Epithelially Restricted Interferon Epsilon Protects Against Colitis

Abstract

Background & aims: Type I interferon (T1IFN) signalling is crucial for maintaining intestinal homeostasis. We previously found that the novel T1IFN, IFNε, is highly expressed by epithelial cells of the female reproductive tract, where it protects against pathogens. Its function has not been studied in the intestine. We hypothesize that IFNε is important in maintaining intestinal homeostasis.

Methods: We characterized IFNε expression in mouse and human intestine by immunostaining and studied its function in the dextran sulfate sodium (DSS) colitis model using both genetic knockouts and neutralizing antibody.

Results: We demonstrate that IFNε is expressed in human and mouse intestinal epithelium, and expression is lost in inflammation. Furthermore, we show that IFNε limits intestinal inflammation in mouse models. Regulatory T cell (Treg) frequencies were paradoxically decreased in DSS-treated IFNε-/- mice, suggesting a role for IFNε in maintaining the intestinal Treg compartment. Colitis was ameliorated by transfer of wild-type Tregs into IFNε-/- mice. This demonstrates that IFNε supports intestinal Treg function.

Conclusions: Overall, we have shown IFNε expression in intestinal epithelium and its critical role in gut homeostasis. Given its known role in the female reproductive tract, we now show IFNε has a protective role across multiple mucosal surfaces.

Keywords: Colitis; Interferon Epsilon; Intestinal Inflammation; Type I Interferons.

Graphical abstract

Figure 1

IFNε expression in human and mouse colon. Human colon biopsies from pediatric control and IBD patients were stained for IFNε or isotype control by immunohistochemistry. IFNε appears brown. Graph shows IFNε staining intensity in control and IBD patients (A). Mouse control and inflamed colon (DSS) were stained for IFNε by immunofluorescence. Colons from IFNε-/- mice were used as negative controls. Graphs show IFNε staining intensity in control and DSS-treated colon (B). Staining intensities were determined using Aperio ImageScope software. Data are depicted as mean staining intensity + standard error of the mean and include individual data. ∗P < .05 by Mann-Whitney U test.

Figure 2

Role for IFNε in intestinal inflammation in the DSS colitis model. WT (n = 8) and IFNε-/- mice (n =8) were treated with DSS in drinking water for 7 days and monitored for development of clinical symptoms. Body weight as percentage of starting weight (A). Colon length at day 7 after start of DSS treatment. Dotted line indicates average colon length of untreated mice (B). Disease activity index (C). Representative pictures of colon H&E staining of untreated control colon (top panel) and after DSS treatment (lower panel). Graph displays histopathology after DSS treatment (D). WT mice were treated intraperitoneally with our IFNε-neutralizing antibody (n = 10) or with isotype control (n = 10) and received DSS in drinking water. Body weight as percentage of starting weight (E). Colon length at day 7 after start of DSS treatment. Dotted line indicates average colon length of untreated mice (F). Disease activity index (G). Histopathology after DSS treatment (H). Data are depicted as mean + standard error of the mean and include individual data (B–D). ∗P < .05,∗∗P < .01,∗∗∗P < .001 by Mann-Whitney U test.

Figure 3

Modulation of Treg by IFNε. WT and IFNε-/- mice were treated with DSS in drinking water for 7 days, and colon FoxP3+ Treg were analyzed by flow cytometry and immunohistochemistry. Representative picture of FoxP3 plotted against CD4 in isolated colon immune cells of WT (left) and IFNε-/- mice after DSS treatment. Graphs display summary data (A). Representative picture of FoxP3 immunohistochemistry staining in colon of WT (left) and IFNε-/- mice after DSS treatment. Graphs show numbers of FoxP3+ cells per field of view for untreated control colon (left) and after DSS treatment (right; B). WT Treg were cocultured with CFSE-labelled IFNAR1-/- CD4 responder T cells (C, E–H), or IFNAR1-/- Treg were cocultured with CFSE-labelled WT CD4 responder T cells (D). All were stimulated with CD3/CD28 beads and interleukin 2. Proliferation was characterized by measuring CFSE dilution in IFNAR1-/- responder cells cocultured with WT Treg in a ratio ranging from 1:2 to 1:64 (C) and WT responder cells cocultured with IFNAR1-/- Treg in a ratio ranging from 1:2 to 1:64 (D). Cell surface expression of CD69 (E) and PD-1 (F) in IFNAR1-/- responder cells cocultured with WT Treg in a ratio ranging from 1:2 to 1:64. Intracellular expression of tumor necrosis factor alpha (E) and interleukin 4 (F) in IFNAR1-/- responder cells cocultured with WT Treg in a ratio ranging from 1:2 to 1:64. Data are depicted as mean + standard error of the mean and include individual data (A and B). ∗P < .05, ∗∗P < .01,∗∗∗P < .001 by Mann-Whitney U test. Data are depicted as mean + standard error of the mean of 8 mice per genotype (C–H).

Figure 4

Colon immune cell frequencies in DSS-treated mice. Flow cytometry was performed on colon immune cells after DSS treatment. Gating strategies are shown in Figures 5 and 6. CD4 T cells as percentage of CD45 and expression of PD-1 and CD69 in the CD4 T cell population (A). CD8 T cells as percentage of CD45 (B). Percentage of monocytes (C) and neutrophils (D). Percentage of CD11c+ DC (E). Percentage of CD11b+ and CD11b+CD11c+ MPh and analysis of MHC II expression in these subsets (F). Data are depicted as mean + standard error of the mean of 5 mice per genotype.

Figure 5

Colon CD4 T-cell gating. For all fluorescence activated cell sorter analysis, doublets were excluded, viable cells were selected, and total immune cells were gated using CD45. CD4 cells were selected and analyzed for PD-1 and CD69 expression as indicated.

Figure 6

Colon myeloid cell gating strategy. For all fluorescence activated cell sorter analysis, doublets were excluded, viable cells were selected, and total immune cells were gated using CD45. Monocytes and neutrophils were gated on the basis of Ly6C and Ly6G expression. Ly6C- cells were subdivided into MPh and DC populations as indicated.

Figure 7

Adoptive transfer of Treg to modulate DSS severity of IFNε-/- mice. On day 4 after start of DSS treatment, IFNε-/- mice were intravenously injected with 1 × 10⁶ WT splenic Treg or saline and monitored for development of clinical symptoms. Body weight as percentage of starting weight (A). Colon length at day 7 after start of DSS treatment. Dotted line indicates average colon length of untreated mice (B). Disease activity index (C). Representative pictures of H&E-stained colon of Treg-treated and saline-treated mice; graphs show summary data (D). Data are depicted as mean + standard error of the mean of n = 8 mice per treatment group and include individual data (B–D). ∗P < .05,∗∗P < .01,∗∗∗P < .001 by Mann-Whitney U test.