C57BL/6J, C57BL/6 Rag1^-/-, CD11c-Cre, Lyz2-Cre mice, and Tlr7^-/- mice (stock # 008380) (43) were purchased from the Jackson Laboratory (Bar Harbor, ME). The generation of Atg16l1^f/f mice was previously described (44). To generate mice with conditional deletion of Atg16l1 specifically in myeloid or DCs, Atg16l1^f/f mice were bred with transgenic mice expressing the Cre recombinase gene under the control of the Lyz2 or Cd11c promoter, Lyz2-Cre and CD11c-Cre mice, respectively. Female mice were used for both chronic and acute DSS-induced colitis experiments, while male mice were used for T cell transfer experiments. Both female and male mice were utilized in steady-state experiments. Mice were maintained under specific pathogen-free conditions in the Animal Care Facility at Cedars-Sinai Medical Center. Murine norovirus (MNV) was detected in feces of all experimental mice using a protocol described previously (45). All mice used in experiments were handled according to the guidelines and approved protocols of the Cedars-Sinai Medical Center Animal Care and Use Committees (IACUC protocol # 9214).

Acute colitis was induced in female mice with 3% (w/v) DSS drinking water ad libitum for 7 days followed by one day of regular drinking water before being euthanized. Mice were checked daily for development of colitis by monitoring body weight, gross rectal bleeding, and stool consistency. Mice were sacrificed at day 8. Chronic colitis was induced by multi-cycle administration of DSS drinking water (46). Female mice of 8 weeks of age received 2.8% (w/v) DSS drinking water (40,000–50,000 MW) (MP Biomedicals, Irvine, CA) water on days 1-5, 8-12, 15-19, and 22-26. Mice were sacrificed at day 29. The MLN, cecum, colon, and rectum were removed. Tissues were fixed in 10% Formalin buffered phosphate, paraffin-embedded, and cross sections were stained with hematoxylin and eosin (H&E). H & E sections of cecum, colon, and rectum were scored by a trained observer blinded to the genotypes and treatments as described previously for acute DSS (47) or chronic DSS (46).

Female C57BL/6 and Tlr7^-/- mice were used for donors. Male Rag1^-/- mice of the C57BL/6 background were used for recipients. Splenic CD4⁺ T cells were negatively selected using the EasySep Mouse CD4⁺ T Cell Enrichment Kit (STEMCELL Technologies Inc., Vancouver, Canada). Cells were labeled with anti-CD4, anti-CD25, and anti-CD45RB. Using the MoFlow cell sorter (Dako Cytomation, Carpinteria, CA), CD4⁺ CD25^- CD45RB^high cells were purified by gating and sorting 40% of the highest fluorescing CD45RB cells. Each recipient mouse was injected i.p. with 0.5 x 10⁶ cells in sterile PBS. Mice were weighed and observed for signs of colitis over a 6–8-week period. Mice were sacrificed at the indicated time points. Cecum, colon, and rectum were collected, formalin-fixed, paraffin-embedded, and stained with hematoxylin and eosin (H&E). Intestinal inflammation was scored by a semi-quantitative scoring system by a trained pathologist blinded to the experimental conditions (48). Single-cell suspensions of spleens, MLN, or lamina propria mononuclear cells (LPMC) were stained with anti-CD4 antibodies. For intracellular staining, cells were re-stimulated with 50 ng/ml PMA and 500 ng/ml ionomycin in the presence of Monensin for 4 h. Cells were stained anti-CD4 antibodies, fixed, permeabilized with Foxp3 staining buffer set (eBioscience), and intracellular stained with antibodies against IFN-γ ( XMG1.2) and IL-17 (eBio17B7) and analyzed using LSR II analyzer (BD Biosciences). Cells were restimulated with anti-CD3ε/anti-CD28 antibodies for 3 days. Cytokine levels in supernatants were measured by ELISA.

Cytokine concentration in culture supernatants was assayed by ELISA for IFN-γ, TNF-α, IL-17A, IL-22, and IL-12p40 (eBioscience, San Diego, CA).

Total RNA was isolated using RNeasy kits and reverse transcribed into cDNA with Omniscript RT kit (both Qiagen). QPCR was performed using the CFX Opus real-time PCR system (Bio-Rad). SsoAdvanced Universal SYBR^® Green Supermix (Bio-Rad) was used. mRNA expression of Ifnb1 was normalized to the expression of Actb. The relative gene expression was calculated by the 2^-ΔΔ^Ct method.

LPMC were isolated from the large intestine as previously described (46, 49). Mesenteric lymph nodes (MLN) were removed from the peritoneal cavity and mesenteric fat was removed. Single cell suspension was isolated by crushing MLN between two glass slides and passing the cell suspension through a 40 µm cell strainer. We routinely recovered LPMC at a viability of 76 - 91% with similar viabilities between genotypes. Single cell suspensions were stained with antibodies against murine CD11c (N418), CD103 (2E7), MHC-II (M5/114.15.2), TCRβ (H57-597), CD8β (H35-17.2), CD62L (MEL-14), CD69 (H1.2F3), and granzyme B (GB11), PD-1 (RMP1-30), CD4 (RM4-5) (all eBioscience). The following gating strategy was used to analyze LPMC and MLN: immune cells were gated based on forward and side scatter excluding cell aggregates. CD8⁺ T_RM subsets were gated first on MHCII^-CD11c⁺CD103⁺ and then on TCRβ⁺ CD8β⁺. CD8⁺ T_RM were identified as CD11c⁺MHCII^-CD103⁺TCRβ⁺CD8β⁺. For intracellular staining, cells were re-stimulated with 50 ng/ml PMA and 500 ng/ml ionomycin in the presence of Monensin for 4 h. Cells were stained with CD8β⁺ or CD4⁺, fixed, permeabilized with Foxp3 staining buffer set (eBioscience) and intracellular stained with antibodies against Granzyme B, IFN-γ, and IL-17 (eBioscience), and analyzed using a flow cytometer. Samples were acquired by flow cytometry using a LSR II analyzer (BD Biosciences) and analyzed using FlowJo software (TreeStar Inc).

Bone-marrow (BM) cells were isolated by flushing femurs with RPMI-1640 medium supplemented with 10% FBS, 50 µg/ml gentamicin, 0.25 µg/ml amphotericin B, 100 U/ml Penicillin G, 0.1 mg/ml Streptomycin, and 50 µM β-mercaptoethanol (Complete RPMI-1640). BM cells were washed twice and cultured for 8 days with complete RPMI-1640 supplemented with 30 ng/ml murine Flt3L to stimulate differentiation to pDCs. After 8 days, cells were harvested, washed, and enriched for pDCs using EasySep™ Mouse plasmacytoid DC isolation kit (Catalog #19724). Enriched pDCs were stimulated for 24 hours with 100 ng/ml LPS or 5µg/ml Imiquimod.

Mice received a preventive treatment of Imiquimod before and during the DSS administration. Imiquimod (R837, InvivioGen) was resuspend in sterile PBS and 300 µg/mouse were i.p. injected 3 days before (D-3), and day 1, 4, and 7 of DSS treatment. Mice were monitored daily for development of colitis by assessing body weight, gross rectal bleeding, and stool consistency.

Data were analyzed using GraphPad Prism software and presented as means ± SD. Differences between two groups were analyzed using the unpaired, 2-tailed Student t test. Differences were considered significant at p < 0.05.

To assess the role of TLR7 signaling in maintaining intestinal homeostasis, we determined the total number mesenteric lymph nodes (MLN) and intestinal lamina propria mononuclear cells (LPMC) in WT and Tlr7^-/- and observed no significant differences in LPMC but a significant increase of MLN cells in Tlr7^-/- mice ( Supplementary Figure S1A ). The percentage of total LP CD8⁺ cells were increased in Tlr7^-/- mice while LP DCs were not different between genotypes ( Supplementary Figure S1B ). Moreover, the percentage and total numbers of tissue-resident memory CD8⁺ T_RM cells (CD11c⁺MHCII^-CD103⁺TCRβ⁺CD8β⁺) were significantly higher in LPMC of Tlr7^-/- compared to WT mice, but similar in MLN, suggesting that the increase in this sub-population was limited to the LP ( Figures 1A–D ). To better characterize these CD8⁺ T_RM cells, we analyzed the percentage of CD62L⁺, CD69⁺, and Granzyme B⁺ CD8⁺ T_RM cells. We observed similar percentages of LP CD62L⁺ CD8⁺ T cells in Tlr7^-/- mice but significantly higher percentages of CD69⁺ and Granzyme B⁺ CD8⁺ cells in Tlr7^-/- compared to WT mice which correspond to the profile of tissue-resident memory CD8⁺ T cells ( Figures 1E–G ). Our data suggest that Tlr7^-/- mice have an increased development of LP CD8⁺ T_RM cells compared to WT mice under steady-state conditions in the absence of histological inflammation ( Supplementary Figure S1C ).

TLR7 is mainly expressed in pDCs and plays an important role in providing protection against viral infections. Next, we assessed if absence of TLR7 signaling and resulting impaired response to enteric viruses could lead to CD8⁺ T_RM cells exhaustion. LP CD8⁺ T_RM cells from Tlr7^-/- mice had significantly higher percentages of the exhaustion marker PD-1 compared to WT mice ( Figure 1I ).

Based on our observation of an accumulation of LP T_RM cells, we hypothesized that Tlr7^-/- mice are more susceptible to acute colitis. We administrated DSS in the drinking water for 6 days following by 1 day of regular drinking water to WT or Tlr7^-/- mice. We did not observe any significant differences in the percentages of body weight loss between WT and Tlr7^-/- mice after administration of DSS ( Figure 2A ). Total number of cells in MLN were significantly higher in Tlr7^-/- compared to WT mice but were similar in LP ( Figure 2B ). Tlr7^-/- mice had increased inflammation in the mid-colon area and decreased inflammation in the rectum compared to WT mice ( Figures 2C, D ). Similar to our observations during steady-state, the numbers of CD8⁺ T_RM were significantly higher in the LP of Tlr7^-/- mice compared to WT mice but was similar in MLN in both genotypes ( Figures 2E, F ). Although the percentages of LP CD4⁺ T cells were significantly higher in Tlr7^-/- compared to WT mice, proliferating CD4⁺ T cells and absolute numbers of LP CD4⁺ T cells and proliferating CD4⁺ T cells were similar between WT and Tlr7^-/- mice ( Supplementary Figures S2A, B ). Next, we analyzed the secretion of T_H1 and T_H17 cytokines. Tlr7^-/- mice had significantly higher secretion of IFN-γ and TNF-α in LP while IL-17A secretion was similar between Tlr7^-/- and WT mice ( Figure 2G ).

To assess the role of TLR7 signaling in the development of chronic colitis we utilized the chronic DSS colitis model by administering four cycles of DSS drinking water to WT or Tlr7^-/- mice. Tlr7^-/- mice developed significantly more severe weight loss compared to WT mice starting during the first cycle of DSS treatment and continued throughout the entire experiment ( Figure 3A ). At the time of sacrifice Tlr7^-/- mice had lost significantly more weight compared to WT mice ( Figure 3B ). We observed significantly higher histoscores characterized by more severe ulcerations, and cellular infiltration into the cecal and colonic LP in Tlr7^-/- compared to WT mice ( Figure 3C ; Supplementary Figure S3 ). Total cell counts were significantly higher in MLN of Tlr7^-/- compared to WT mice ( Figure 3D ). We did not observe significant differences in total cell counts for LP in Tlr7^-/- compared to WT mice ( Figure 3D ). However, when we analyzed the cellular composition, we observed significant increases in the percentages of CD4⁺ T cells in LP in Tlr7^-/- mice ( Figure 3E ) and increases of IFN-γ⁺, IL-17A⁺, and IFN-γ⁺ IL-17A⁺ double-positive CD4⁺ T cells ( Figure 3F ). We observed significant increases in the secretion of IL-17A, IFN-γ, and IL-22 from LPMC that were re-stimulated with anti-CD3/anti-CD28 ( Figure 3G ). Interestingly, we did not observe any differences in the secretion of these cytokines in MLN cells (data not shown). These data suggest that TLR7-deficiency leads to an exacerbation of localized intestinal T_H1 and T_H17 responses.

To determine if T cell intrinsic TLR7 signaling contributes to the development of chronic colitis, we transferred CD4⁺ CD45RB^hi naive T cells isolated from wild-type (WT) or Tlr7^-/- mice into Rag1^-/- mice and monitored for development of inflammatory disease. Mice that received Tlr7^-/- CD4⁺ T cells displayed a similar degree of weight loss at the time of sacrifice than WT recipients ( Supplementary Figure S4A ). We did not observe any significant differences in the degree of intestinal inflammation in mice receiving WT or Tlr7^-/- T cells ( Supplementary Figures S4B, C ). We observed a significant increase in LP cell numbers in mice receiving Tlr7^-/- T cells while the cell numbers in MLN, and spleens of mice receiving WT or Tlr7^-/- T cells were similar ( Supplementary Figure S4D ). Analysis of cellular composition revealed no significant differences in the percentages of CD4⁺ T cells in spleens but a significant increase of MLN cells in mice receiving Tlr7^-/- T cells ( Supplementary Figure S4E ). However, we did observe any differences in the percentages of IL-17⁺IFN-γ⁺ CD4⁺ T cells in MLN or spleens of mice receiving WT or Tlr7^-/- T cells ( Supplementary Figure S4F ). Next, we analyzed cytokine production upon TCR re-stimulation in LP and MLN. Although we did observe significantly increased numbers of LPMC we did not observe any differences in the secretion of IL-17A or IFN-γ in re-stimulated LPMC or MLN cells between mice receiving WT or Tlr7^-/- T cells ( Supplementary Figures S4G, H ). We also did not observe any significant differences in the percentage of activated CD4⁺ T cells or percentage of CD4⁺ T cells expressing the gut-homing chemokine receptor CCR9 or the chemokine receptor CCR6 that identifies T_H17 cells in LPMC or MLN from mice receiving WT or Tlr7^-/- T cells (data not shown). These data suggest that TLR7 signaling in T cells does not contribute to the development of T cell-driven chronic colitis.

Chronic viral infections are associated with an increase in the numbers of CD8⁺ T_RM cells and a shift towards an exhausted phenotype of CD8⁺ T cells, which has been associated with defective pDC cytokine secretions (29, 50). We hypothesized that TLR7-deficiency leads to increased numbers of CD8⁺ T_RM cells due to defective anti-viral responses of pDC. To test our hypothesis, we used two conditional knock-out mouse strains with specific deletion of Atg16l1 in myeloid cells (Atg16l1^ΔLyz2 ) or DCs (Atg16l1^ΔCd11c ). At steady-state, Atg16l1^ΔCd11c and Atg16l1^ΔLyz2 developed a similar phenotype than Tlr7^-/- mice with increased percentage and total number of CD8⁺ T_RM cells in the LP ( Figure 4A ). CD8⁺ T_RM cells in Atg16l1^ΔCd11c and Atg16l1^ΔLyz2 mice had similar percentages of CD62L⁺ but higher percentages of CD69⁺ and granzyme B⁺ cells compared to WT mice ( Figure 4B ).

In order to determine if direct activation of TLR7 using the agonist Imiquimod could bypass the requirement for ATG16L1 in myeloid cells, we cultured bone marrow-derived pDCs isolated from WT, Tlr7^-/- , Atg16l1^ΔCd11c , and Atg16l1^ΔLyz2 mice and stimulated them with LPS or Imiquimod in vitro. pDCs isolated from all genotypes had increased secretion of IL12p40 after stimulation with LPS, demonstrating their capacity to respond to a TLR4 ligand ( Figure 5A ; Supplementary Figure S5 ). Stimulation with Imiquimod increased secretion of IL12p40 in pDCs isolated from WT, Atg16l1^ΔCd11c , and Atg16l1^ΔLyz2 mice but not from Tlr7^-/- mice as expected ( Figure 5A ; Supplementary Figure S5 ). Similarly, Imiquimod increased secretion of IL12p40 and TNF-α in bone marrow-derived macrophages isolated from WT and Atg16l1^ΔCd11c mice but not from Tlr7^-/- mice as expected ( Figures 5B, C ). These data demonstrates that stimulation with Imiquimod in vitro could bypass the defect in ATG16L1 and activate TLR7 signaling in pDCs and macrophages.

To further investigate the relationship between TLR7 and ATG16L1 signaling, we administrated DSS in the drinking water of WT, Tlr7^-/- , Atg16l1^ΔLyz2 , and Atg16l1^ΔCd11c mice and treated Atg16l1^ΔCd11c mice with Imiquimod for 10 days ( Figure 6A ). Tlr7^-/- , Atg16l1^ΔLyz2 , and Atg16l1^ΔCd11c mice were more susceptible to DSS as apparent by increased disease activity index and histoscores compared to WT mice ( Figures 6B, C ; Supplementary Figure S6A ). Tlr7^-/- and Atg16l1^ΔLyz2 mice had increased total cell numbers in MLN compared to WT mice, but similar total cell numbers in the LP ( Figure 6D ). Percentage and total number of LP CD8⁺ T_RM cells and secretion of IFN-γ and TNF-α were increased in Tlr7^-/- , Atg16l1^ΔLyz2 , and Atg16l1^ΔCd11c compared to WT mice ( Figures 6E, F ). In contrast, colonic Ifnb1 mRNA expression was similar between WT, Tlr7^-/- , and Atg16l1^ΔCd11c but significantly reduced in Atg16l1^ΔLyz2 compared to WT mice ( Supplementary Figure S6B ). Atg16l1^ΔCd11c mice treated with Imiquimod showed attenuated DSS-induced inflammation compared to untreated Atg16l1^ΔCd11c mice with significantly decreased disease activity index, histoscores, and percentages and total numbers of LP CD8⁺ T_RM cells ( Figures 6B, C, E ). Secretion of IFN-γ and TNF-α were decreased but did not reach statistical significance ( Figure 6F ). Our data demonstrate that ATG16L1-deficiency in DCs promotes the development of LP CD8⁺ T_RM cells and increases susceptibility to colitis which can be attenuated by stimulation of the TLR7 signaling pathway.