IL-4-producing ILC2s are required for the differentiation of TH2 cells following Heligmosomoides polygyrus infection

Immunity to many human and murine gastrointestinal helminth parasites requires interleukin-4 (IL-4)-directed type 2 helper (TH2) differentiation of CD4+ T cells to elicit type-2 immunity. Despite a good understanding of the inflammatory cascade elicited following helminth infection, the initial source of IL-4 is unclear. Previous studies using the rat helminth parasite Nippostronglyus brasiliensis, identified an important role for basophil-derived IL-4 for TH2 differentiation. However, basophils are redundant for TH2 differentiation following infection with the natural helminth parasite of mice Heligmosomoides polygyrus, indicating that other sources of IL-4 are required. In this study using H. polygyrus, which is controlled by IL-4-dependent immunity, we identified that group-2 innate lymphoid cells (ILC2s) produced significant amounts of IL-4 and IL-2 following H. polygyrus infection. Leukotriene D4 was sufficient to stimulate IL-4 secretion by ILC2s, and the supernatant from activated ILC2s could potently drive TH2 differentiation in vitro in an IL-4-dependent manner. Furthermore, specific deletion of IL-4 from ILC2s compromised TH2 differentiation in vivo. Overall, this study highlights a previously unrecognized and important role for ILC2-derived IL-4 for TH2 differentiation in a natural TH2-dependent model of human helminthiasis.


INTRODUCTION
Intestinal helminth infections remain a significant global burden with devastating economic and health impacts, 1 highlighting clear gaps in our knowledge and translation of anti-helminth immunity. Most small animal model systems are not permissive hosts for human helminths, limiting our ability to study natural, coadapted host-pathogen interactions. Instead, natural murine helminths, such as Heligmosomoides polygyrus, which establish chronic infections in inbred mice similar to their human counterparts, provide an appropriate experimental system to study natural host-pathogen interactions.
Immunity to H. polygyrus has been shown to be dependent on CD4 þ T cells and interleukin (IL-4) 2 orchestrating a polarized type-2 immune response, activating and mobilizing a suite of innate immune cells and local tissue responses. Early innate responses can determine the outcome, severity, and persistence of infection; 3 however, our understanding of these early events is incomplete. Specifically, IL-4-dependent differentiation of CD4 þ type 2 helper (T H 2) cells is an essential component of immunity to H. polygyrus; 4 however, the in vivo source of IL-4 remains elusive. A series of studies identified that basophils were an important source of IL-4 for T H 2 differentiation during murine infection with the rat parasite Nippostrongylus brasiliensis. 5 However, given that IL-4 is not required for immunity to N. brasiliensis, 6 this experimental system may not be the most appropriate to identify the early sources of IL-4 during natural immunity. Basophils and eosinophils express and secrete IL-4 during H. polygyrus infection; 7 however, T H 2 differentiation and type-2 immunity was not impaired in basophil-deficient mice, basophil-depleted mice, or mice given anti-IL-5 to deplete eosinophils. 2,8 Thus, although IL-4 is critical for T H 2 differentiation during H. polygyrus infection, neither basophils nor eosinophils are essential sources of IL-4.
Group-2 innate lymphoid cells (ILC2s) have a variety of important functions including the secretion of potent type-2 cytokines IL-13, IL-5, and IL-9, which contribute to antihelminth immunity as well as the pathogenesis of allergy. 9,10 Recent evidence identified ILC2s as antigen-presenting cells able to process and present antigen to CD4 þ T cells and relay signals to the adaptive immune system. 11 ILC2s also contribute to the maintenance of other innate cells in the tissue, 12,13 support tissue remodeling and repair following injury, 14 and maintain metabolic homeostasis. 15 ILC2s have been shown to transcribe 16 and secrete IL-4; 17 however, the functional relevance of ILC2-derived IL-4 has not been tested.
In this study, we report that ILC2s expand early during H. polygyrus infection in both the intestinal lamina propria (LP) and in the draining mesenteric lymph nodes (MLNs) in concert with early T H 2 cell differentiation. ILC2 expansion with IL-2 immune complexes invoked a local type-2 response in the tissue and provided some protection from chronic H. polygyrus in the absence of Rag-dependent cells. Furthermore, LP ILC2s expressed Il4 and Il2 transcripts and secreted large amounts of IL-4 and IL-2 protein. Functionally, ILC2-derived IL-4 was required for the recruitment of ILC2s to the tissue following H. polygyrus infection and to drive optimal CD4 þ T H 2 cell differentiation. These data identify a previously overlooked and important role for ILC2-derived IL-4 for T H 2 cell differentiation following H. polygyrus infection. Harnessing the influence of ILC2s may, therefore, support T-cell-mediated vaccine approaches against intestinal helminth infections.

RESULTS
Development of type-2 responses following H. polygyrus infection correlates with the expansion of ILC2s in the local lymph node and SI ILC2s are recruited to local lymph nodes of mice early after infection with the natural murine helminth, H. polygyrus, 18 with the total number of ILC2s correlating with resistance to H. polygyrus. 19 However, the functional relevance of ILC2s in H. polygyrus infection has not been tested. Wild-type (WT) C57BL/6 mice infected with H. polygyrus had an elevated local type-2 response in the small intestine (SI) from day 5 following infection, with elevated expression of Il33 and type-2 cytokines (Il13, Il5, and Il4) (Figure 1a). Anti-helminth type-2 innate effector pathways were also increased, including macrophage-associated Retnla and goblet cell-derived Retnlb (Figure 1a). Concomitant with this early type-2 response was an increase in the total number of Lin -Thy1.2 þ ILCs and KLRG1 þ ILC2s in the LP (Figure 1b) and MLN (Figure 1c), with increased mucus production in the local tissue ( Figure 1d). KLRG1 þ ILC2s expressed mRNA and protein for Gata3, canonical type-2 cytokines Il5 and Il13 (Figure 1e,f and Supplementary Figure 1A online), high levels of CD25 (Supplementary Figure 1B), and were morphologically similar to purified Il4-GFP þ CD4 þ T H 2 cells (Figure 1g). ILC2s did not express mRNA or protein for Rorct (Figure 1e and Supplementary Figure 1A) or express Il17a or Il22 (Figure 1e). Thus, KLRG1 expression on Lin -Thy1.2 þ cells faithfully marked GATA-3 þ RORgTcells (Supplementary Figure 1A,C) as reported previously, 20 and GATA-3 þ RORgTcells were significantly expanded in the MLN at day 5 postinfection (Supplementary Figure 1D). IL-4- (Figure 1h) and IL-13-secreting ( Figure 1i) CD4 þ T H 2 cells were also significantly increased as early as day 5 postinfection and further increased at days 7 and 10 postinfection. Taken together, these data highlight a concomitant innate and adaptive type-2 response developing early following H. polygyrus infection.  Figure 2A). The number of Il4-GFP þ ILC2s were significantly expanded following H. polygyrus infection (Figure 2b), in accordance with observed increases in KLRG1 þ cells. Il4-GFP expression does not always correlate with IL-4 protein secretion. 7 To determine whether ILC2s were secreting IL-4 in vivo, we infected IL4-GFP/KN2 mice, which express human CD2 on the surface of cells that have translated and secreted IL-4, and analyzed KN2 expression following infection. Both the proportion (Supplementary Figure 2C) and total number (Figure 2c) of Il4-GFP þ KN2 þ cells were increased 5 days following infection. Furthermore, LP ILC2s secreted large amounts of IL-4 protein following ex vivo stimulation with phorbol 12-myristate 13-acetate (PMA) and ionomycin ( Figure 2d). Although basophils have been described as a dominant Il4-GFP þ population early in the tissue during N. brasiliensis infection, 5 following H. polygyrus infection we observed significantly more Il4-GFP þ ILC2s in the LP compared with Il4-GFP þ basophils (Supplementary Figure 2B) Harvest Lin -Thy1.2 þ KLRG1 þ ) in the LP and MLN of naïve Rag2 À / À and Il2 À / À Rag2 À / À mice. Rag2 À / À or Il2 À / À Rag2 À / À mice were infected with 200 H. polygyrus larvae and harvested at day 5 postinfection. (h) Total number of ILC2s (Lin -Thy1.2 þ KLRG1 þ ) in the MLN of naïve Rag2 À / À (white bars) and infected Rag2 À / À and Il2 À / À Rag2 À / À mice (black bars). Data are representative of three independent experiments with three to five mice per group. ILC2s were expanded using IL-2 complex (IL-2c) and purified from the MLN of Rag2 À / À or Il2 À / À Rag2 À / À mice. Purified ILC2s were stimulated with PMA þ I for 3 h, washed, and then plated in complete media for an additional 21 h. cytokines and lipid mediators known to activate ILC2s. Stimulation with IL-2, IL-33, and IL-25 did not induce IL-4 secretion ( Figure 2e); however, stimulation with leukotriene D4 (LTD 4 ) induced significant IL-4 secretion (Figure 2e), as reported previously, 17 but did not stimulate IL-5 and IL-13 secretion (Supplementary Figure 3A). Previous reports identified that ILC2s also secrete or express IL-2; 12,20 however, it remains unclear whether ILC2-derived IL-2 is required for ILC2 function. We FACS-purified H. polygyrus-elicited KLRG1 þ ILC2s and confirmed that ILC2s transcribed Il2 and secreted large amounts of IL-2 protein following PMA and ionomycin stimulation (Figure 2f). Stimulation with IL-25/IL-33 or LTD 4 did not induce IL-2 secretion from ILC2s (Supplementary Figure 3A), suggesting that other, as yet unknown, factors may stimulate IL-2 secretion from ILC2s in vivo. To test the requirements of IL-2 for ILC2 development and IL-4 production, we infected Il2 À / À mice, which were also backcrossed onto a Rag2 À / À background, with H. polygyrus.

Heligmosomoides polygyrus
In both the LP and the MLN of naïve mice, the frequency and total number of ILC populations were comparable between Il2 À / À Rag2 À / À and Rag2 À / À mice ( Figure 2g). Similarly, following H. polygyrus infection, IL-2 was not required for the expansion of KLRG1 þ ILC2s in the MLN (Figure 2h). Thus, unlike IL-7, which is essential for the development of all ILC populations, 9 IL-2 appears to be redundant for ILC development and ILC2 expansion, 5 days following H. polygyrus infection. Stimulation of FACS-purified ILC2s with IL-2 alone was insufficient for IL-4 secretion (Supplementary Figure 3B); however, to determine whether IL-2 was required for IL-4 secretion, we isolated ILC2s from the H. polygyrus-infected Il2 À / À Rag2 À / À and Rag2 À / À mice and restimulated these cells ex vivo. IL-2 deficiency had no significant impact on IL-4, IL-13, or IL-5 secretion from ILC2s (Figure 2i), suggesting that IL-2 signaling is not fundamentally required for type-2 cytokine production by ILC2s. ILC2s can directly 11,21 and indirectly 22 support T H 2 cell differentiation. Following the observation that H. polygyruselicited ILC2s secreted both IL-4 ( Figure 2d) and IL-2 ( Figure 2e), key cytokines for T H 2 cell differentiation, 23,24 we hypothesized that ILC2s may be important sources of these cytokines for T H 2 cell differentiation. We first tested whether the supernatant from H. polygyrus-elicited ILC2s could support T H 2 differentiation in vitro by culturing naïve CD4 þ CD25 -CD44 þ Il4-GFP -T cells with ILC2-derived supernatant (model; Figure 2j). Strikingly, the supernatant from ILC2s, in combination with T-cell receptor (TCR) ligation, potently differentiated naïve T cells into Il4-GFP þ T H 2 cells in an IL-4-dependent manner (Figure 2j). Of note, when we cultured naïve Il4-GFP -T cells with the supernatant from Il2 À / À ILC2s, we observed a slightly lower frequency of Il4-GFP þ T H 2 cells (Figure 2j). Taken together, these data indicate that ILC2-derived IL-4 can drive T H 2 differentiation in a contact-independent manner and that IL-2-competent ILC2s are required for optimal ILC2-mediated T H 2 differentiation in vitro. Furthermore, given the nonessential role for basophils in early T H 2 responses to H. polygyrus 8 and the minimal increase in basophils at day 5, these in vitro data suggest that ILC2s may also be an important source of IL-4 for T H 2 differentiation in vivo. 25 IL-2-expanded ILC2s provide partial immunity to H. polygyrus IL-4 is necessary and sufficient to expel H. polygyrus 26,27 and H. polygyrus-elicited ILC2s produce IL-4 ( Figure 2). Following published observations that IL-2:anti-IL-2 cytokine complex (IL-2c) treatment can significantly expand ILC2s in vivo, 28,29 we tested whether IL-2c-expanded ILC2s could orchestrate the expulsion of H. polygyrus in the absence of CD4 þ T cells. IL-2c treatment expanded the total number of ILC2s in the LP and MLN at day 5 postinfection (Figure 3a,b). Furthermore, IL-2c treatment increased ILC2 proliferation, determined by Ki67 expression (Figure 3c), and further enhanced CD25 expression on ILC2s (Figure 3d). Expansion of ILC2s correlated with the increased expression of type-2 cytokine genes (Il5, Il13, and Il4) in the SI as well as in increased expression of Retnla, Retnlb, and Arg1 (Figure 3e), indicating that downstream type-2 anti-helminth immune-driven pathways were activated. Consequently, we observed more worms trapped in the wall at day 9 that failed to emerge and develop into mature adults, resulting in a significant reduction of adult worms in the lumen at day 22 (Figure 3f), suggesting that IL-2c-expanded ILC2s provided partial protection from H. polygyrus infection. However, many worms did emerge and persisted in IL-2ctreated mice, suggesting that despite a fivefold increase in ILC2s (Figure 3b) and the significant activation of type-2 immune pathways, ILC2s alone provided significant, but limited protection from H. polygyrus infection.
ILC2s are required for T H 2 differentiation in vivo, with ILC2-derived IL-4 contributing significantly to ILC2 expansion and T H 2 differentiation following H. polygyrus infection The observation that IL-2c-expanded ILC2s could provide significant, but limited, immunity to a primary infection with H. polygyrus in the absence of CD4 þ T cells (Figure 3) was in stark contrast to the ability of ILC2s to drive complete expulsion of N. brasiliensis following similar IL-2c treatment. 11 We therefore explored whether ILC2s could instead be supporting the development of adaptive immune responses in vivo, and in particular whether ILC2-derived IL-4 contributed to T H 2 differentiation in vivo following H. polygyrus infection. To test the requirement of ILC2s and ILC2-derived IL-4, we reconstituted ILC-deficient mice (Il7r À / À ) 30 with either ILC2-deficient BM (Rora sg/sg ; 31 ) alone or in a 80:20 ratio with either Il4 À / À or WT BM (model; Figure 4a). These chimeric mice allowed us to test whether ILC2s or IL-4 production by ILC2s contributed to the differentiation of T H 2 cells in vivo. Following H. polygyrus infection, the proportion of ILCs in the LP did not differ significantly between chimeric mice; however, as expected, Il7r À / À mice reconstituted with Rora sg/sg BM alone (group ''-'') were devoid of KLRG1 þ ILC2s in the LP and MLN, but had normal frequencies of other ILCs (Figure 4b,c and Supplementary Figure 4F). Although IL-4 is not required for ILC2 development, 32 IL-4 deficiency in ILC2s also impaired their recruitment to the SI and MLN (Figure 4b,c and Supplementary Figure 4A). Total splenic CD4 þ T cells were comparable between chimeric mice ( Figure 4d); however, chimeric mice devoid of ILC2s had lower frequencies and lower total number of IL-4-, IL-13-, and IL-5-secreting CD4 þ CD44 þ T cells in the spleen and MLN following H. polygyrus infection (Figure 4e,f and Supplementary Figures 4G,H). Furthermore, mice reconstituted with Il4 À / À BM also displayed fewer IL-4-, IL-13-, and IL-5-secreting CD4 þ CD44 þ T cells in the spleen and MLN (Figure 4e,f and Supplementary Figures 4B,C). Whether the reduction of T H 2 cells in chimeric mice with Il4 À / À ILC2s was because of fewer ILC2 numbers (Figure 4b,c) is currently unclear. Nevertheless, these in vitro and in vivo observations clearly identify a significant role for ILC2s in promoting T H 2 cell differentiation following H. polygyrus infection.

DISCUSSION
The differentiation of T H 2 cells and subsequent activation of B cells secreting high-affinity immunoglobulin E and immunoglobulin G1 antibodies has been the cornerstone of

Il5
Rel. to PBS Il13 Rel. to PBS

Il4
Rel. to PBS

Rel. to PBS Arg1
Rel. to PBS P B S anti-helminth vaccine efforts. However, the precise mechanisms involved in T H 2 cell differentiation in vivo remain poorly understood. Although a clear role for IL-4R signaling in T cells has been well established in vitro and in vivo, 33 the source of IL-4 in vivo is still unclear. 25 The ability of ILC2s to secrete type-2 cytokines 9,10 and present antigen to T cells 11 has established an important role for ILC2s in initiating type-2 immunity and supporting adaptive immunity.
The majority of our current understanding of the role of ILC2s in anti-helminth immunity is derived from murine infections with the rat helminth, N. brasiliensis. Although this has been a useful model to study ILC2 biology, whether murine infections with the mouse-adapted N. brasiliensis reveal all of the characteristics of ILC2s is unclear. Furthermore, immunity to N. brasiliensis does not require IL-4 (ref. 6) and therefore this model system may not be the most appropriate to identify critical sources of IL-4 for anti-helminth immunity. In this study, using the natural murine parasite, H. polygyrus, we identified that ILC2s secreted significant amounts of IL-4 in vivo and ex vivo, which was required to stimulate T H 2 differentiation in vitro and in vivo. Restricting Il4 deficiency to ILC2s in vivo revealed an important role of ILC2s for the differentiation of IL-13 þ and IL-5 þ T H 2 cells following infection with H. polygyrus.   Eosinophils and basophils have also been reported to secrete IL-4 following H. polygyrus infection. 7 Although both of these sources of IL-4 are redundant for T H 2 differentiation during H. polygyrus infection, 2,8 the interplay between these three innate cell populations (basophils, ILCs, and eosinophils) has recently been reported. Specifically, basophil-derived IL-4 was required to activate ILC2s 34 and ILC2-derived IL-5 was sufficient to support eosinophilia. 12 Thus, these three innate cell populations can establish an early type-2 inflammatory landscape, activating innate and adaptive cells and reorganizing local tissue. Whether such interactions are required for optimal IL-4 secretion from ILC2s during H. polygyrus is unclear.
In this study, we identified that LTD 4 , but not IL-2, IL-25, or IL-33, stimulated IL-4 secretion by ILC2s. Instead, IL-2, IL-25, and IL-33 stimulated IL-5 and IL-13 secretion, as reported by many others. 9,10 These results suggest that different activating signals stimulate differential secretion of cytokines from ILC2s. In particular, IL-4 secretion appears to be uncoupled from IL-5 and IL-13 secretion in ILC2s. Mast cells, which are highly responsive to tissue damage-associated IL-33 35 and form close associations with ILC2s in vivo, 29 are an important source of LTD 4 . 36 Following H. polygyrus infection, early activation of mast cells 37 was required for T H 2 differentiation, 37 although the precise mechanisms were unclear. From these reports and our results here, we propose a model where mast cell-derived LTD 4 , following the secretion of tissue damage-associated IL-33, stimulates IL-4 secretion by ILC2s, supporting early T H 2 differentiation. Indeed, using Il4-GFP/KN2 reporter mice, we confirmed that ILC2s secreted IL-4 in the MLN and LP, early during H. polygyrus infection. Of note, a recent report identified that Kit-deficient mice (Kit W /Kit W-v ), which are devoid of mature mast cells, were also deficient in ILC2s. 38 If Kit-deficient mice are indeed deficient in ILC2s, then studies investigating the role of mast cells using these mice may need to be readdressed. In addition to LTD 4 , we observed that Il2 À / À ILC2s secreted slightly less IL-4 and were not as efficient at driving T H 2 differentiation in vitro. These subtle defects may be because of the absence of an IL-2 autocrine feedback loop supporting IL-4 production, given that ILC2s express high levels of the high-affinity IL-2Ra (CD25) 9 and are highly responsive to IL-2 in vitro 9 and in vivo. 39 A similar IL-2 autocrine loop supporting IL-4 production has been reported in T cells following TCR engagement. 23 Alternatively, ILC2derived IL-2 may facilitate T H 2 differentiation by providing IL-2 directly to T cells, as IL-2 signaling in T cells is required for optimal T H 2 differentiation. 40 Finally, IL-4 can also activate ILC2s, 34 providing an additional feedback loop, with ILC2derived IL-4 facilitating local ILC2 activation.
During H. polygyrus infection, we observed a significant increase in IL-4 þ ILC2s in the tissue (Figure 2). In addition to the requirement of ILC2-derived IL-4 for T H 2 differentiation, whether ILC2-derived IL-4 activates other innate, adaptive, or stromal cells is unclear. Indeed, IL-4R-mediated activation of tissue macrophages 41 and epithelial cells 42 is required for control of H. polygyrus. It has also been reported that IL-4R signaling in dendritic cells is required for optimal T-cell differentiation in vivo. 43 Furthermore, ILC2s can activate macrophages in vitro 44 and support epithelial cell repair in the airways. 14 We found that IL-2c-mediated expansion of ILC2s in the absence of adaptive immunity led to a significant increase in epithelial cell-associated Retnlb 27 and macrophage-associated Arg1 expression (Figure 3), suggesting that ILC2-derived IL-4 or IL-13 may have a role in activating these anti-helminth pathways. It has previously been reported that IL-1b may impede ILC2 expansion in WT mice by regulating IL-25 and IL-33 secretion following H. polygyrus infection. 3 In this study, we report that expansion of ILC2s following IL-2c treatment correlated with a reduction in worm establishment, indicating that ILC2s can support immunity following H. polygyrus infection. Indeed, ILC2 frequency was shown to correlate with increased type-2 immunity and resistance to H. polygyrus in genetically resistant mice. 19 In conclusion, using a natural murine helminth infection system to model chronic human helminth infection, we have identified an important and previously unappreciated role for ILC2-derived IL-4 for T H 2 cell differentiation in vitro and in vivo. It was recently reported that helminth-infected children had reduced ILC2s, 45 and given that human ILC2s also secrete IL-4, 46 strategies aimed at enhancing ILC2 function may increase the efficacy of anti-helminth vaccine approaches by boosting T H 2 cell responses and type-2 immunity.

METHODS
Animals and generation of BM chimeric mice. All mice used in this study were maintained under specific pathogen-free conditions at the Mill Hill Laboratory, The Francis Crick Institute (London, UK). C57BL/6, Rag2 À / À , Il2 À / À Rag2 À / À , and TRE-IL-7R.B IL-7RecKO rtTA.C (Il7r À / À ) mice were bred and maintained at The Francis Crick Institute. Il4-GFPFoxp3-RFP mice were generated by crossing 4get 7 and FIR 47 mice at The Francis Crick Institute. C57BL/6 Il4 À / À BM was kindly provided by Judi E Allen (University of Edinburgh, Edinburgh, UK). Rora sg/sg mice were obtained from Jackson Labs (Farmington, CT) and KN2 (ref. 7) mice were provided by Andrew MacDonald (University of Manchester, Manchester, UK) and crossed with 4get 7 mice at The Francis Crick Institute. BM cells were isolated either by flushing adult bones or gently crushing bones from neonatal mice using a pestle and mortar. BM cells were filtered through a 40 mM filter, mixed at the required ratios as indicated, and diluted in sterile phosphate-buffered saline (PBS) for intravenous delivery. A total of 3-5 Â 10 6 cells were transferred per mouse and left to reconstitute for 7-8 weeks before the start of the experiment. Animal experiments were performed according to institutional guidelines and following UK Home Office regulations (project license 80/2506) and were approved by The Francis Crick Institute Ethical Review Panel.
H. polygyrus infection and in vivo treatments. Mice were infected with 200 H. polygyrus L3 larvae by oral gavage. Worms were counted in the lumen or wall of the intestine at days 9 and 22 following infection. For the expansion of ILCs, Rag2 À / À were treated with IL-2 complex (IL-2c) formed of recombinant IL-2 (R&D, Abingdon, UK) and anti-IL-2 antibody (clone JES6-1A12; BioXcell, West Lebanon, NH). They were prepared at a 1:10 ratio of IL-2:anti-IL-2 in sterile PBS for intraperitoneal delivery. Two doses were used in this study, IL-2c low (0.5 mg:5 mg) and IL-2c high (2.5 mg:25 mg). Mice were given three intraperitoneal doses of IL-2c on days 0, 2, and 4 (as specified in figure legends).
Cell isolation, RNA extraction, and quantitative real-time polymerase chain reaction. MLN cells were made into a single-cell suspensions by gently mashing through a 40 mM filter (Thermo-Scientific, Loughborough, UK) and prepared for FACS analysis or sorting. For the isolation of LP cells, Peyer's patches and adipose tissues were removed from the SI, dissected longitudinally to remove faecal contents, and cut into B2-in segments. Mucus was scraped off before resuspension in PBS containing 5% fetal bovine serum (FBS) and 25 mM HEPES (Lonza, Slough, UK). The IEL fraction was removed by incubating the intestines in PBS containing 10% FBS, 15 mM HEPES, 5 mM EDTA (Life Technologies, Paisley, UK), and 1 mM dithiothreitol (Sigma, Gillingham, UK) for 25 min at 37 1C. IEL fractions were discarded through a wide mesh into a beaker. The remaining LP was incubated in cIMDM (complete Iscove's modified Dulbecco's medium (Gibco, Loughborough, UK) containing 10% FBS, 1 mM EDTA, 100 U ml À 1 penicillin and 100 mg ml À 1 streptomycin (Gibco), 8 mM L-glutamine (Gibco), and 0.05 mM 2-mercaptoethanol (Gibco)) containing 0.5 mg ml À 1 Liberase TL (Roche, Burgess Hill, UK) and 60 mg ml À 1 DNAse (Roche) for 25-30 min in a 37 1C shaker. Cells were layered onto 40% isotonic Percoll (GE Healthcare, Little Chalfont, UK) to recover leukocytes from digested tissue. Cells were resuspended in cIMDM and prepared for FACS analysis or sorting. For cytospins, 200 000 purified cells were fixed onto a slide and stained with a modified Giemsa stain (Sigma). Cells were identified based on cell morphology and staining characteristics. Slides were scanned (60X/1.4 magnification) using an Olympus IX71 inverted microscope (Olympus, Southend, UK), captured with a Camera QIClick colour CCD camera (QImaging, Surrey, BC, Canada) and processed using Image Pro-Plus software (Media Cybernetics, Marlow, UK). and images were analyzed using Fiji (PMID 22743772) and Photoshop (Adobe, Maidenhead, UK). For quantitative real-time polymerase chain reaction, tissue samples were harvested into RNAlater (Life Technologies). Tissue samples were homogenized in 500 ml Qiazol using a Precellys Homogenizer (Precellys, Stretton, UK). Two hundred microliters of chloroform was added and the samples were shaken and left at room temperature for 10 min. The samples were spun in a tabletop microcentrifuge at a maximum speed for 15 min. Supernatants were added to EtOH and samples were processed using the RNAeasy Kit (Qiagen, Manchester, UK) according to the manufacturer's instructions and eluted in 30-50 ml RNAse/DNAse-free water. Sortpurified cells were harvested and stored in RLT lysis buffer (Qiagen) at À 20 1C. RNA extractions were performed as above. For quantitative real-time polymerase chain reaction, RNA concentrations were quantified using a Nanodrop 1000 (Thermo Scientific). 0.1-1 mg of RNA was reverse transcribed using the Qiagen Quantitect Reverse Transcription Kit (Qiagen) according to the manufacturer's instructions. cDNA was used for real-time PCR analysis using Power SYBR Green Master Mix and analyzed using an Applied Biosystems 7900HT Fast Real-Time PCR System (both Applied Biosystems, Loughborough, UK). Gene expression was normalized to the housekeeping gene hypoxanthine-guanine phosphoribosyl transferase and expressed as fold change relative to day 0 or PBS treated for some experiments (as detailed in the figure legends). Sequences for primers used are listed in Table 1.
Statistical analysis. Data sets were compared by Mann-Whitney test using GraphPad Prism 5 (LaJolla, CA). Differences were considered significant at Pr0.05.
SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/mi