Leukocyte-derived IFN-α/β and epithelial IFN-λ constitute a compartmentalized mucosal defense system that restricts enteric virus infections

Abstract

Epithelial cells are a major port of entry for many viruses, but the molecular networks which protect barrier surfaces against viral infections are incompletely understood. Viral infections induce simultaneous production of type I (IFN-α/β) and type III (IFN-λ) interferons. All nucleated cells are believed to respond to IFN-α/β, whereas IFN-λ responses are largely confined to epithelial cells. We observed that intestinal epithelial cells, unlike hematopoietic cells of this organ, express only very low levels of functional IFN-α/β receptors. Accordingly, after oral infection of IFN-α/β receptor-deficient mice, human reovirus type 3 specifically infected cells in the lamina propria but, strikingly, did not productively replicate in gut epithelial cells. By contrast, reovirus replicated almost exclusively in gut epithelial cells of IFN-λ receptor-deficient mice, suggesting that the gut mucosa is equipped with a compartmentalized IFN system in which epithelial cells mainly respond to IFN-λ that they produce after viral infection, whereas other cells of the gut mostly rely on IFN-α/β for antiviral defense. In suckling mice with IFN-λ receptor deficiency, reovirus replicated in the gut epithelium and additionally infected epithelial cells lining the bile ducts, indicating that infants may use IFN-λ for the control of virus infections in various epithelia-rich tissues. Thus, IFN-λ should be regarded as an autonomous virus defense system of the gut mucosa and other epithelial barriers that may have evolved to avoid unnecessarily frequent triggering of the IFN-α/β system which would induce exacerbated inflammation.

Fig 1. Intestinal epithelial cells minimally express IFN-α/β receptor and do not respond to type I IFN.

(A) RT-qPCR analysis of IFN-α/β receptor chains (Ifnar1 and Ifnar2) and IFN-λ receptor chains (Ifnlr1 and Il10r2) in intestinal epithelial cells (IEC) and lamina propria lymphocytes (LPL) isolated from whole intestinal tissue of adult wild-type mice (n = 4–8). (B) IFNAR1 expression analyzed by flow cytometry on IEC or LPL fractions harvested from wild-type or Ifnar1 ^-/- mice. (C) RT-qPCR analysis of two representative ISGs at steady state in IEC and LPL isolated from wild-type, Ifnar1 ^-/- and Ifnlr1 ^-/- mice (n = 3). (D) Adult Ifnar1 ^-/- and Ifnlr1 ^-/- mice were treated twice subcutaneously with 1 μg of mouse IFN-λ2 or human IFN-αB/D, respectively, at 24 h and 12 h prior to sacrifice as indicated. IFN-induced Mx1 in tissue sections from the gastrointestinal tract was visualized by immunofluorescence. IFN-responsive cells contain nuclear Mx1 (dotty structures in green). Epithelial cells express E-cadherin (red). DAPI (blue) stains nuclei. Data is representative for two to three independent experiments. Mean ± SEM. Bar = 100 μm. ns = non-significant, * p<0.05, ** p<0.01, *** p<0.001.

Fig 2. IECs are potent producers of IFN-λ but not type I IFN.

(A) Base line expression of Ifna5, Ifnb and Ifnl2/3 genes in IEC and LPL isolated from adult wild-type mice (n = 3) assessed by RT-qPCR. (B) Adult wild-type (n = 3) mice were injected intraperitoneally with 100 μg of polyI:C and intestinal tissue was harvested at 2 and 6 h post-treatment. Expression of Ifna5, Ifnb and Ifnl2/3 was analyzed by RT-qPCR in IEC and LPL fractions. (C) Steady state IFN-λ gene expression analysis by RT-qPCR in FACS-sorted EpCAM⁺CD45^- and EpCAM^-CD45⁺ cells from the epithelial fraction. Data is representative for two to three independent experiments. Mean ± SEM. * p<0.05, ** p<0.01. Different letters above bars mark significant differences (p<0.05).

Fig 3. IFN-λ restricts reovirus replication in the epithelium and determines virus shedding in feces, whereas type I IFN blocks replication in the lamina propria and inhibits systemic dissemination.

(A-D) Adult wild-type (n = 25), Ifnar1 ^-/- (n = 18) and Ifnlr1 ^-/- (n = 27) mice were inoculated intragastrically with 10⁸ pfu of reovirus T3D. Data pooled from several independent experiments. (A) At day 4 post-infection, reovirus replication in small intestinal tissue was analyzed by virus titration. (B) Reovirus titers in feces samples collected from wild-type and mutant mice at day 4 post-infection. (C, D) Immunostaining for reovirus antigen (green), E-cadherin (red) and DAPI (blue) in (C) small intestinal tissue or (D) Peyer’s patches. Images are representative of three independent experiments. Bar = 100 μm. ns = non-significant, ** p<0.01, *** p<0.001.

Fig 4. Reovirus replicates extensively in epithelial cells of intestine and biliary tract and induces fatal liver disease in suckling mice lacking functional IFN-λ receptors.

Suckling wild-type (n = 7), Ifnar1 ^-/- (n = 8) and Ifnlr1 ^-/- (n = 11) mice were infected orally with 5 x 10⁶ pfu of reovirus T3D. Data from two independent experiments were pooled. (A) Reovirus titers in the small intestine on day 4 post-infection. (B) Immunostaining of small intestinal tissue at day 4 post-infection for reovirus antigen (green), E-cadherin (red) and DAPI (blue). (C) Survival kinetics of wild-type (n = 11), Ifnar1 ^-/- (n = 13) and Ifnlr1 ^-/- (n = 15) mice. (D) Immunostaining of liver tissue harvested on day 4 post-infection. Reovirus antigen (green), cytokeratin (red) and DAPI (blue). Arrows point to intrahepatic bile ducts. (E) Immunostaining of extrahepatic bile ducts for reovirus antigen (green), cytokeratin (red) and DAPI (blue) at day 4 post-infection. (F) Immunostaining for reovirus antigen of an extrahepatic bile duct from a diseased Ifnlr1 ^-/- mouse on day 8 post-infection. Note that the duct is filled with material seemingly originating from virus-infected cells. Bar = 100 μm. ** p<0.01, *** p<0.001.

Fig 5. Timely IFN-λ production by epithelial cells drives rapid clearance of intestinal reovirus infection.

Suckling wild-type, Ifnar1 ^-/- and Ifnlr1 ^-/- mice (n = 3–4) were orally infected with 5 x 10⁶ pfu of reovirus T3D, and tissue was harvested at either day 1 or day 4 post-infection. (A) Kinetics of reovirus replication by titration. (B) Expression of IFN-responsive genes Isg15 and Oasl2 in whole intestinal tissue analyzed by RT-qPCR. (C) Expression of Ifna5, Ifnb and Ifnl2/3 genes in the IEC fraction of wild-type and Ifnlr1 ^-/- mice assessed by RT-qPCR at day 1 post-infection. (D) Immunostaining of small intestinal tissue for reovirus antigen (red), Mx1 (green) and DAPI (blue). Dotted line marks the border of villi. Bar = 50 μm. d.p.i. = days post infection. Data representative for two individual experiments are shown. Mean ± SEM. Different letters above bars mark significant differences (p<0.05). * p<0.05.