key: cord-0006189-xfyavk3m
authors: Gil-Cruz, Cristina; Perez-Shibayama, Christian; Onder, Lucas; Chai, Qian; Cupovic, Jovana; Cheng, Hung-Wei; Novkovic, Mario; Lang, Philipp A; Geuking, Markus B; McCoy, Kathy D; Abe, Shinya; Cui, Guangwei; Ikuta, Koichi; Scandella, Elke; Ludewig, Burkhard
title: Fibroblastic reticular cells regulate intestinal inflammation via IL-15-mediated control of group 1 ILCs
date: 2016-10-31
journal: Nat Immunol
DOI: 10.1038/ni.3566
sha: fc5c48ff3c5d19b3ce17e01a9df5727cf1aa3d7e
doc_id: 6189
cord_uid: xfyavk3m

Fibroblastic reticular cells (FRCs) of secondary lymphoid organs form distinct niches for interaction with hematopoietic cells. We found here that production of the cytokine IL-15 by FRCs was essential for the maintenance of group 1 innate lymphoid cells (ILCs) in Peyer's patches and mesenteric lymph nodes. Moreover, FRC-specific ablation of the innate immunological sensing adaptor MyD88 unleashed IL-15 production by FRCs during infection with an enteropathogenic virus, which led to hyperactivation of group 1 ILCs and substantially altered the differentiation of helper T cells. Accelerated clearance of virus by group 1 ILCs precipitated severe intestinal inflammatory disease with commensal dysbiosis, loss of intestinal barrier function and diminished resistance to colonization. In sum, FRCs act as an 'on-demand' immunological 'rheostat' by restraining activation of group 1 ILCs and thereby preventing immunopathological damage in the intestine. SUPPLEMENTARY INFORMATION: The online version of this article (doi:10.1038/ni.3566) contains supplementary material, which is available to authorized users.

The intestine harbors a vast number of commensal microorganisms and is frequently confronted with infectious agents, including pathogenic viruses. Moreover, food antigens that can induce severe allergic reactions enter the body via the intestinal mucosa. Hence, the immunoregulation of tolerance to food allergens and commensal microbes must be initiated and maintained by the gut immune system, while forceful protective immunity to pathogens should be induced following their detection in gut-associated lymphoid tissues 1,2 . Peyer's patches (PPs) are central structures of the gut-associated lymphoid tissues and serve as the main initiation site of intestinal immune responses. Moreover, antigen and immune cells from PPs reach the mesenteric lymph nodes (mLNs) that serve as a second major line of defense against pathogens 3 .

Diverse layers of protective mechanisms, including mucosal immunoglobulin A (IgA) 4 and innate lymphoid cells (ILCs) 5 , contribute to the maintenance of gut homeostasis and responses to intestinal pathogens. Various ILC subsets (group 1 (ILC1), group 2 (ILC2) and group 3 (ILC3)) have distinct functions in mucosal tissues, such as interferon-γ (IFN-γ)-mediated clearance of intracellular pathogens (ILC1) 6 or support of the expulsion of helminths by secretion of T helper type 2 cytokines (ILC2) 7 . A third type of ILCs contribute to the elimination of Gram-negative bacteria through the production of IL-22 (ILC3) 8 . In addition, group 3 ILCs direct the development of PPs through interaction with fibroblastic stromal cells in the embryo 9 .

It is not clear however whether group 3 ILCs or other ILC subsets maintain their cross-talk with fibroblastic stromal cells in adult PPs.

The activation of immune responses in PPs and other secondary lymphoid organs (SLOs) depends on the establishment of specialized niches for hematopoietic cells that facilitate optimized antigen acquisition, activation of T cells in the T cell zone and induction of antibody responses in the B cell zone 10 . The particular confines of these immune reactions are generated by fibroblastic reticular cells (FRCs) that not only provide structural support and guidance to immune cells but also actively participate in shaping immune responsiveness 11, 12 . For example, FRCs of the T cell zone in lymph nodes (LNs) regulate the migration and survival of T cells by producing homeostatic chemokines such as CCL19 and CCL21, both crucial for the attraction and retention of T cells [13] [14] [15] . Podoplanin (PDPN)-expressing FRCs located at the border between the T cell zone and B cell zone and in the B cell follicles provide B cell-stimulating factors to foster antibody responses 16 and generate the B cell-attracting chemokine CXCL13 during infection 17 . FRCs situated along the LN subcapsular sinus, commonly known as 'marginal reticular cells' , are characterized by expression of the adhesion molecule MAdCAM-1 (ref. 18) . Although the phenotype and function of FRCs in peripheral LNs have been studied in detail 11, 12 , the precise role of FRCs in PPs in immunological function and PP structure is not yet known, nor is it fully clear whether FRC subsets found in LNs are present in PPs.

We found here that genetic ablation of innate immunological sensing dependent on the adaptor MyD88 in CCL19-expressing FRCs precipitated severe intestinal inflammatory disease in the aftermath of infection with an enteropathogenic virus. Delineation of the underlying mechanism revealed that unleashed trans-presentation of IL-15 on MyD88-deficient FRCs during viral infection fostered the activation of cytotoxic and non-cytotoxic group 1 ILCs and substantially accelerated clearance of the virus. However, enhanced vigilance against the viral infection came at the cost of dysregulated T cell responses with a shift from Foxp3 + regulatory T cells toward IFN-γ-producing type 1 helper T cells. We conclude that FRCs in PPs and mLNs actively regulate intestinal homeostasis by controlling local ILC1 activation and subsequent helper T cell differentiation.

MyD88-independent FRC subset specification and homeostasis FRCs generate the cellular scaffold required for the activation of immune cells in LNs 15, 16 , express multiple Toll-like receptors (TLRs) and respond to stimulation by the innate immune system 19 . We therefore reasoned that FRCs in PPs and mLNs might sense mucosal pathogens and affect immunity by directing those immune cells that respond first to the intrusion. To address this issue, we selectively abolished MyD88-dependent innate immunological sensing in FRCs through the use of mice in which loxP-flanked alleles encoding MyD88 undergo FRC-specific deletion by a transgene expressing Cre recombinase under the control of the Ccl19 promoter (Ccl19-Cre), called 'Myd88-conditional-knockout' (Myd88-cKO) mice here. FRCs in PPs can be highlighted in vivo through expression of enhanced yellow fluorescent protein (EYFP) from the ubiquitous Rosa26 locus (R26R) in Ccl19-CreR26R-EYFP mice (called 'Ccl19 EYFP mice' here) 15 (Fig. 1a) . Ablation of MyD88 in FRCs did not affect PP formation (Supplementary Fig. 1a) , nor did it alter the size or structural organization of PPs (Fig. 1b) or their immune-cell content (Supplementary Fig. 1b-e) . Likewise, the structural organization of mLNs (data not shown) and their immune-cell composition (Supplementary Fig. 1b-e) were not affected by this FRC-specific ablation of MyD88. High-resolution in situ analysis by confocal laserscanning microscopy revealed that the PP FRCs of both Ccl19 EYFP mice and Ccl19 EYFP Myd88-cKO mice formed the typical network structure and expressed the FRC marker PDPN (Fig. 1a,b) , CCL21, smooth muscle actin-α and the intercellular adhesion molecule ICAM1 (Supplementary Fig. 1f) . Moreover, we found that the lack of innate immunological sensing in Myd88-cKO mice did not affect expression of the Ccl19-Cre transgene in CD31 − PDPN + stromal cells of PPs or mLNs (Fig. 1c,d) . The appearance of FRC subsets such as CD35 + follicular dendritic cells (Fig. 1e ) or CD157 + MAdCAM-1 + marginal reticular cells was independent of MyD88 in PPs and mLNs ( Fig. 1f and Supplementary Fig. 1g) . Likewise, the expression of other canonical FRC markers was not altered by the absence of MyD88 in EYFP + cells of PPs or mLNs (Fig. 1g) MyD88 signaling in FRCs controls antiviral ILC1 responses To assess whether an invasive enteric pathogen would substantially alter the activity of FRCs, we infected MyD88-sufficient mice and mice with FRC-specific MyD88 deficiency with mouse hepatitis virus (MHV). This cytopathic coronavirus is recognized via the TLR7-MyD88 pathway 20 , 'preferentially' targets macrophages in SLOs 21 and causes severe inflammatory disease in the intestine following uptake via the oral route 22 . Here, we used a dose of 5 × 10 4 infectious particles, which led to substantial viral replication on days 3 and 6 after infection in the PPs and mLNs of MyD88-sufficient mice (Fig.  2a,b) but spared other regions of their intestine (data not shown). Viral titers were significantly lower in Myd88-cKO mice than in Myd88sufficient mice on day 3 after infection, and infectious particles were almost completely eliminated from the Myd88-cKO mice on day 6 ( Fig. 2a,b) . This finding was unexpected, because global deficiency in MyD88 or TLR7 resulted in uncontrolled viral spread, with viral titers more than 10,000-fold higher in mice with such global deficiency than in Myd88-cKO mice (Fig. 2a,b) . Moreover, the virus was purged from the spleen and liver of Myd88-cKO mice at day 6 ( Supplementary  Fig. 2a) , which indicated that potent immunological effector mechanisms had prevented systemic spread of the pathogen.

The accelerated viral control in Myd88-cKO mice as early as on day 3 after infection suggested that innate antiviral immune cells had been activated by the FRC-specific MyD88 deficiency. Indeed, we found significantly more cells expressing the activating natural killer cell (NK cell) receptors NK1.1 and NKp46 in the PPs and mLNs of Myd88-cKO mice than in those of Myd88-sufficient early after infection (Fig. 2c,d) . Moreover, production of the antiviral effector cytokine IFN-γ was much greater in the NK1.1 + cells of Myd88-cKO mice than in their counterparts from Myd88-sufficient mice (Fig. 2e) . Histological analysis of PPs from infected mice at day 3 after infection with MHV revealed that NK1.1 + cells were located mainly in the interfollicular regions (Fig. 2f) . Moreover, these cells were in close contact with FRCs expressing the Ccl19-Cre transgene, both in the presence of MyD88 (Fig. 2g) and in its absence (Fig. 2h) in FRCs. The frequency and number of both non-cytotoxic group 1 ILCs expressing the cytokine receptor IL-7Rα and IL-7Rα-negative NK cells were increased when immunological sensing was altered in the FRCs of PPs ( Fig. 2i,j and Supplementary Fig. 2b ,c) and mLNs ( Supplementary  Fig. 2b,c) . Control staining revealed that NK1.1 + IL-7Rα − group 1 ILCs, but not IL-7Rα + group 1 ILCs, expressed the signature transcription factor Eomes (Supplementary Fig. 2d) . FRC-specific MyD88 deficiency had a positive effect on the population expansion and activation of group 1 ILCs, whereas the size of the ILC2 and ILC3 compartments in PPs (Fig. 2i,j) and mLNs (Supplementary Fig. 2b,c) was slightly reduced relative to their size in Myd88-sufficient mice. To assess whether this enhanced antiviral immunity was exclusively dependent on group 1 ILCs, we ablated NK1.1-expressing cells using a well-established depletion protocol 23 (Supplementary Fig. 2e,f) . Antibody-mediated depletion of group 1 ILCs completely restored viral replication in all organs of Myd88-cKO mice but had little effect on viral replication in Myd88-sufficient mice (Fig. 2k) . These data indicated that the FRCs in PPs and mLNs were able to function as potent regulators of antiviral immunity and that this function was exerted via control of group 1 ILCs.

The differentiation and activation of cytotoxic and non-cytotoxic ILCs is stringently regulated by cytokines 24 . In addition, NK cell activity is controlled by inhibitory receptors that bind to major histocompatibility complex class I molecules 25 . Since MyD88-deficient A r t i c l e s FRCs in PPs and mLNs did not show altered expression of major histocompatibility complex class I during intestinal infection with MHV (data not shown), we established an in vitro cell-culture system (Fig. 3a) to probe FRC cytokine responses following exposure to R848, a synthetic agonist of TLR7 and TLR8. We found that MyD88sufficient FRCs responded to stimulation of TLR7 with considerable production of the inflammatory mediators IL-6 and CCL2, whereas FRCs lacking MyD88 failed to respond to stimulation with R848 (Fig. 3b) . Exposure to this TLR7 ligand led to a substantial reduction in the production of IL-15 by MyD88-sufficient FRCs, whereas MyD88-deficient FRCs continued to produce large amounts of this NK cell-and ILC1-activating cytokine (Fig. 3b) . Since IL-15 acts mainly in a cell-contact-dependent manner via trans-presentation by IL-15 receptor α-chain (IL-15Rα) 26 , we used expression of IL-15Rα as an additional marker of differential FRC activation. Expression of IL-15Rα in MyD88-sufficient FRCs was significantly reduced after stimulation with the TLR7 ligands R848 or singlestranded RNA or with IL-1β ( Fig. 3c and Supplementary Fig. 3) .

MyD88-dependent regulation of the production of IL-15 in vivo was confirmed by RT-PCR analysis showing significantly higher expression of Il15 mRNA in CD45 − PP stromal cells from MyD88-cKO mice than in those from Myd88-sufficient mice (Fig. 3d) and in EYFP + cells sorted from the PPs of Ccl19 EYFP Myd88-cKO mice than in those sorted from the PPs of Ccl19 EYFP mice (Fig. 3e) , on day 3 after infection. Moreover, flow cytometry revealed that IL-15 was detectable on the surface of EYFP + cells expressing the cell-adhesion molecule CD44 (Fig. 3f) . The lack of MyD88 substantially increased not only the amount of surface-bound IL-15 but also expression of the IL-15-trans-presenting molecule IL-15Rα (Fig. 3f) . Back-gating showed that the enhanced IL-15 production under conditions of MyD88 deficiency could be attributed to PDPN dim CD44 hi FRCs (Fig. 3g) , which suggested that a distinct FRC subpopulation activated group 1 ILCs via trans-presentation of IL-15.

To assess whether IL-15 is the dominant FRC-derived factor that mediates the population expansion of group 1 ILCs and controls viral replication, we applied a neutralizing antibody to IL-15 (anti-IL-15) during the course of the infection starting 1 d before infection with MHV. We found that treatment with anti-IL-15 efficiently blunted the exaggerated population expansion of group 1 ILCs in Myd88-cKO PPs (Fig. 4a,b) . Notably, neutralization of IL-15 led to a reduced frequency (Fig. 4a) and absolute number (Fig. 4b) of group 1 ILCs in Myd88-sufficient mice and diminished the ILC3 population both in Myd88-sufficient PPs and in Myd88-cKO PPs (Fig. 4c) . These data confirmed the importance of IL-15 for ILC homeostasis in PPs 24 and indicated that innate immunological signaling in FRCs functioned as an important regulatory switch for this IL-15-driven ILC1 proliferation. Assessment of the expression of CD122, a component of the IL-15 receptor, on Eomes1 + and Eomes1 − ILC1 subsets revealed no major change in expression of the IL-15 receptor β-chain (CD122) under conditions of FRC-specific MyD88 deficiency (Fig. 4d) , indicative of an ILC1extrinsic regulation circuit. Notably, in vivo neutralization of IL-15 disinhibited viral replication in Myd88-cKO mice, while Myd88-sufficient mice showed only minor changes in viral load (Fig. 4e) . Together these data revealed that the ILC1 activity in PPs and mLNs was controlled almost exclusively via a single cytokine derived from a particular FRC subset. Moreover, it appeared that innate immunological activation of FRCs via direct recognition of viral RNA or through IL-1β-mediated stimulation was critical for the adjustment of ILC1 reactivity. A r t i c l e s FRCs regulate homeostatic ILC1 and NK cell maintenance To assess whether FRC-derived IL-15 not only controls the proliferation and activity of group 1 ILCs and/or NK cells during an infection but also contributes to their homeostasis, we selectively abolished IL-15 expression in FRCs by generating Ccl19-CreIl15 fl/fl mice (called 'Il15-cKO' mice here). We found that FRC-specific ablation of IL-15 affected neither PP formation (Supplementary Fig. 4a ) nor the composition of T lymphocytes and B lymphocytes in this SLO (Supplementary Fig. 4b) . NK1.1 + NKp46 + cells were almost completely absent under conditions of FRC-specific loss of IL-15, whereas neither the ILC2 compartment nor the ILC3 compartment was significantly affected (Fig. 5a,b and Supplementary Fig. 4c,d) .

Since the Ccl19-Cre transgene was active in less than 0.08% of nonendothelial bone marrow stromal cells (Supplementary Fig. 4e) , we concluded that the IL-15-expressing FRCs in gut-associated SLOs generated an essential niche for the maintenance of group 1 ILCs and NK cells under homeostatic conditions.

Next we assessed to what extent such specific ablation of Il15 in FRCs affected the activation of ILC1-and/or NK cell-mediated antiviral immunity. Even under inflammatory conditions, we found an almost complete absence of NK1.1 + NKp46 + cells in the PPs (Fig. 5c) and mLNs (Fig. 5d) of Il15-cKO mice. Moreover, other CD45 + Lin − cells (i.e., group 2 and group 3 ILCs) did not produce IFN-γ during the early phase of the viral infection in Il15-cKO mice (Fig. 5c,d) , which indicated that other sources of IL-15, such as dendritic cells or macrophages 27, 28 , failed to compensate for the lack of this growth factor in the FRC niche. Consistent with the results of the anti-NK1.1 ablation experiment (Fig. 2k) , we found only a moderate effect of the selective loss of group 1 ILCs and/or NK cells on the control of viral replication (Fig. 5e) . Thus, we concluded that the FRCs built and maintained an exclusive IL-15-dependent niche for the maintenance of group 1 ILCs and/or NK cells in the PPs and mLNs and controlled antiviral immunity in gut-associated SLOs through cessation of IL-15 production.

Early IFN-γ production by NK cells has been shown to be important for the differentiation of helper T cell subsets 29 . Accordingly, the greater abundance of IFN-γ-secreting group 1 ILCs in Myd88-cKO mice was associated with an elevated frequency of antiviral CD4 + T cells that secreted IFN-γ in the PPs and mLNs on day 10 after infection with MHV (Fig. 6a,b and Supplementary Fig. 5a,b) but only moderate proliferation of IL-17-producing helper T cells (Supplementary Fig. 5c) . Notably, MyD88 deficiency in FRCs precipitated a significantly lower abundance of regulatory T cells expressing the transcription factor Foxp3 in the PPs of Myd88-cKO mice A r t i c l e s than that in the PPs of Myd88-sufficient mice (Fig. 6c) , which suggested that immunological regulation in the small intestine might have been compromised. Indeed, the lack of innate immunological sensing in FRCs led to substantial weight loss in MHV-infected Myd88-cKO mice (Fig. 6d) , despite their accelerated viral clearance (Fig. 2a) . As expected, global deficiency in MyD88, which permitted almost unrestricted replication of the cytopathic virus ( Fig. 2b and Supplementary Fig. 5d) , was associated with a worsened clinical appearance (Fig. 6d) . Gross pathological analysis revealed that the intestinal wall of Myd88-cKO mice was considerably inflamed on day 10 after MHV infection, with a significantly shorter colon length than that of Myd88-sufficient mice (Fig. 6e) . Histopathological examination confirmed that the enhanced immunopathology in the small intestine of Myd88-cKO mice (Fig. 6f) included an edematous muscular layer and blunting of villi (Fig. 6g) . Those pathological changes were associated with a greater antibody response to intestinal microbiota such as Escherichia coli than that of Myd88-sufficient mice (Fig. 6h) , which indicated that the integrity of the epithelial barrier was compromised in infected Myd88-cKO mice. Such weakening of the epithelial barriers in MHV-infected Myd88-cKO mice was accom- panied by pronounced changes in the composition of the microbiome ( Fig. 6i and Supplementary Fig. 5e) , whereas the composition of the commensal flora in naive mice was not affected by FRC-specific MyD88 deficiency (Supplementary Fig. 5e ). The finding that Myd88-cKO mice showed significantly lower resistance to colonization by other intestinal pathogens such as Citrobacter rodentium on day 12 after infection with MHV, relative to that of Myd88-sufficient mice (Fig. 6j) , further emphasized the importance of immunoregulatory functions of FRCs and their ability to adjust global immune responsiveness in the intestine.

Genetic association studies have provided evidence that multiple TLRs are involved in balancing the sensing of microbes in chronic inflammatory disease of the intestine 30, 31 . Moreover, experimental studies have revealed that innate signal integration via the shared adaptor MyD88 promotes epithelial integrity 32, 33 . It has been suggested that ILCs secreting cytokines involved in chronic intestinal inflammation, such as IL-17, IL-22 or IFN-γ, serve as critical sentinels by translating innate immunological signals for the activation of adaptive immunity 34 . Our study has identified a MyD88-dependent pathway of ILC regulation that is called into action once the epithelial barrier has been breached by a pathogen. Our results showed that antiviral ILC1 and NK cell responses in PPs and mLNs were efficiently regulated through limiting of the provision of IL-15 by FRCs. Both direct recognition of viral RNA via TLR7 and indirect activation of FRCs via IL-1β led to activation of this potent control mechanism. It appears that this direct and simple regulatory pathway acts proximal to more elaborate immunoregulatory circuits that involve the balancing of helper T cell differentiation and the generation of regulatory T cells.

ILCs are recognized as the main drivers of the diverse immune responses needed to maintain mucosal surface integrity and protection 5 . However, these cells are almost completely unresponsive to microbial ligands 35 , which suggests that ILCs utilize indirect means to regulate their activation and functionality after microbial invasion. The data presented here suggest that FRCs in gut-associated SLOs function as an intermediate cell to regulate ILC1 function specifically via IL-15. The importance of IL-15 production by FRCs was emphasized by the almost complete absence of group 1 ILCs and/or NK cells in mice that lacked Il15 expression in FRCs. Thus, it appears that FRCs in PPs and mLNs form a highly specialized niche for group 1 ILCs and/or NK cells. That interpretation is in line with the finding that IL-15 production by myeloid cells is not an exclusive regulator of the homeostasis of group 1 ILCs and/or NK cells in SLOs 27 . Thus, it appears that IL-15 production by FRCs secures the maintenance of group 1 ILCs and/or NK cells in SLOs and prevents exaggerated innate immune reactions through MyD88-dependent restriction of IL-15 production under conditions in which myeloid cells start to provide IL-15 via its trans-presentation by IL-15Rα 36 to other activated immune cells.

The conceptual framework of intestinal immunoregulation outlined by published studies 1,37 predicts that innate immune signaling in intestinal tissues needs to be equilibrated at levels that preserve both tissue repair and host defense. Indeed, the absence of MyD88 signaling in all cells rendered the host highly susceptible to infection with cytopathic murine coronavirus. This virus exhibits strong tropism for myeloid cells, and the provision of type 1 interferons is needed to prevent cell death 20, 21 . Moreover, type 1 interferons are important stimulators of NK cells 38 and potentiate early antiviral immune responses, for example, through the stimulation of IL-15 production by dendritic cells 39, 40 . Interestingly, IL-15-mediated counter-balancing of this protective mechanism was induced in FRCs immediately after the recognition of viral RNA via TLR7, following exposure to IL-1β and the transmission of these innate immunological signals via MyD88. Our finding that the proliferation and activity of group 1 ILCs and NK cells in PPs and mLNs was controlled almost exclusively via IL-15 is in line with the finding that intraepithelial group 1 ILCs are highly responsive to . It remains to be determined whether the ILC1 subset, and other ILC subsets, outside of PPs and mLNs are subject to regulation by stromal cells. It is possible that mesenchymal stromal cells of the lamina propria, such as myofibroblasts or mural cells, exert immunoregulatory functions comparable to those of FRCs in PPs. Clearly, lamina propria mesenchymal stromal cells not only form the scaffold of the tissue but also provide a plethora of positive and negative regulatory factors that impinge on innate and adaptive immune processes 42 .

In sum, our study has extended the current paradigm of immunoregulation in the intestine and has revealed innate immunological sensing in FRCs as a key mechanism for the maintenance of intestinal homeostasis. It will be important to further delineate the FRC-ILC axis of intestinal immunoregulation to facilitate the targeting of processes that efficiently restrain intestinal immunopathology.

Methods and any associated references are available in the online version of the paper. 

Type 2 innate lymphoid cells control eosinophil homeostasis

Nfil3 is crucial for development of innate lymphoid cells and host protection against intestinal pathogens

Development of secondary lymphoid organs

Form follows function: lymphoid tissue microarchitecture in antimicrobial immune defence

Lymph node fibroblastic reticular cells in health and disease

Fibroblastic reticular cells: organization and regulation of the T lymphocyte life cycle

Fibroblastic reticular cells in lymph nodes regulate the homeostasis of naive T cells

Restoration of lymphoid organ integrity through the interaction of lymphoid tissue-inducer cells with stroma of the T cell zone

Maturation of lymph node fibroblastic reticular cells from myofibroblastic precursors is critical for antiviral immunity

B cell homeostasis and follicle confines are governed by fibroblastic reticular cells

Identification of a new stromal cell type involved in the regulation of inflamed B cell follicles

Marginal reticular cells: a stromal subset directly descended from the lymphoid tissue organizer

Transcriptional profiling of stroma from inflamed and resting lymph nodes defines immunological hallmarks

Control of coronavirus infection through plasmacytoid dendritic-cell-derived type I interferon

Type I IFN-mediated protection of macrophages and dendritic cells secures control of murine coronavirus infection

The organ tropism of mouse hepatitis virus A59 in mice is dependent on dose and route of inoculation

Natural killer cell activation enhances immune pathology and promotes chronic infection by limiting CD8+ T-cell immunity

The biology of innate lymphoid cells

Dissecting natural killer cell activation pathways through analysis of genetic mutations in human and mouse

The biology of interleukin-2 and interleukin-15: implications for cancer therapy and vaccine design

Macrophage-and dendritic-cell-derived interleukin-15 receptor alpha supports homeostasis of distinct CD8 + T cell subsets

Dendritic cells support the in vivo development and maintenance of NK cells via IL-15 trans-presentation

Induced recruitment of NK cells to lymph nodes provides IFN-γ for T H 1 priming

The role of the Toll receptor pathway in susceptibility to inflammatory bowel diseases

Crohn's disease is associated with a toll-like receptor-9 polymorphism

Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis

Innate immune signalling at intestinal mucosal surfaces: a fine line between host protection and destruction

The unusual suspects--innate lymphoid cells as novel therapeutic targets in IBD

Transcription factors controlling innate lymphoid cell fate decisions

IL-15Rα chaperones IL-15 to stable dendritic cell membrane complexes that activate NK cells via trans presentation

An innately dangerous balancing act: intestinal homeostasis, inflammation, and colitis-associated cancer

Initial and innate responses to viral infections--pattern setting in immunity or disease

Dendritic cells prime natural killer cells by trans-presenting interleukin 15

Differential responses of immune cells to type I interferon contribute to host resistance to viral infection

Intraepithelial type 1 innate lymphoid cells are a unique subset of IL-12-and IL-15-responsive IFN-γ-producing cells

Mesenchymal cells of the intestinal lamina propria

ANOVA with Tukey's post-test or two-way ANOVA with Benferroni's post-test. Statistical significance was defined as P < 0

Generalized lacZ expression with the ROSA26 Cre reporter strain

Identification of coronavirus non-structural protein 1 as a major pathogenicity factor-implications for the rational design of live attenuated coronavirus vaccines

Intimin-specific immune responses prevent bacterial colonization by the attaching-effacing pathogen Citrobacter rodentium

The authors declare no competing financial interests.reprints and permissions information is available online at http://www.nature.com/ reprints/index.html. 

Mice. C57BL/6/N (B6) mice were purchased from Charles River Laboratories (Germany). Myd88 −/− and Tlr7 −/− mice were obtained from the Institute for Laboratory Animal Sciences at the University of Zürich. BAC-transgenic C57BL/6N-Tg(Ccl19-Cre)489Biat (Ccl19-Cre) mice have been previously described 15 and were crossed with R26R-EYFP mice 43 . To specifically ablate MyD88 or IL-15 in FRCs, Ccl19-Cre and Ccl19 EYFP mice were crossed with mice with loxP-flanked Myd88 alleles (Jackson Laboratory) or with mice with loxP-flanked Il15 alleles. Constitutive ablation of Il15 was achieved by flanking exon 5 with two loxP sequences by standard germline recombination in embryonic stem cells. All mice were on the C57BL/6N genetic background and were maintained in individually ventilated cages and were used between 8 and 10 weeks of age. As control animals, co-housed and Cre-negative littermate mice have been used in all experiments. Experiments were performed in accordance with federal and cantonal guidelines (Tierschutzgesetz) under permission numbers SG11/05, SG09/14 and SG10/14 following review and approval by the Cantonal Veterinary Office (St. Gallen, Switzerland).

Mice were infected with MHV A59, orally by gavage with 5 × 10 4 plaque-forming units as previously described 21 . Mice were sacrificed and organs were stored at −70 °C until further analysis. MHV titers were determined by standard plaque assay using L929 cells 44 Table 1 ). Cells were acquired with a FACSCanto (BD Biosciences) and analyzed using FlowJo software (Treestar Inc.). Analysis of EOMES, GATA-3, TBET, Foxp3 and RORγt expression was performed using the Foxp3/transcription factor staining buffer set from eBioscience, according the manufacturer's instructions. Analysis of NK1.1 + and CD4 + T cells responses was performed using cytokine production after stimulation with phorbolmyristateacetate (PMA, 50 ng/ml) and ionomycin (500 ng/ml; both purchased from Sigma) in the presence of brefeldin A (5 µg/ml) for 3 h at 37 °C. For peptide-specific cytokine production, cells were restimulated with M133 peptide (TVYVRPIIEDYHTLT; GenScript) in the presence of brefeldin A (5 µg/ml) for 4 h at 37 °C. For intracellular staining, restimulated cells were surface-stained and fixed with cytofixcytoperm (BD Biosciences) for 20 min. Fixed cells were incubated at 4 °C for 40 min with permeabilization buffer (2% FCS, 0.5% Saponin in PBS) containing antibodies to IFN-γ and to IL-17A (Supplementary Table 1 ). Samples were analyzed by flow cytometry using a FACSCanto (Becton Dickinson), data were analyzed using FlowJo software (Tree Star, Inc.).Histology. PPs and mLNs were fixed overnight in freshly prepared 4% paraformaldehyde (Merck) at 4 °C under agitation. Fixed organs were embedded in 4% low melting agarose (Invitrogen) in PBS and sectioned with a vibratome (Leica VT-1200). 20-to 30-µm-thick sections were blocked in PBS containing 10% FCS, anti-Fcγ receptor (Supplementary Table 1 ) and 0.1% Triton X-100 (Sigma). Sections were incubated over night at 4 °C with the following antibodies: anti-PDPN, anti-B220, anti-CD31, anti-CD4, anti-smooth muscle actin-α, anti-EYFP and anti-CCL21 (Supplementary Table 1) . Unconjugated antibodies were detected with the following secondary antibodies: Dylight649conjugated anti-rat-IgG, Alexa488-conjugated anti-rabbit-IgG, Dylight549conjugated anti-syrian hamster-IgG and Dylight549-conjugated Streptavidin (Supplementary Table 1) . Microscopy was performed using a confocal microscope (Zeiss LSM-710) and images were processed with ZEN 2010 software (Carl Zeiss, Inc.) and Imaris (Bitplane).

Intestinal tissues were removed and cleaned and the distal ileum was immediately fixed in neutral buffered 10% formalin solution. Two to four 5-µm-thick sections of the ileum from each mouse were stained with hematoxylin and eosin. Each sample was graded for the following four criteria on a scale of 0-3 (0, none; 1, mild; 2, intermediate; 3 severe): leukocyte infiltration in the lamina propria; iedema; villus distortion; and presence of markers of severe inflammation such as crypt abscesses, submucosal inflammation, and ulcers. Scores for each criterion were added to generate an overall inflammation score for each sample. Sections were scored in a blinded fashion by two independent observers. Table 1 ). Samples were analyzed by flow cytometry using a FACSCanto (BD Biosciences). Data was analyzed using FlowJo software (Tree Star, Inc.).Quantitative real-time PCR. Total cellular RNA was extracted from homogenized tissues and sorted cells using TRIZOL reagent (Invitrogen) following the manufacturer's protocol. cDNA was prepared using cDNA archive kit (Applied Biosystems), and quantitative RT-PCR was performed using the Light Cycler-FastStart DNA Master SYBR Green I kit (Roche Diagnostics) on a LightCycler machine (Roche Diagnostics). mRNA expression was measured using the QuantiTect SYBR Green PCR primers (Qiagen).

For microbial composition analysis, the ileal content of naive or MHV-infected Myd88-cKO and their Cre-negative littermates was used. Microbial composition was assessed by a 16S high-throughput amplicon analysis. The 16S rRNA gene segments spanning the variable V5 and V6 regions were amplified from DNA from ileal content samples, using a multiplex approach with the barcoded forward fusion primer 5′-CCATCTCA TCCCTGCGTGTCTCCGACTCAG BARCODE ATTAGATACCCYGGTAGT CC-3′ in combination with the reverse fusion primer 5′-CCTCTCTATGGG CAGTCGGTGATACGAGCTGACGACARCCATG-3′. The PCR-amplified 16S V5-V6 amplicons were purified and prepared for sequencing on the Ion torrent PGM system according to the manufacturer's instructions (Life Technologies). Samples with over 500 reads were accepted for analysis. Data analysis was performed using the QIIME pipeline version 1.8.0. Operational taxonomic units were picked using UCLUST with a 97% sequence identity threshold, followed by taxonomy assignment using either the latest Greengenes database.Statistical analysis. Statistical analyses were performed with Graphpad Prism 5.0 using an unpaired two-tailed Student's t-test or Mann-Whitney test. Longitudinal comparison between different groups was performed by