|Home | About | Journals | Submit | Contact Us | Français|
CD4pos T helper (Th) 2 cells secrete interleukin (IL)-4, IL-5 and IL-13 and are required for immunity to gastrointestinal helminth infections1. However, Th2 cells also promote chronic inflammation associated with asthma and allergic disorders2. The non-hematopoietic cell-derived cytokines thymic stromal lymphopoietin (TSLP), IL-33 and IL-25 (IL-17E) have been implicated in inducing Th2 cell-dependent inflammation at mucosal sites3-6, but how these cytokines influence innate immune responses remains poorly defined. Here we show that IL-25, a member of the IL-17 cytokine family, promotes the accumulation of a lineage negative (Linneg) multi-potent progenitor (MPP) cell population in the gut-associated lymphoid tissue (GALT) that promotes Th2 cytokine responses. The IL-25-elicited cell population, termed MPPtype2 cells, was defined by expression of Sca-1 and intermediate expression of c-kit (c-kitint) and exhibited multi-potent capacity, giving rise to cells of monocyte/macrophage and granulocyte lineages both in vitro and in vivo. Progeny of MPPtype2 cells were competent antigen presenting cells and adoptive transfer of MPPtype2 cells could promote Th2 cytokine responses and confer protective immunity to helminth infection in normally susceptible Il17e-/- mice. The ability of IL-25 to induce the emergence of an MPPtype2 cell population identifies a link between the IL-17 cytokine family and extramedullary hematopoiesis and suggests a previously unrecognized innate immune pathway that promotes Th2 cytokine responses at mucosal sites.
Epithelial cell-derived TSLP, IL-33 and IL-25 promote the development of mucosal Th2 cytokine responses through the induction of IL-4 in distinct innate immune cell lineages7-12. TSLP elicits MHC class IIpos basophils that promote Th2 cytokine responses, while IL-33 treatment can activate basophils13, eosinophils14 and natural helper cells (NHCs)15. In contrast, IL-25 is thought to promote IL-4 production in either NKT cells or an undefined non-B/non-T (NBNT) c-kitpos cell population7, 14, 16. Based on c-kit expression, IL-25-elicited NBNT cells were suggested to be a mast cell or mast cell precursor population14, 16. Hematopoietic stem cells (HSCs) are known to express c-kit, circulate through peripheral tissues and differentiate in response to microbial signals17, 18. However, whether IL-25 influences the population expansion or differentiation of peripheral HSC and whether these cells can influence CD4pos Th2 cell responses has not been examined.
Following administration of IL-25 to IL-4/eGFP reporter mice19, a 1.8-fold increase in the total cell numbers was observed in the mesenteric lymph nodes (mLN), with equivalent increases in the total numbers of CD4pos and CD8pos T cells, B cells and macrophages (Fig. 1a). In contrast, IL-25-treatment resulted in a 56-fold increase in a NBNT c-kitpos cell population (Fig. 1b), indicating that the NBNT c-kitpos cells exhibited the greatest relative population expansion following IL-25 administration. The IL-25-mediated population expansion of the NBNT c-kitpos cells was associated with increased expression of Il4, Il5 and Il13 mRNA in the large intestine (Supplementary Fig. 1a), elevated levels of serum IgE (Supplementary Fig. 1b) and increased mucin production in the intestine (Supplementary Fig. 1c) 5.
Analysis of the IL-25-elicited cells revealed that in comparison to c-kitpos mast cells, this cell population exhibited intermediate expression of c-kit (c-kitint) (Supplementary Fig. 2a). Delivery of IL-25 elicited increased frequencies of c-kitint cells in W-sash (Wsh) mice (Supplementary Fig. 2b), which lack classical mast cell populations20 and induced equivalent expression of Il13 mRNA and mucin responses in wild type (WT) and Wsh mice (Supplementary Fig. 2c and d), indicating that IL-25 promotes Th2 cytokine responses independently of mast cells.
Compared to control-treated animals (Supplementary Fig. 3a-c), administration of IL-25 increased the frequency of c-kitint cells in all compartments of the GALT examined, including the mLN (Fig. 1c), the Peyer's patches (Fig. 1d) and cecal patch (Fig. 1e). However, IL-25 did not elicit this population in the spleen or bone marrow (data not shown), suggesting that IL-25-responsive cells may be located in the GALT. Further, analysis of IL-25-elicited c-kitint cells in the GALT revealed two distinct cell populations distinguished by expression of IL-4/eGFP (Fig. 1c-e, right panels), indicating that the IL-25-elicited c-kitint cells are a heterogeneous population.
Previous studies reported elevated expression of IL-25 and increased frequencies of a c-kitpos cell population following exposure to the helminth parasite Nippostrongylus brasiliensis16, 21. Compared to uninfected controls (Supplementary Fig. 3d), increased frequencies and absolute numbers of c-kitint cells were observed in the mLN of WT mice following infection with Nippostrongylus (Fig. 1f and g). Mice lacking expression of either Il17rb or Il17ra failed to exhibit IL-25-elicited population expansion of the c-kitint cells (Supplementary Fig. 4a) or the development of IL-13 and mucin responses (Supplementary Fig. 4b and c), indicating that both IL-17RB and IL-17RA are required for the IL-25-mediated induction of this cell population. Furthermore, the total number of c-kitint cells induced following Nippostrongylus infection were reduced following administration of αIL-25 mAb (infected + control IgG, 58981 ± 4975; infected + αIL-25 mAb, 26109 ± 3039).
To test whether IL-25-elicited c-kitint cells influenced the development of antigen-specific or protective Th2 cytokine responses in vivo, CFSE-labeled ovalbumin (OVA)-specific TCR transgenic T cells were transferred alone or in combination with IL-25-elicited c-kitint cells into naïve congenic recipients. As early as 48 hours–post OVA immunization, donor CD4pos T cells began to proliferate (Fig. 2a, left panel) and accumulated at the site of immunization (Fig. 2b). In mice that also received IL-25-elicited c-kitint cells, OVA-specific CD4pos T cell proliferation was augmented (Fig. 2a, right panel) and there was a significant increase in the accumulation of antigen-specific cells (Fig. 2b). Further, mLN cells isolated from recipients of both OVA-specific CD4pos T cell and IL-25-elicited c-kitint cells secreted elevated levels of IL-13 compared to controls (Fig. 2c). Trichuris-infected Il17e-/- mice produced low levels of Th2 cytokines and parasite-specific IgG1 (Fig. 2d and e), impaired mucin responses (Fig. 2f) and were susceptible to infection (Fig. 2g)22. However, adoptive transfer of IL-25-elicited c-kitint cells from WT mice into infected Il17e-/- mice resulted in elevated production of IL-4, IL-5 and IL-13 and parasite-specific IgG1 (Fig. 2d and e), increased mucin responses (Fig. 2f) and host protective immunity (Fig. 2g). Taken together, these data indicate that IL-25-elicited c-kitint cells could promote antigen-specific Th2 cytokine responses and protective immunity to helminth infection.
IL-25-elicited c-kitint-GFPneg or c-kitint-GFPpos cells lacked expression of CD4 (Supplementary Fig. 5a), indicating that they were distinct from the CD4pos NKT cell populations7, and did not express other surface markers associated with CD4pos T cells (Supplementary Fig. 5a). Further, the IL-25-elicited c-kitint populations did not express B cell-, basophil- or eosinophil-associated surface markers (Supplementary Fig. 5b and c). The majority of IL-25-elicited c-kitint-GFPneg cells were T1/ST2neg/lo and IL-7Rαneg, while IL-25-elicited c-kitint-GFPpos cells were T1/ST2neg/lo and heterogeneous for expression of IL-7Rα (Supplementary Fig. 5d), suggesting that both populations are distinct from NHCs15. Consistent with this, both c-kitint populations expressed little or no mRNA encoding Gata3, Junb, Maf, Stat6 and Il1rl1 (Supplementary Fig. 5e). Delivery of IL-25 resulted in increased frequencies of c-kitint cells in the peritoneum and mesentery (Supplementary Fig. 6a and b). However, while IL-25 treatment increased the cellularity in the mesentery, no changes were observed in the frequency of NHCs or in their expression of CD44 or Thy1.2 (Supplementary Fig. 6c). Taken together, these data indicate that IL-25-elicited c-kitint cells are a unique population and are not T- or B-lymphocytes, NKT cells, basophils, eosinophils, mast cells or NHCs.
Hematopoietic stems cells (HSCs) and multi-potent progenitors (MPPs) express c-kit and Sca-1 and are characterized as lineageneg 23, 24. While HSCs are primarily localized in the bone marrow, they can circulate in the periphery25-28 and have been implicated in immunosurveillance17, 18. IL-25-elicited c-kitint-GFPneg and c-kitint-GFPpos populations were Linneg/lo (Supplementary Fig. 7), and the majority of the IL-25-elicited c-kitint-GFPneg and c-kitint-GFPpos cells expressed Sca-1, were CD150neg, and exhibited heterogeneous expression of CD34 (Fig. 3a-c). Therefore the IL-25-elicited cell populations exhibited a surface phenotype consistent with a MPP-like cell. Although administration of IL-25 induced MPP-like cells in the GALT, the frequencies of MPPs, short-term and long-term HSCs in the BM were unchanged following IL-25-treatment (Supplementary Fig. 8a and b).
To assess the capacity of the c-kitint MPP-like cell population to exhibit multi-potent potential, IL-25-elicited c-kitint-GFPneg or c-kitint-GFPpos cells were sorted and cultured in vitro in the presence of SCF and IL-3 (Fig. 3c-f). Un-fractionated bone marrow cells from naïve mice differentiated into a CD11bpos macrophage-like population (Supplementary Fig. 9a, orange gate) and a CD11bneg granulocyte population that could be identified as mast cells or basophils based on expression of c-kit and FcεRI in addition to cell morphology (Supplementary Fig. 9a and b). Sorted IL-25-elicited c-kitint-GFPpos cells gave rise to a CD11bneg c-kitpos FcεRIpos mast cell population (Fig. 3c, red gate), but failed to give rise to CD11bpos progeny. Consistent with this, the progeny of c-kitint-GFPpos cells were morphologically similar to mast cells (Fig. 3d). IL-25-elicited c-kitint-GFPneg input cells also gave rise to a CD11bneg mast cell population (Fig. 3e, red gate). In addition, the c-kitint-GFPneg input cells gave rise to c-kitneg FcεRIpos basophils (Fig. 3e, blue gate) and CD11bpos macrophages (Fig. 3e, orange gate). The multi-potent potential of the c-kitint-GFPneg cell population was confirmed by cell morphology (Fig. 3f). c-kitint-GFPneg cells cultured in the presence of SCF and IL-3 or using OP9 stromal cells29 also gave rise to a CD11bpos M-CSFRpos Ly6Cpos MHC class IIpos cell population (Supplementary Fig. 10a and b), consistent with myeloid potential. IL-25-elicited c-kitint-GFPneg cells were isolated from WT CD45.2 donor mice and adoptively transferred into naïve CD45.1 congenic recipients. Six days following transfer, c-kitint-GFPneg donor cells differentiated into CD11bpos cells as well as CD11bneg c-kitpos FcεRIpos cells (Supplementary Fig. 11). Collectively, these data indicate that the c-kitint-GFPneg cells may represent a previously unrecognized IL-25-responsive MPP-like cell population in the GALT that selectively exhibits multi-potent potential in vitro and in vivo.
IL-25-elicited c-kitint cell populations were sorted (Fig. 4a, left panel) and cultured in vitro in the presence of SCF and IL-3 for 8 days. The majority of the progeny derived from c-kitint-GFPpos cells were IL-4/eGFPpos MHC class IIneg, while the c-kitint-GFPneg-derived progeny contained IL-4/eGFPpos and MHC class IIpos populations (Fig. 4a, right panels). To test whether these cells could influence T cell proliferation and/or differentiation, progeny from sorted c-kitint-GFPneg- or c-kitint-GFPpos cells were pulsed with OVA and co-cultured with CFSE-labeled OVA-specific TCR transgenic CD4pos T cells. In the absence of antigen, T cells exhibited minimal proliferation and Th2 cytokine production (1% CFSEdim, Fig. 4b (shaded histograms) and c). Antigen-pulsed c-kitint-GFPpos-derived progeny cells failed to induce T cell proliferation (Supplementary Fig. 12, black histogram) or production of Th2 cytokines. In contrast, antigen-pulsed progeny derived from c-kitint-GFPneg cells induced MHC class II-dependent T cell proliferation and production of IL-4 and IL-13 (Fig. 4b and c). Inclusion of anti-IL-4Rα mAb did not affect T cell proliferation (Fig. 4b), but resulted in decreased production of IL-4 and IL-13 (Fig. 4c), indicating that both MHC class II and IL-4R signaling are required for the c-kitint-GFPneg-derived progeny to influence Th2 cell differentiation. No IFN-γ was detected in any culture conditions (data not shown). These results indicate that IL-25-elicited c-kitint cells contain a population of progenitors with multi-potent capacity, termed MPPtype2 cells, whose progeny act to promote CD4pos Th2 cell differentiation.
Collectively, these findings indicate that in addition to a c-kitint-GFPpos cell population that differentiates into mast cells, IL-25 elicits a MPP-like cell population that can differentiate into monocyte/macrophage and granulocyte lineages (Supplementary Fig. 13). Coupled with reports that peripheral HSCs express TLRs and can respond to microbial stimulation17, 18, these findings indicate a previously unrecognized pathway in which peripheral MPPtype2 cells can promote type 2 inflammation at mucosal sites and suggest an evolutionarily conserved pathway between the IL-17 cytokine family, extramedullary hematopoiesis and adaptive immunity.
Mice were treated i.p. with PBS or 0.4 μg recombinant IL-25 daily for 4 days or infected with Nippostrongylus brasiliensis. Mesenteric lymph nodes from IL-4/eGFP reporter mice were separated from the mesentery, homogenized and stained with anti-mouse fluorochrome-conjugated monoclonal antibodies against combined lineage markers and c-kit, FcεRIα, T1/ST2, and IL-7Rα. IL-25-elicited c-kitint cells were sorted using a FACSAria and transferred into OVA/IFA-immunized WT mice or Trichuris-infected Il17e-/- mice and Th2 cytokine responses measured. IL-25-elicited c-kitint-GFPpos and c-kitint-GFPneg cell populations were either sorted and transferred into CD45.1 congenic recipients to assess in vivo differentiation or incubated in vitro in the presence of SCF (50 ng mL-1) and IL-3 (10 ng mL-1). Progeny were analyzed by cytospin or flow cytometry for surface expression of lineage-specific markers. Some progeny were co-cultured with CFSE-labeled OVA-specific TCR transgenic CD4pos T cells in the presence of OVA peptide with or without the inclusion of blocking antibodies against MHC class II or IL-4Rα. CFSE dilution in CD4pos T cells was assessed by flow cytometry and cell-free supernatants were analyzed for secretion of IL-4 and IL-13 by sandwich ELISA.
Balb/c, C57BL/6 and Wsh mice were obtained from Jackson Laboratory and IL-4/eGFP reporter mice were obtained from M. Mohrs (Trudeau Institute). Il17e-/- mice were provided by R.A. Kastelein (Schering-Plough Biopharma). Il17ra-/- and Il17rb-/- mice were provided by J.E. Tocker and A.L. Budelsky (Amgen). Animals were bred and housed in specific pathogen-free conditions at the University of Pennsylvania. All experiments were performed under Institutional Animal Care and Use Committee (IACUC) approved protocols and in accordance with the guidelines of the IACUC of the University of Pennsylvania. All mice used were 4-12 weeks of age. Mice were treated intraperitoneally with PBS or recombinant IL-25 (IL-17E) (0.4 μg; R&D Systems) daily for 4 days.
Mesenteric lymph nodes from IL-4/eGFP reporter mice were separated from the mesentery, homogenized by passing through a 70 μm nylon mesh filter and stained with anti-mouse fluorochrome-conjugated monoclonal antibodies against CD3ε, CD4, CD8, TCRβ, TCRγδ, B220, CD19, CD11b, CD11c, MHC class II, Gr-1, NK1.1, Ter119, FcεRIα, c-kit, Sca-1, CD150, CD34, CD62L, CD69, CD127 (IL-7Rα), CD45.2, CD49b and CCR3 (eBioscience and BD Bioscience). T1/ST2 staining was performed using T1/ST2 biotinylated mAb (MD Biosciences) and PE- or eFluor450-conjugated streptavidin (eBioscience). Peritoneal exudate cells (‘peritoneum’) were collected using peritoneal lavage with injection of 10 mL PBS and aspirated using the same syringe. Mesentery was processed as previously described15. Cells were run on a BD FACSCanto II using DiVa software (BD Bioscience) and analyzed with FlowJo software (Version 8.7.1; Tree Star, Inc.). IL-25-elicited c-kitint (GFPpos or GFPneg) cell populations were sorted using a FACSAria (BD Bioscience) and for in vitro differentiation studies were incubated in the presence of SCF (50 ng mL-1; R&D systems) and IL-3 (10 ng mL-1; R&D systems). Following in vitro culture, progeny were assessed for expression of CD11b, MHC class II, CD115 (M-CSFR), Ly6C, FcεRIα, and c-kit (eBioscience) by flow cytometry as described above. Naïve OVA-specific CD4pos T cells were isolated from DO11.10 mice as previously described14. T cells were stained with fluorochrome-conjugated monoclonal antibodies against CD4, CD62L, and CD44 (eBioscience), re-suspended in 2% FBS in HBSS with 2 mM EDTA (Gibco) with DAPI (1 μg mL-1; Molecular Probes) and naïve T cells sorted based on live cells (DAPIneg) CD4pos CD62Lhi CD44lo using a FACSAria (BD Bioscience). T cells were labeled with CFSE (Molecular Probes) and co-cultured in the presence of OVA peptide (1 μg mL-1) and blocking antibodies to MHC class II (5 μg mL-1; M5/114; eBioscience) or IL-4Rα (5 μg mL-1; mIL4R-M1; BD Bioscience) in complete medium (DMEM Iscove's with 10% (vol/vol) heat-inactivated FBS, 2 mM glutamine, 100 U mL-1 of penicillin, 100 μg mL-1 of streptomycin, 25 mM HEPES and 50 μM β-mercaptoethanol). Cell-free supernatants were assessed for cytokine production by standard sandwich ELISA (eBioscience) following 4h stimulation with 50 ng mL-1 PMA, 750 ng mL-1 ionomycin (Sigma Aldrich).
3-5 ×106 CFSE-labeled OVA-specific CD45.2 CD4pos T cells were transferred i.v. into CD45.1 congenic recipient mice and 24 hours later immunized i.p. with 100 μg OVA emulsified in IFA, with one cohort receiving 5×104 IL-25-elicited c-kitint cells i.p. Proliferation of OVA-specific CD4pos T cells in the spleen was assessed two days post-immunization and cytokine production measured by ELISA. Trichuris muris infections were performed as previously described22. Trichuris-infected Il17e-/- mice were left untreated or given 5×104 IL-25-elicited c-kitint cells at day 10 post-infection. Worm counts were performed at day 20 post-infection. Mesenteric lymph nodes cells were collected at necropsy, plated in medium alone or polyclonally stimulated with 1 μg ml-1 each of αCD3 and αCD28 (eBioscience). Following 48 h, cell-free supernatants were assessed for cytokine production by sandwich ELISA (eBioscience). Trichuris-specific IgG1 antibody titers were analyzed by ELISA as described previously22. Total serum IgE was measured using the OptEIA IgE ELISA kit according to the manufacturer's instructions (BD Biosciences). For Nippostrongylus brasiliensis infections, WT mice were infected s.c. with 500 infective third-stage larvae (L3) and treated with 0.5 mg of anti-IL-25 or control IgG (from J.E.T. and A.L.B.) on days 0, 2, 4, 6 and 8. Mesenteric LN cells from infected mice were assessed at day 10 post-infection for the induction of c-kitint cells. For in vivo differentiation assays, 3×104 CD45.2 IL-25-elicited c-kitint-GFPneg cells were FACS-purified and transferred i.p. into CD45.1 congenic recipient mice. Recipient mice were treated 4 times with 1 μg each of SCF and IL-3 and differentiation of donor cells was assessed on day 6 post-transfer.
FACS-purified cell populations were plated onto semi-confluent OP9 stromal cells (ATCC # CRL-2749), as previously described30. Monolayers were irradiated (3000 rad) prior to co-culture with FACS-purified populations. Cells were incubated in the presence of SCF (50 ng mL-1; R&D systems) and IL-3 (10 ng mL-1; R&D systems).
Colon sections were fixed in 4% (vol/vol) paraformaldehyde and embedded in paraffin wax. 4 μm sections were stained with Periodic acid-Schiff/Alcian blue. Purified cell populations were subjected to cytospin (Cytospin 3, Thermo Fisher Scientific) and stained by Diff-quick for analysis of cellular morphology.
RNA from colonic tissues of mice was isolated by TRizol extraction (Invitrogen) or collected from sorted cell populations using RNeasy Mini kit (Qiagen). Whole tissues were homogenized with a tissue homogenizer (TissueLyzer; Qiagen) and cDNA was prepared with SuperScript Reverse Transcriptase (Invitrogen). Quantitative real-time PCR analysis used commercial QuantiTect primer sets for Il4, Il5, Il13, Gata3, Maf, Junb, Stat6 and Il1rl1 (Qiagen) and SYBR Green chemistry (Applied Biosystems). All reactions were run on an ABI 7500 Fast Real-Time PCR System (Applied Biosystems). Target genes were normalized for endogenous β-actin levels and relative quantification of samples were compared to controls.
Results are shown as means ± s.e.m for individual animals. Statistical significance was determined by Student's t-test. Results were considered significant at P < 0.05.
We thank members of the Artis laboratory for constructive discussions and Michael Abt, David Hill, Gregory Sonnenberg, Paul Giacomin, Meera Nair and Kevin Walsh for critical reading of the manuscript, and T. Chi and J.J. Bell for invaluable assistance with reagents and experimental design. Work in the Artis lab is supported by the National Institutes of Health (AI61570, AI074878 and AI083480 to D.A., F31 training grant GM082187 to S.A.S. and T32 training grant AI007532-08 to J.G.P.), the Burroughs Wellcome Fund (D.A.), a National Institute of Diabetes and Digestive Kidney Disease Center Grant (P30 DK50306) and pilot grants from the University of Pennsylvania (URF, VCID and PGI) (to D.A.).
Author Contributions S.A.S., M.C.S., J.G.P., S.P.S., T.K., A.B. and D.A. designed and performed the research. J.F.U., J.E.T., A.L.B., M.A.K. and R.A.K. provided new reagents. S.A.S., M.C.S., J.G.P. and D.A. analyzed the data. S.A.S and D.A. wrote the paper.
Competing Interest statement J.E.T. and A.L.B. are stockholding employees of Amgen. M.A.K. and R.A.K. are employees of SPB, a subsidiary of Merck&Co.
Unless otherwise stated, the authors declare no competing financial interests.