Myeloid Cells and Sphingosine-1-Phosphate Are Required for TCRαβ Intraepithelial Lymphocyte Recruitment to the Colon Epithelium

Abstract

Intraepithelial lymphocytes (IELs) are T cells important for the maintenance of barrier integrity in the intestine. Colon IELs are significantly reduced in both MyD88-deficient mice and those lacking an intact microbiota, suggesting that MyD88-mediated detection of bacterial products is important for the recruitment and/or retention of these cells. Here, using conditionally deficient MyD88 mice, we show that myeloid cells are the key mediators of TCRαβ+ IEL recruitment to the colon. Upon exposure to luminal bacteria, myeloid cells produce sphingosine-1-phosphate (S1P) in a MyD88-dependent fashion. TCRαβ+ IEL recruitment may be blocked using the S1P receptor antagonist FTY720, confirming the importance of S1P in the recruitment of TCRαβ+ IELs to the colon epithelium. Finally, using the TNFΔARE/+ model of Crohn's-like bowel inflammation, we show that disruption of colon IEL recruitment through myeloid-specific MyD88 deficiency results in reduced pathology. Our results illustrate one mechanism for recruitment of a subset of IELs to the colon.

Figure 1:. Myeloid cells require MyD88 for TCRαβ+ IEL recruitment to the colon.

IELs were harvested from colon epithelial tissue of 8–10-week-old male and female (A) Lck-Cre⁺ MyD88^fl/fl mice (n=5) and Cre⁻ littermates (n=7) and (B) Villin-Cre⁺ MyD88^fl/fl mice (n=8) and Cre⁻ littermates (n=10) and enumerated by flow cytometry for CD3+ IELs. (C-G) IELs were harvested from colon epithelial tissue of 8–10-week-old male and female LysM-Cre⁺ MyD88^fl/fl mice (n=10) and Cre⁻ littermates (n=9) and enumerated by flow cytometry for (C) TCRαβ+, (D) TCRγδ+, (E) TCRαβ+ CD4+, (F) TCRαβ+ CD8+, and (G) TCRαβ+ CD4− CD8− DN IELs. Data are shown as individual animals across three separate experiments with bars as mean ± SEM. Littermate groups are indicated with matching symbols. *, P<0.05 as determined by Student’s t-test.

Figure 2:. S1P is a potential IEL chemoattractant.

Colon IELs and lamina propria T cells (LPLs) were harvested from 8–10-week-old male and female C57Bl/6 mice after treatment for one week with broad-spectrum antibiotics followed by recolonization for one week by housing with dirty bedding from unmanipulated littermates. (A) RNA expression of cell surface receptors and proteins involved in migration were assayed by microarray. The top genes enriched >5-fold in colon IELs versus LPLs from 10 pooled mice are shown. (B) Expression of S1pr1 normalized to β-actin in IELs (n=6) versus LPLs (n=7) was confirmed by qPCR. Symbols are the individual sample fold change ΔΔCt to the mean ΔΔCt for LPLs. Bars are the mean ± SEM. (C) Whole colon tissue was assayed for S1P by liquid chromatography-mass spectrometry (LC-MS) in specific pathogen free (SPF), antibiotic-treated SPF, and germ-free 8–10-week-old C57Bl/6 male and female mice. *, P<0.05; **, P<0.01 as determined by Student’s t-test.

Figure 3:. Myeloid cells produce S1P in response to stimulation with cecal contents in a MyD88-dependent fashion.

Bone marrow derived myeloid cells (BMDMs) were stimulated with 5 ng/mL of heat-killed cecal contents. (A) At the indicated times after stimulation, BMDMs were harvested, RNA was isolated, and Sphk1 expression was assayed by qPCR and normalized to β-actin. Data are the mean ± SEM fold change from time point 0 hours. (B) BMDMs generated from LysM-Cre⁺ MyD88^fl/fl mice and Cre⁻ littermate controls were stimulated with cecal contents for 12 hours and assayed by qPCR for Sphk1 expression normalized to β-actin. Data are the mean ± SEM fold change from Cre⁺ mice. (C) T cell migration towards stimulated versus unstimulated C57Bl/6 BMDMs was assayed using a transwell system. Data are expressed as percent of T cells recovered in the bottom chamber out of total T cells from both chambers. Symbols represent individual data points and bars are the mean ± SEM. (D) The transwell assay was repeated using BMDMs from LysM-Cre⁺ MyD88^fl/fl and LysM-Cre⁻ MyD88^fl/fl mice. Additionally, for one condition with LysM-Cre⁻ MyD88^fl/fl BMDMs, T cells were pre-treated with the S1P receptor inhibitor FTY720. Data are expressed as percent of T cells recovered in the bottom chamber out of total T cells from both chambers. Symbols represent individual data points and bars are the mean ± SEM. (E) 8–10-week-old male and female germ-free C57Bl/6 mice were colonized with cecal contents from C57Bl/6 SPF mice, and viable CD3+ TCRαβ+ IELs enumerated by flow cytometry at the indicated time points. Symbols represent individual mice while bars are the mean ± SEM. (F) CD11c+ myeloid cells were magnetically isolated from the lamina propria of colonized germ-free mice at the indicated times and assayed by qPCR for Sphk1 expression relative to β-actin. Data are the fold change in Sphk1 expression compared to germ-free mice with symbols representing individual mice and the bars as the mean ± SEM.*, P<0.05; **, P<0.01 as determined by Student’s t-test.

Figure 4:. Blockade of S1P receptors significantly reduces recruitment of TCRαβ+ IELs to the colon.

Male and female 8–10-week-old C57Bl/6 mice were treated with broad-spectrum antibiotics to deplete their microbiomes and colon IELs. At the start of recolonization stimulated by housing with dirty bedding of unmanipulated littermates, mice were treated with 1 mg/kg FTY720 (n=5) or 10% DSMO (n=5). After seven days, colon IELs were harvested from and evaluated by flow cytometry. Absolute numbers of (A) total TCRαβ+, (B) TCRαβ+ CD4+, (C) TCRαβ+ CD8+, (D) TCRαβ+ CD4− CD8− DN IELs, and (E) TCRγδ+ IELs are shown for each mouse (symbols) and the mean ± SEM (bars). *, P<0.05; **, P<0.01 as determined by Student’s t-test while ns=not significant. Littermates are indicated by matching symbols.

Figure 5:. Loss of MyD88 in myeloid cells ameliorates ileitis and colitis in TNF^ΔARE/+ mice.

Male and female LysM-Cre⁺ MyD88^fl/fl TNF^ΔARE/+ (n=16) and LysM-Cre⁻ MyD88^fl/fl TNF^ΔARE/+ (n=10) littermate controls were euthanized at 14 weeks of age and tissues were processed for histologic analysis. (A-F) Histology of the ileum was evaluated for active and chronic inflammation (A-B), villus architecture (C), and overall injury score (D) in three to five fields viewed at 200x. The average injury score for each individual mouse is represented as a symbol while the bars are the mean ± SEM for the group. Representative images of the ileum from a LysM-Cre⁻ MyD88^fl/fl TNF^ΔARE/+ (E) and a LysM-Cre⁺ MyD88^fl/fl TNF^ΔARE/+ (F) mouse at 200x are shown. (G-L) Histology of the colon was evaluated for inflammation severity (G), injury extent (H), crypt damage (I), and total injury score (J) along the entire length of the colon tissue at 200x. Representative images of the colon from a LysM-Cre⁻ MyD88^fl/fl TNF^ΔARE/+ (K) and a LysM-Cre⁺ MyD88^fl/fl TNF^ΔARE/+ (L) mouse at 200x are shown. Size bar indicates 100 μm.

Figure 6:. IELs alone are not sufficient to drive ileitis and colitis in Rag^−/− mice.

IELs from TNF^ΔARE/+ (n=3) and TNF^+/+ (n=4) littermate controls were isolated and transferred into Rag^−/− recipients by tail vein injection. After ten weeks, experimental animals and Rag^−/− controls (n=4) were euthanized and tissues were processed for histologic analysis. Tissues were evaluated for total injury score (A-B) as in Figure 5 and number of isolated lymphoid follicles (ILFs) per field (C-D). Histology of the ileum was assessed in at least three fields viewed at 200x (E-G). Colon histology was assessed at 200x along the entire colon length (H-J). *, P<0.05 or as otherwise indicated as determined by Student’s –t-test. Size bar indicates 100 μm.