Type 2 cytokines act on enteric sensory neurons to regulate neuropeptide-driven host defense

Abstract

Enteric nervous system (ENS)-derived neuropeptides modulate immune cell function, yet our understanding of how inflammatory cues directly influence enteric neuron responses during infection is considerably lacking. Here, we characterized a primary enteric sensory neuron (PSN) subset producing the neuropeptides neuromedin U (NMU) and calcitonin gene-related peptide β (CGRPβ) and coexpressing receptors for the type 2 cytokines interleukin-4 (IL-4) and IL-13. Type 2 cytokines amplified NMU and CGRPβ expression in PSNs both in vitro and in vivo, and this was abrogated by PSN-specific Il13ra1 deletion. Deletion of Il13ra1 in PSNs impaired host defense to the gastrointestinal helminth Heligmosomoides polygyrus and blunted muscularis immune responses. Co-administration of NMU23 and CGRPβ rescued helminth clearance deficits and restored anti-helminth immunity, highlighting the essential bidirectional neuroimmune cross-talk regulating intestinal type 2 inflammation.

Fig. 1.. Type 2 cytokines directly amplify PSN neuropeptide expression.

(A and B) scRNA-seq reanalysis of mouse SI enteric neurons integrated from three independent studies (–12). (A) Uniform manifold approximation and projection (UMAP) and (B) violin plots showing indicated gene expression. (C) scRNA-seq of human colon enteric neurons from Drokhlyansky et al. (10). (D and E) Bulk RNA-seq analysis of in vitro enteric neuroglial cultures treated for 24 hours with indicated cytokines (50 ng ml⁻¹ per cytokine). (D) Schematic and volcano plot. (E) DESeq2-normalized gene counts; lower cutoff was set to 1 for log-scale visualization (n = 5 to 9 mice). (F and G) C57Bl/6J mice were administered IL-4 complex intraperitoneally for 3 days (F) or 4 (G) days. (F) Duodenal (Duo), ileal (Ile), and colonic (Col) qPCR gene expression (n = 10 mice). (G) NMU23 and CGRP enzyme immunoassay (EIA) concentrations in ileal lysates. Shown is the fold change to bovine serum albumin (BSA) average per experiment (n = 8 to 9). (H and I) Calb2^ΔIl13ra1 cKO mice and littermate controls administered IL-4 complex. (H) Schematic. (I) Duodenal qPCR (n = 4 to 6 mice). (J and K) iCGRPβ^ΔIl13ra1 cKO mice and littermate controls gavaged with three or five consecutive doses of tamoxifen and administered IL-4 complex for 4 days. (J) Schematic. (K) Duodenal qPCR (n = 7 mice). Data were pooled from two independent experiments [(F), (G), (I), and (K)] or are representative of three independent experiments (E). Data points represent single mice [(F), (G), (I), and (K)] or distinct in vitro cultures derived from single neonatal mice [(D) and (E)]. qPCR gene expression is relative to Actb, represented as 2^–ΔCt × 1000 [(I) and (K)] or fold change to the BSA group per experiment (F). Each graph indicates the mean ± SEM of replicates. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, two-way ANOVA with Sidak’s correction [(F), (I), and (K)] or two-tailed Welch’s t test (G). PIN, putative interneuron; PIFN, putative intestinofugal neuron; PEMN, putative excitatory motor neuron; PIMN, putative inhibitory motor neuron; PSVN, putative secretomotor/vasodilator neuron.

Fig. 2.. PSN-specific Il13ra1 ablation impairs host defense to H. polygyrus.

(A) Confocal imaging and three-dimensional rendering (Aivia) of Ce3D-cleared duodenal muscularis at 6 d.p.i. Autofluorescent H. polygyrus larvae are pseudocolored green, white, and pink; enteric neurons/β3-tubulin are shown in red; and CGRP is shown in cyan. Insets (yellow boxes) show magnified white larvae from the top (left inset) and the side (right inset). (B) Duodenal qPCR of muscularis (black) and lamina propria/submucosa (red) at the indicated d.p.i. Gene expression is relative to Tubb3, represented as fold change to the water group per organ. Statistics compare indicated d.p.i. against the water group per organ. Data were pooled from four independent time points (n = 3 to 4 mice per time point). (C to E) H. polygyrus–infected Calb2^ΔIl13ra1 cKO and littermate control (Calb2^cre/+, Il13ra1^flox, or heterozygous) mice at 14 d.p.i. (C) Schematic. (D) Fecundity (n = 12 to 13 mice). (E) Worm count (n = 15 to 20 mice). Squares indicate mice that received daily PBS. (F to H) H. polygyrus–infected tamoxifen-induced iCGRPβ^ΔIl13ra1 cKO and littermate control mice. (F) Schematic. (G) Fecundity (n = 12 mice). (H) Worm count (n = 12 to 13 mice). (I to K) H. polygyrus-infected Phox2b^ΔIl13ra1 cKO and littermate control mice at 14 d.p.i. (I) Schematic. (J) Fecundity (n = 11 to 12 mice). (K) Worm count (n = 11 to 12 mice). Data were pooled from two [(D), (H), (J), and (K)], three (G), or four (E) independent experiments. Each data point represents a single mouse [(C) to (K)], and each bar and scatter plot indicate the mean ± SEM of replicates. ns, not significant; *P < 0.05; **P < 0.01, two-tailed Welch’s t test [(E), (H), and (K)], two-tailed Mann-Whitney test [(D), (G), and (J)], or two-way ANOVA with Dunnett’s correction (B).

Fig. 3.. Type 2 cytokines maintain PSN population and neuropeptide expression.

(A to C) Duodenal qPCR data from the indicated Calb2^ΔIl13ra1 cKO mice (n = 8 to 14 mice) (A), 14 d.p.i. tamoxifen-induced iCGRPβ^ΔIl13ra1 cKO mice (n = 12 to 13 mice) (B), and 14 d.p.i. Phox2b^ΔIl13ra1 cKO mice (n = 7 mice) (C) and respective littermate controls. (D and E) Scatter plots correlating 14 d.p.i. duodenal qPCR gene expression (x axes) and fecundity (y axes) (D) or duodenal qPCR expression of indicated genes from matched experiments (E) (n = 41 mice). (F and G) Calb2^ΔIl13ra1 cKO mice administered the indicated neuropeptides intraperitoneally (2 nmoles each per injection) or PBS daily. (F) Fecundity on indicated d.p.i. (G) 14 d.p.i. worm count (n = 4 to 11 mice). (H to K) Confocal microscopy of duodenal cross-sections. (H) Representative villi at 7 d.p.i. (I) Percentage CGRP⁺ area of total 4ʹ,6-diamidino-2-phenylindole (DAPI) area per villus (left) and per mouse (right) (n = 4 mice). (J) Representative villi. (K) Percentage calretinin⁺ area of total DAPI area per villus (left) and per mouse (right) (n = 4 to 6 mice). (L to O) Whole-mount confocal imaging of 7 d.p.i. duodenal myenteric plexi. (L) Calretinin⁺p75NTR⁺ neurons per square millimeter of β3-tubulin⁺ area (n = 6 to 7 mice). (M) Representative 7 d.p.i. myenteric plexi immunostained for β3-tubulin (yellow), calretinin (magenta), and ANNA-1 (neuronal nuclei/cell body marker; cyan). Arrows indicate large calretinin⁺ PSNs. (N) Scatter plot of pooled segmented neurons at 7 d.p.i. from littermate control (n = 5) and Calb2^ΔIl13ra1 cKO (n = 4) mice. (O) Percentage of large calretinin⁺ neurons (n = 4 to 6 mice). Data were pooled from two [(B), (K), (O)], three [(D) to (G)], or six (A) independent experiments or are representative of two or more independent experiments [(C), (I), and (L)]. Data points represent single mice or individual villi [(I), left; (K), left]. Scale bars, 100 μm [(H), (J), and (M)]. qPCR gene expression is relative to Actb, represented as 2^–ΔCt × 1000 [(D) and (E)] or fold change to the cKO group per experiment [(A) to (C)]. Each bar graph indicates the mean ± SEM of replicates. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, two-way ANOVA with Sidak’s correction [(A), (B), and (O)] or Tukey’s correction (K), mixed-effects analysis with Tukey’s correction (F), one-way ANOVA with Tukey’s correction (G), two-tailed t test [(C), (I), and (L)], or simple linear regression [(D) and (E)].

Fig. 4.. PSN IL-4/IL-13 sensing promotes muscularis-specific type 2 immune responses.

(A) Duodenal qPCR gene expression at 14 d.p.i. (n = 9 to 19 mice). (B) Duodenal qPCR gene expression in mice with daily CGRPβ and NMU23 co-administration as in Fig. 3, F and G (n = 3 to 4 mice). (C) DESeq2-normalized bulk RNA-seq counts of ileal muscularis and LP ILC2s at 7 d.p.i. as gated in fig. S11A (n = 4 to 5 mice). (D and E) Flow cytometric analysis of ileal muscularis and LP immune cells at 7 d.p.i. (D) Representative flow plots gated on single live CD45⁺ CD90.2^–Lineage(Ter-119, NK1.1, CD19, B220)⁻ cells. (E) Percentage of muscularis and LP eosinophils (as gated in fig. S11A) of all CD45⁺ cells (n = 4 to 5 mice). (F to H) scRNA-seq of ileal muscularis (input 50% enriched for stroma or glia as in fig. S13A) and duodenal epithelium at 7 d.p.i. (F) UMAP embedding of 8270 individual cells pooled from two mice per group using general-level cluster annotation. (G) Gene set enrichment analysis (GSEA) plots of general MMφ cluster, depicting Gene Ontology gene sets from MSigDB. (H) Violin plots of general MMφ clusters. (I) Duodenal muscularis qPCR gene expression at indicated d.p.i. (n = 3 to 6 mice per time point). (J and K) Flow cytometry of duodenal MMφ (as gated in fig. S11C) at 3 d.p.i. (J) Percentage CD301b⁺ of CD45⁺ cells in indicated MMφ subsets at 3 d.p.i. (n = 4 to 6 mice). (K) Percentage Arg1⁺ MMφ (MHC2⁺Ly6C^–) of CD45⁺ immune cells (n = 3 to 6 mice per time point). (L and M) Flow cytometry of ileal muscularis and LP immune cells at 14 d.p.i. (L) Representative flow plots; shown are total immune cells (top) and macrophages (bottom). (M) Frequency of Arg1⁺ cells in indicated MMφ subsets (n = 5 to 14 mice). Data were pooled from two [(A), (K), and (M)] or four (I) independent experiments or are representative of two or more independent experiments [(B) and (E)]. Each data point represents a single mouse. qPCR gene expression is relative to Actb, represented as 2^–ΔCt × 1000 (I) or fold change to the cKO group (A) or PBS-treated cKO group (B) per experiment. Each bar graph indicates the mean ± SEM of replicates. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, two-tailed t test (A), two-way ANOVA with Sidak’s correction [(B), (C), (E), (I) to (K), and (M)], or Wilcoxon rank-sum test adjusted P value (H).

Fig. 5.. CGRPβ synergizes with type 2 cytokines to drive MMφ Arg1.

(A) CGRP EIA peptide concentration (n = 3 to 8 mice). (B) scRNA-seq dot plot across general immune cell clusters (including lymphoid cell subclusters) from the dataset shown in Fig. 4F. (C to E) qPCR gene expression in duodenal SI-Mφ cultures (n = 3 to 5 wells) (C) or BMDM cultures (n = 4 to 8 wells) (D and E) treated for 8 hours with or without indicated neuropeptides (250 nM each) or IL-4 concentrations. (F to H) Flow cytometry of 14 d.p.i. ileal muscularis immune cells. (F) Representative flow plots depicting percentage Arg1⁺ MMφ (single live CD45⁺ CD3ε^–Ly6G^–Siglec-F^–CD64⁺ CD11b⁺). (G) Percentage Arg1⁺ muscularis immune cells (left) and MMφ (right) (n = 8 mice). Solid and dotted lines with grayed area represent the mean ± SEM of control mice at 14 d.p.i. pooled from four independent experiments (n = 16 mice). (H) Scatter plot correlating 14 d.p.i percentage Arg1⁺ muscularis immune cells (x axis) with worm count (y axis) (n = 16 mice). (I) The 14 d.p.i. fecundity of Ramp1 KO and littermate control (Ramp1^+/−) mice (n = 5 to 8 mice). Data were pooled from two independent experiments [(G) to (I)] or are representative of two or more independent experiments [(A), (C), (D), and (E)]. Data points represent individual mice [(A) and (G) to (I)] or culture wells pooled from two to five mice per experiment [(C) to (E)]. qPCR gene expression is relative to Actb, represented as fold change to media control group per experiment [(C) to (E)]. Each graph indicates the mean ± SEM of replicates. ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 using two-way ANOVA with Tukey’s correction (A), one-way ANOVA with Tukey’s correction [(D) and (E)], two-tailed unpaired t test (C), two-tailed Mann-Whitney test (G), two-tailed Welch’s t test (I), or simple linear regression (H).