T-cell derived acetylcholine aids host defenses during enteric bacterial infection with Citrobacter rodentium

Abstract

The regulation of mucosal immune function is critical to host protection from enteric pathogens but is incompletely understood. The nervous system and the neurotransmitter acetylcholine play an integral part in host defense against enteric bacterial pathogens. Here we report that acetylcholine producing-T-cells, as a non-neuronal source of ACh, were recruited to the colon during infection with the mouse pathogen Citrobacter rodentium. These ChAT+ T-cells did not exclusively belong to one Th subset and were able to produce IFNγ, IL-17A and IL-22. To interrogate the possible protective effect of acetylcholine released from these cells during enteric infection, T-cells were rendered deficient in their ability to produce acetylcholine through a conditional gene knockout approach. Significantly increased C. rodentium burden was observed in the colon from conditional KO (cKO) compared to WT mice at 10 days post-infection. This increased bacterial burden in cKO mice was associated with increased expression of the cytokines IL-1β, IL-6, and TNFα, but without significant changes in T-cell and ILC associated IL-17A, IL-22, and IFNγ, or epithelial expression of antimicrobial peptides, compared to WT mice. Despite the increased expression of pro-inflammatory cytokines during C. rodentium infection, inducible nitric oxide synthase (Nos2) expression was significantly reduced in intestinal epithelial cells of ChAT T-cell cKO mice 10 days post-infection. Additionally, a cholinergic agonist enhanced IFNγ-induced Nos2 expression in intestinal epithelial cell in vitro. These findings demonstrated that acetylcholine, produced by specialized T-cells that are recruited during C. rodentium infection, are a key mediator in host-microbe interactions and mucosal defenses.

Fig 1. Infection with C. rodentium recruits ChAT+ T-cells to the colon.

(A) Confocal microscopy was conducted on colonic tissues from LB or C. rodentium infected ChAT-GFP reporter mice using DAPI, anti-CD3, and anti-GFP. (B) These CD3+ ChAT-GFP+ cells (inset “I”) where then quantified in 5 random fields from LB control, or 6, 10, 21, or 30 days post-infection with C. rodentium. * P < 0.05 ANOVA, with n = 5–6 mice per timepoint.

Fig 2. Flow cytometric characterization of ChAT-GFP+ T-cells.

(A) Lamina propria lymphocytes were isolated from ChAT-GFP mice 10 days after C. rodentium infection, or (LB gavaged) controls. Representative gating strategy depicts analysis of (lymphocytes, single, cells, live CD3+). CD4+ ChAT-GFP+ or CD4+ ChAT-GFP- cells expressing IFNγ, IL-17A, or IL-22 in each of these populations (light grey: non-stimulated control, green; ChAT-GFP+ cells stimulated with PMA/ionomycin, dark grey: ChAT-GFP- cells stimulated with PMA/ionomycin). The amount of each cytokine was quantified by MFI (B), and the frequency of these cells per live CD3+ cell (C). * P < 0.05 ANOVA, with n = 4–6 mice per group.

Fig 3. Increased C. rodentium colonization in ChAT T-cell cKO mice without increased histopathology.

(A) The number of colonic tissue adherent bacteria were assessed in WT and ChAT T-cell cKO mice 10 days p.i. (B) Confocal imaging of uninfected and infected WT and ChAT T-cell cKO mice demonstrates increased adherent bacteria in the luminal surface of intestinal epithelial cells (IEC, CDH1+ DAPI+). (C) Infection of WT and ChAT T-cell cKO mice induces crypt hyperplasia as indicated by significantly increased crypt lengths. (D&E) and Ki67+ IEC compared to uninfected controls. Data are representative from 3 separate experiments with 3–5 mice per group. ND = not detected, *, # P < 0.05 ANOVA.

Fig 4. Conditional ablation of ChAT in T-cells increases C. rodentium-induced innate inflammatory genes.

The immune response during infection was assessed through qRT-PCR conducted on colonic tissues from LB or C. rodentium infected WT or ChAT T-cell cKO mice. Expression of the cytokines Il1-β, Il-6, and Tnfα, Ifnγ, Il-17a and Il-22 was evaluated in naïve and infected WT and ChAT T-cell cKO mice. *, # P < 0.05 ANOVA, with n = 8–10 mice per group.

Fig 5. Loss of T-cell derived ACh reduces Nos2 expression during C. rodentium infection.

Expression of genes characteristic of macrophage polarization were assessed by qRT-PCR on colonic tissue from LB or C. rodentium infected WT or ChAT T-cell cKO mice. *, # P < 0.05 ANOVA, with n = 8–10 mice per group.

Fig 6. NOS2 is reduced in C. rodentium infected Intestinal epithelial cells from ChAT T-cell cKO mice.

(A) Localization and quantification of NOS2 was conducted on fixed colonic tissues from LB and C. rodentium infected WT and ChAT T-cell cKO mice, 10 days p.i. Confocal microscopy was conducted using DAPI and antibodies directed against epithelial e-cadherin (CDH1), and NOS2. (B) Expression of NOS2 per IECs (DAPI+ CDH1+ cells) was then quantified using Imaris. (C) In a separate cohort of mice, NOS2 expression was evaluated 10 days p.i in isolated IEC by flow cytometry. IEC were identified as live/single/CD45- EpCAM+ cells. Expression of NOS2 was determined by MFI and is summarized in (D). *P <0.05 ANOVA n = 4–8 mice/group.

Fig 7. Nos2 expression induced by IFNγ in IEC is enhanced by cholinergic agonists.

CMT-93 cells were stimulated with vehicle (control), Carbachol (‘Cch’ 100 μM), IFNγ (1 ng/mL), or IFNγ (1ng/mL) + Cch (100 μM) for 3h. Expression of IFNγ-induced genes Ciita, Irf1 and Nos2 was quantified by qRT-PCR. *, # P<0.05, **P < 0.01, ANOVA, n = 4 independent experiments.