medians are indicated. (G and H) Weight loss (G) and clinical colonic and lymphoid organ features (H) of co-housed sex-matched M-TG (green circles, n = 5 or 6) and N-TG littermate controls (black circles, n = 6) after aDSS or in healthy controls (day 0 in H). Means ± SEM (weight loss) and medians (colon, MLN, and spleen weight) are indicated. Data are representative of 3 independent experiments. (I) Flow-cytometric quantification of colonic neutrophil frequency (top) and absolute neutrophil count (bottom) in N-TG and M-TG mice at day 21 after aDSS (n = 6 per group) versus controls (day 0) (n = 5–6 per group). Medians are indicated. (J) Pro-IL-1β expression by colonic CX3CR1⁺ MNP subsets in N-TG and M-TG mice treated as in (I) (n = 5 or 6 per group). (K) Frequency (top) and absolute cell count (bottom) of pro-IL-1β-expressing MNP subsets in N-TG and M-TG mice treated as in (I). Medians are indicated. Data are representative of two independent experiments. p values calculated using a two-way ANOVA with Bonferroni’s multiple comparisons test (A and G), or the nonparametric Mann-Whitney U test (C-F, H-K). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S5.

Figure 6

MNP FcγR A:I Ratio Modulates Intestinal Type 17 Immunity (A) Type-17-associated cytokine production by day-14 LPMCs stimulated with Ova and Ova-IC for 16 h (n = 5–9). Paired samples represent cells isolated from the same mouse. Data are pooled from 2 independent experiments. (B) qPCR of type 17 cytokines in WT and Fcgr2b^−/− whole colonic tissue after cDSS versus controls (n = 4–9 per group). Data are normalized to uninflamed healthy colon. Medians are indicated. Data are representative of 2 independent experiments. (C) Flow-cytometry plots of colonic IL-17A-expressing T cell subsets in co-housed sex-matched WT and Fcgr2b^−/− mice at day 21 after aDSS (n = 6–8 per group) versus controls (day 0) (n = 3–5 per group). (D) Quantification of absolute numbers of colonic IL-17A-producing T cells shown in (C). Medians are indicated. Data are representative of 3 independent experiments. (E) Quantification of absolute numbers of colonic IL-17A-producing T cell subsets in co-housed WT and Fcgr2b^−/− mice at day 15 after aDSS and weekly treatment with anti-IL-1R1 IgG-blocking antibody or control IgG (n = 5–7 per group). Medians are indicated. Data are representative of two independent experiments. (F) Colonic IL-17A-expressing T cell subsets in M-TG and N-TG littermate controls at day 21 after aDSS (n = 5 or 6 per group) versus controls (n = 5 or 6 per group). Data are representative of two independent experiments. (G) Quantification of absolute cell counts of colonic IL-17A-producing T cell subsets as shown in (F). Medians are indicated. (H) Weight loss of Fcgr2b^−/− mice treated with control or anti-IL-1R1 IgG antibodies after aDSS treatment (n = 5–7 per group). Mean ± SEM are indicated. Data are representative of two independent experiments. (I) Colonic neutrophil infiltration in WT and Fcgr2b^−/− mice treated as in (H) (n = 5–7 per group). Medians are indicated. p values were calculated with a ratio paired t test (A), the nonparametric Mann-Whitney U test (B–G and I), or a two-way ANOVA (H). ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001. See also Figure S6.