Report

β2-adrenergic receptor–mediated negative regulation of group 2 innate lymphoid cell responses

See allHide authors and affiliations

Science  02 Mar 2018:
Vol. 359, Issue 6379, pp. 1056-1061
DOI: 10.1126/science.aan4829

An off switch for helminth immunity

Group 2 innate lymphoid cells (ILC2s) are involved in responses to helminths, viruses, and allergens. Moriyama et al. found that ILC2s interact with the nervous system to modulate helminth immunity. ILC2s from the small intestine expressed the β2-adrenergic receptor (β2AR), which normally interacts with the neurotransmitter epinephrine. Inactivating β2AR resulted in lower helminth burden and more ILC2s, eosinophils, and type 2 cytokine production in mice. Conversely, treatment of helminth-infected mice with a β2AR agonist enhanced worm burden and reduced proliferation of ILC2s. Thus, β2AR negatively regulates ILC2-driven protective immunity.

Science, this issue p. 1056

Abstract

The type 2 inflammatory response is induced by various environmental and infectious stimuli. Although recent studies identified group 2 innate lymphoid cells (ILC2s) as potent sources of type 2 cytokines, the molecular pathways controlling ILC2 responses are incompletely defined. Here we demonstrate that murine ILC2s express the β2-adrenergic receptor (β2AR) and colocalize with adrenergic neurons in the intestine. β2AR deficiency resulted in exaggerated ILC2 responses and type 2 inflammation in intestinal and lung tissues. Conversely, β2AR agonist treatment was associated with impaired ILC2 responses and reduced inflammation in vivo. Mechanistically, we demonstrate that the β2AR pathway is a cell-intrinsic negative regulator of ILC2 responses through inhibition of cell proliferation and effector function. Collectively, these data provide the first evidence of a neuronal-derived regulatory circuit that limits ILC2-dependent type 2 inflammation.

The type 2 inflammatory response is a highly conserved module of the innate and adaptive immune system that is elicited after exposure to infectious and environmental triggers, such as helminth infections, allergens, venoms, and other stimuli (13). The hallmarks of type 2 responses include activation of T helper 2 (TH2) cells and release of type 2 cytokines, such as interleukin-4 (IL-4), IL-5, IL-9, and IL-13; production of immunoglobulin E (IgE); and the activation of multiple effector cells, such as basophils, mast cells, and eosinophils. This cascade of immunologic events is associated with increased mucus production from goblet cells and induction of smooth-muscle contractility that contributes to the elimination of pathogens or allergic stimuli (4, 5). Recent studies identified group 2 innate lymphoid cells (ILC2s) as a potent source of type 2 cytokines that contribute to the type 2 inflammatory responses (511). Although considerable advances have been made to define the cytokines, growth factors, and environmental stimuli that trigger ILC2 responses, the regulatory mechanisms that constrain ILC2 responses and type 2 inflammation at mucosal sites remain incompletely understood.

Mucosal and lymphoid tissues are highly innervated (1214), and immune cells—including mast cells, macrophages, ILCs, and T cells—can be regulated by neuronal-derived bioactive molecules in certain circumstances (1523). However, our knowledge of the complex interactions between the nervous system and ILCs in the context of type 2 inflammation is still limited. RNA sequencing (RNA-seq) of sorted murine ILC subsets revealed that small intestinal (SI) ILC2s exhibited higher levels of the β2AR gene (Adrb2) mRNA expression compared to SI ILC3s, but lower levels of other adrenergic receptors (Fig. 1A). Consistent with these results, ILC2s sorted from SI lamina propria (SILP), gut-draining mesenteric lymph node (mLN), colon LP, and lung exhibited higher levels of Adrb2 mRNA expression compared to SILP ILC3s (Fig. 1B and fig. S1A). ILC2s from mesenteric white adipose tissue exhibited lower levels of Adrb2 mRNA expression compared to ILC2s from other tissue. We also detected ADRB2 mRNA in ILC2s and CD4+ T cells sorted from human lung and peripheral blood mononuclear cells (Fig. 1C). These results show that the β2AR gene is expressed in murine and human ILC2s.

Fig. 1 Adrb2 expression in ILCs and localization of ILC2s in gut-associated tissues.

(A) Normalized counts of adrenergic receptor expression in SI ILC2s and ILC3s by RNA-seq are shown as a heat map. Adra genes encode α-ARs. (B) Relative expression of Adrb2 in immune cells sorted from indicated tissues. SILP, mLN, and colon LP ILC2s were sorted as KLRG1+CD127+CD90+Lineage negative (Lin)CD45+, and lung and mesenteric white adipose tissue (mWAT) ILC2s were sorted as ST2+CD127+CD90+LinCD45+. SILP ILC3s were sorted as CCR6+CD127+CD90+LinCD45+. SPL, spleen; NK, natural killer cells; DC, dendritic cells; Eo, eosinophils. Data represent mean ± SEM of five pooled experiments. (C) Relative expression of ADRB2 in CRTH2+CD127+LinCD45+ ILC2s and CD4+ T cells sorted from human lung and peripheral blood mononuclear cells (PBMCs). Data represent mean ± SEM of two pooled experiments. Each circle represents data from one donor. (D and E) Representative sections of SI from a Bcl11b-tdTomato (red) reporter mouse stained for KLRG1 (green), CD3ε (blue), NKp46 (cyan), and TH (white). V, villi; S, submucosa; M, muscularis. Arrows show ILC2s. Data are representative of two experiments. (F) Representative section of SI from a Il13–fate mapping (red) mouse stained for KLRG1 (green), CD3ε (blue), and TH (white). Data are representative of two experiments. (G and H) Relative expression of Th and Dbh in intact SI tissue and in the indicated fractions of SI. LPL, lamina propria leukocytes. ***P < 0.001 and ****P < 0.0001 by one-way analysis of variance (ANOVA) with Dunnett’s multiple comparison. (I to K) Representative section of mLN from a Bcl11b-tdTomato (red) reporter mouse stained for KLRG1 (green), CD3ε (blue), and IgD (cyan). Arrows show ILC2s. Data are representative of two experiments.

Examination of the localization of ILC2s (defined as KLRG1+Bcl11b+CD3εNKp46) and adrenergic neurons [marked by tyrosine hydroxylase (TH) expression] in the SI microenvironment revealed that ILC2s were found in close proximity to TH+ neurons in the villi and submucosa but not in the muscularis (Fig. 1, D and E, and fig. S1B). To further examine this, we generated Il13–fate mapping mice by crossing Il13cre and Ai14 (tdTomato) mice and found that ILC2s (defined as KLRG1+tdTomato+CD3ε) colocalize with TH+ neurons (Fig. 1F and fig. S1, C and D). Notably, the highest expression levels of genes encoding enzymes involved in norepinephrine synthesis, Th and Dbh (dopamine β-hydroxylase), were found in the parenchyma fraction in the SI (Fig. 1, G and H), which collectively suggests that ILC2s may respond to neuronal-derived norepinephrine, a β2AR ligand. ILC2s were also observed in the interfollicular region (24), paracortical region, and the medulla of the draining mLN, which are anatomical sites where sympathetic adrenergic neurons are abundant (12, 13) (Fig. 1, I to K, and fig. S1, E and F).

To analyze the effect of β2AR signaling on ILC2 development at steady state, β2AR-sufficient (Adrb2+/+) and β2AR-deficient (Adrb2−/−) mice were analyzed. Adrb2+/+ and Adrb2−/− mice had comparable numbers of ILC2 progenitors (ILC2Ps) (25) in the bone marrow (BM, fig. S2A). Additionally, Adrb2+/+ and Adrb2−/− mice had similar frequencies of all ILC subsets in mLNs and in SILP (fig. S2, B to D), indicating that β2AR deficiency does not result in impaired ILC2 development or homeostasis at steady state.

Next, we examined whether β2AR signaling regulates ILC2 responses and type 2 inflammation after exposure to the gastrointestinal helminth Nippostrongylus brasiliensis. This infection induces potent ILC2 responses that play an important role in the expulsion of the parasite through production of IL-5 and IL-13 (10, 26, 27). After infection, Adrb2−/− mice exhibited increased frequencies and numbers of ILC2s and increased numbers of IL-13–producing ILC2s, but comparable frequencies of ILC1s and ILC3s, in mLNs and in SILP compared to Adrb2+/+ mice (Fig. 2, A and B, and fig. S2E). The expression of inflammatory cytokines encoded by genes Il25, Il33, and Tslp, which stimulate ILC2s, was comparable between Adrb2+/+ and Adrb2−/− mice in the SI after infection (fig. S2F). However, exaggerated eosinophilia and goblet cell hyperplasia (Fig. 2, C and D) and reduced worm burdens (Fig. 2E) were observed in Adrb2−/− mice. Of note, the amount of norepinephrine in the SI was comparable at 0, 4, and 7 days postinfection (fig. S2G), suggesting that expression of this ligand is not altered during infection. Notably, although ILC2s still expressed Adrb2 after infection, expression was at a lower level than observed in naive mice (fig. S2H), suggesting that ILC2-intrinsic β2AR expression is dynamically regulated in response to inflammatory cues within the tissue microenvironment. Taken together, these results indicate that the β2AR pathway may be an important negative regulator of ILC2 responses.

Fig. 2 Regulation of anti-helminth responses by β2AR-deficiency or agonist treatment.

(A to E) Adrb2+/+ and Adrb2−/− mice were analyzed 4 days [(A) and (B)] and 7 days [(C) to (E)] after N. brasiliensis infection. Shown are (A) flow cytometry plots of LinCD45+ cells and enumeration of ILC2 percentages and numbers; (B) cytokine production; and (C) Siglec F+SSChi eosinophil percentages in mLNs. Also shown are (D) representative SI sections with periodic acid–Schiff (PAS)–Alcian blue staining and enumeration of goblet cell numbers and (E) worm burdens in SI. Data are representative of three experiments. (F to J) Β6 mice treated with β2AR agonist clenbuterol (Clen) or vehicle (Veh) were analyzed 7 days after N. brasiliensis infection. Shown are (F) flow cytometry plots of LinCD45+ cells and enumerations of ILC2 percentages and numbers; (G) cytokine production; and (H) Siglec F+SSChi eosinophil numbers in mLNs. Also shown are (I) representative SI sections with PAS–Alcian blue staining and enumeration of goblet cell numbers and (J) worm burdens in SI. Data are representative of three experiments. (K to M) B6 mice treated with anti-CD4 (α-CD4) mAb together with clenbuterol or vehicle were analyzed 10 days after N. brasiliensis infection. Shown are (K) enumerations of ILC2 percentages, (L) cytokine production, and (M) worm burdens in SI. Data are representative of two experiments. (N and O) Il7rcre/+ and Il7rcre/+ Adrb2f/f mice treated with α-CD4 mAb were analyzed 10 days after N. brasiliensis infection. Shown are (N) enumerations of ILC2 percentages and (O) worm burdens in SI. Data are pooled from two experiments. For panels (A) to (O), each circle represents data from one mouse, the numbers in flow cytometry plots represent mean ± SEM in each gate, and bar graphs represent mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 by unpaired two-tailed Student’s t test. (P and Q) Rag2−/− Il2rg−/− mice reconstituted with ILC2Ps from Adrb2+/+ or Adrb2−/− mice were analyzed 7 days after N. brasiliensis infection. Shown are enumerations of (P) ILC2 percentages and (Q) Siglec F+SSChi eosinophil percentages in SI. Each circle represents data from one mouse. Bar graphs represent mean ± SEM. Data are representative of three experiments. *P < 0.05 and **P < 0.01 by one-way ANOVA with Dunnett’s multiple comparison.

Given that β2AR-deficient mice had exaggerated ILC2 responses, we next sought to test whether treatment with a β2AR agonist, clenbuterol, would inhibit ILC2 responses and dampen type 2 inflammation. Although the proportion of ILC subsets was not changed by β2AR agonist treatment at steady state (fig. S2I), β2AR agonist–treated wild-type C57BL/6 (B6) mice exhibited significantly fewer ILC2s in the mLNs compared to vehicle-treated mice after N. brasiliensis infection (Fig. 2F). Further, agonist treatment was associated with significantly fewer IL-5– and IL-13–producing ILC2s, reduced eosinophilia, and diminished goblet cell responses (Fig. 2, G to I, and fig. S2J). Consistent with these results, agonist-treated mice had significantly higher worm burdens (Fig. 2J). We also treated the mice with another β2AR agonist, salmeterol, during N. brasiliensis infection and found reduced anti-helminth responses (fig. S2K), indicating that both β2AR agonists have an inhibitory effect on type 2 inflammation.

Because type 2 cytokine production from ILC2s was reduced after simultaneous exposure to a β2AR agonist and helminth infection in vivo (Fig. 2G and fig. S2J), we next sought to test whether ILC2 effector function is regulated by short-term β2AR stimulation in vitro. SILP cells from Il13–YFP (yellow fluorescent protein) reporter mice were cultured with IL-33 for 4 hours in the presence of a β2AR agonist or control vehicle. Agonist-treated cells exhibited lower frequencies of YFP-expressing ILC2s compared to vehicle-treated controls (fig. S2L), suggesting that signaling through β2AR negatively regulates acute activation and IL-13 production from ILC2s.

In addition to ILC2s, β2AR agonists can alter the TH1/TH2 cell balance by inhibiting the development of TH1 cells (28). Therefore, we examined whether β2AR stimulation by the agonist could inhibit ILC2 responses and inflammation in the absence of CD4+ T cells. After treatment of B6 mice with a CD4–depleting monoclonal antibody (mAb) and the β2AR agonist, fewer ILC2s were observed after N. brasiliensis infection than were observed in vehicle-treated mice (Fig. 2K). Additionally, mice treated with both anti-CD4 mAb and agonist exhibited fewer cytokine-producing ILC2s and increased worm burdens compared to vehicle-treated mice (Fig. 2, L and M). Together, these data indicate that β2AR agonist–mediated inhibition of ILC2 responses and inflammation occurs in the absence of CD4+ T cells.

To further analyze the effects of β2AR deletion on ILCs, we crossed Adrb2-flox (Adrb2f/f) mice with transgenic mice expressing a Cre recombinase in the Il7r locus (Il7rcre/+) to generate β2AR conditional knockout mice. After N. brasiliensis infection together with anti-CD4 mAb treatment, Il7rcre/+ Adrb2f/f mice exhibited increased ILC2 frequencies and reduced worm burdens compared to control Il7rcre/+ mice (Fig. 2, N and O). β2AR agonist treatment of Il7rcre/+ Adrb2f/f mice did not induce significant changes in eosinophilia and worm burden after helminth infection (fig. S2, M and N). Furthermore, irradiated Adrb2+/+ and Adrb2−/− mice reconstituted with Adrb2+/+ BM cells had similar worm burdens after helminth infection (fig. S2O). These results indicate that β2AR signaling on CD127 (IL-7 receptor α, IL-7Rα)–expressing CD4 hematopoietic cells, including ILC2s, is important during anti-helminth responses but not essential for radio-resistant cells and CD127-negative cells.

To further address the importance of β2AR signaling directly on ILC2s, we reconstituted ILC-deficient Rag2−/− Il2rg−/− mice with Adrb2+/+ or Adrb2−/− ILC2s by transferring ILC2Ps from Adrb2+/+ or Adrb2−/− mice and then infecting the mice with N. brasiliensis. After 7 days of infection, there was a trend toward increased SI ILC2s, and we observed increased eosinophilia in Adrb2−/− ILC2–reconstituted mice compared to Adrb2+/+ IL7C2–reconstituted mice (Fig. 2, P and Q), further supporting the importance of β2AR on ILC2s in controlling anti-helminth responses.

Because lung ILC2s exhibited high levels of Adrb2 mRNA expression (Fig. 1B) and lung tissue is also highly innervated (15, 19), we examined if β2AR signaling controls ILC2 responses in the lung as it does in the intestine. After inoculation, N. brasiliensis larvae migrate to and induce inflammation in the lung before reaching the intestine. Adrb2−/− mice exhibited increased frequencies of ILC2s in the lung compared to Adrb2+/+ mice after infection (Fig. 3A). As an additional approach to induce lung inflammation, we administered the alarmin IL-33 intranasally and observed increased frequencies of ILC2s in Adrb2−/− mice compared to that in Adrb2+/+ mice (Fig. 3B). Further, ILC2 frequencies and cytokine production after IL-33 administration were inhibited by β2AR agonist treatment (Fig. 3, C and D). Similarly, when the mice received extract of the fungal allergen Alternaria alternata intranasally, β2AR agonist–treated mice had reduced frequencies of ILC2s in the lungs than did vehicle-treated mice (Fig. 3E). In addition, Il7rcre/+ Adrb2f/f mice exhibited increased ILC2 frequencies and cytokine production compared to Il7rcre/+ mice after the Alternaria extract administration together with anti-CD4 mAb treatment (Fig. 3, F and G). Collectively, these results indicate that β2AR signaling is an evolutionarily conserved regulatory pathway serving to dampen ILC2 responses against diverse inflammatory stimuli at multiple mucosal barrier surfaces.

Fig. 3 β2AR signaling inhibits ILC2 responses in lung inflammation.

(A) Lung cells from Adrb2+/+ and Adrb2−/− mice were analyzed 4 days after N. brasiliensis infection. Shown is an enumeration of CD127+CD90+LinCD45+ ILC2 percentages. (B to D) Adrb2+/+ and Adrb2−/− mice (B) and B6 [(C) and (D)] mice treated with β2AR agonist salmeterol (Salm) or vehicle were intranasally (i.n.) administered IL-33 for 3 days and analyzed 4 days later. Shown are flow cytometry plots of LinCD45+ cells [(B) and (C)] and CD127+CD90+LinCD45+ ILC2s (D). Also shown are enumerations of CD127+CD90+LinCD45+ ILC2 percentages [(B) and (C)] and cytokine production (D). The numbers in flow cytometry plots represent percentages in each gate. (E) B6 mice treated with β2AR agonist salmeterol or vehicle were intranasally administered an Alternaria extract for 3 days and then analyzed 4 days later. Shown is an enumeration of CD127+CD90+LinCD45+ ILC2 percentages. (F and G) Il7rcre/+ and Il7rcre/+ Adrb2f/f mice treated with α-CD4 mAb were intranasally administered an Alternaria extract for 3 days and then analyzed 4 days later. Shown are enumerations of cytokine production. For all panels, each circle represents data from one mouse, bar graphs represent mean ± SEM, and data are representative of two experiments. *P < 0.05 and **P < 0.01 by unpaired two-tailed Student’s t test.

These data provoke the hypothesis that the signaling through β2AR on ILC2s negatively regulates ILC2 responses and type 2 inflammation after exposure to helminth infection and allergens. To investigate the mechanisms underlying this effect, we performed RNA-seq analysis on ILC2s sorted from N. brasiliensis–infected Adrb2+/+ mice with or without β2AR agonist treatment. Although ILC2s from both groups exhibited similar expression levels of genes encoding ILC2-associated transcription factors (Id2, Gata3, and Rora) and cytokine receptors (Il1rl1, Crlf2, and Il4ra) (Fig. 4A), we also observed significant down-regulation of a number of genes in the agonist-treated groups compared to that in controls. Gene set enrichment analysis (GSEA) of the down-regulated genes revealed significant enrichment in gene ontology (GO) terms associated with the cell cycle and cell proliferation (Fig. 4, B and C, and fig. S3A). Taken together with the reduced ILC2 numbers observed in these mice after infection (Fig. 2F), these results collectively suggest a role for β2AR signaling in limiting the proliferation and accumulation of ILC2s.

Fig. 4 β2AR stimulation negatively regulates ILC2-intrinsic cell proliferation.

(A to C) KLRG1+CD127+CD90+LinCD45+ mLN ILC2s sorted from N. brasiliensis–infected Adrb2+/+ mice treated with or without clenbuterol were analyzed by RNA-seq. Shown are (A) a scatter plot showing mean of normalized counts, (B) bar graphs of the significantly enriched [false discovery rate (FDR) < 0.01] GO terms with the highest normalized enrichment scores (NES), and (C) a GSEA plot of the GO term “cell cycle.” (D and E) Β6 mice treated with clenbuterol or vehicle were analyzed for percentages of Ki67+ in ILC2s (D) 4 days after N. brasiliensis infection and (E) 8 days after H. polygyrus infection. Data are representative of two experiments. (F) Adrb2+/+ and Adrb2−/− mice intraperitoneally (i.p.) administered IL-33 were analyzed for Ki67+ in ILC2s. Data are representative of two experiments. (G) Sorted SI ILC2s stained with cell-proliferation dye were cultured with IL-2, IL-7, IL-33, and β2AR agonist salmeterol or vehicle for 4 days, and the enumeration of proliferation dye geometric mean fluorescent intensity (GMFI) is shown. n = 3. Data are representative of three experiments. For panels (D) through (G), each circle represents data from one mouse, and bar graphs represent mean ± SEM; *P < 0.05 and **P < 0.01 by unpaired two-tailed Student’s t test. (H) Chimeric mice reconstituted with congenic Adrb2+/+ and Adrb2−/− BM cells were analyzed at steady state and 7 days after N. brasiliensis infection. Flow cytometry plots and enumerations of ILC2 percentages for each genotype are shown. Connected symbols represent data from one chimeric mouse. Data are representative of two experiments. **P < 0.01 by paired two-tailed Student’s t test.

To directly test whether β2AR signaling regulates ILC2 proliferation in vivo, we analyzed the expression of the proliferation marker Ki67 during N. brasiliensis infection in mice treated with the β2AR agonist. After treatment, ILC2s from these mice exhibited reduced frequencies of Ki67-expressing ILC2s, but comparable frequencies of apoptotic ILC2s, compared to those of vehicle-treated control mice after infection (Fig. 4D and fig. S3B), suggesting that β2AR stimulation suppresses ILC2 proliferation but does not regulate apoptosis. Further, this regulation was not restricted to N. brasiliensis, because it was also observed after infection with another gastrointestinal helminth, Heligmosomoides polygyrus bakeri (Fig. 4E). In addition, when Adrb2+/+ and Adrb2−/− mice were injected with IL-33 intraperitoneally, the frequencies of Ki67+ ILC2s were significantly higher in Adrb2−/− mice than in Adrb2+/+ mice (Fig. 4F). Together, these data indicate that β2AR signaling negatively regulates ILC2 proliferation in vivo during type 2 inflammation.

To examine whether β2AR-dependent inhibition of ILC2 proliferation is cell intrinsic, sorted ILC2s were stained with a proliferation dye (CellTrace Violet) and then cultured in vitro with a β2AR agonist. Agonist-treated ILC2s exhibited significantly higher levels of CellTrace Violet intensity compared to vehicle-treated ILC2s, suggesting that the reduced proliferation after β2AR agonist treatment occurs through a cell-intrinsic mechanism (Fig. 4G). To further test this in vivo, we reconstituted irradiated B6 mice with congenically marked Adrb2+/+ and Adrb2−/− BM cells and compared the proliferation of Adrb2+/+ and Adrb2−/− ILC2s in the same chimeric mouse (Fig. 4H). At steady state, the chimeric mice had similar frequencies of Adrb2+/+ and Adrb2−/− ILC2s, in accordance with comparable ILC2 frequencies in naive Adrb2+/+ and Adrb2−/− mice (fig. S2, B to D). However, after N. brasiliensis infection, the frequencies of Adrb2−/− ILC2s were increased compared to the frequencies of Adrb2+/+ ILC2s (Fig. 4H), indicating that β2AR signaling inhibits ILC2 proliferation in a cell-intrinsic manner in vivo.

Collectively, these results reveal a previously unrecognized regulatory circuit that operates between the adrenergic nervous system and the innate immune system to control type 2 inflammation at multiple mucosal sites. Specifically, signaling through β2AR negatively regulates ILC2 proliferation and effector function. Given the importance of ILC2s in driving type 2 immune responses (10, 26, 27) and the altered β2AR expression levels on ILC2s during inflammation, it appears that β2AR may function as a molecular rheostat to fine-tune the ILC2 response, thus controlling the balance between promotion of host-protective acute ILC2 responses and prevention of chronic pathologic type 2 inflammation. Intriguingly, in contrast to the adrenergic-mediated negative regulation observed here, recent studies have identified cholinergic neurons as potent activators of intestinal and lung ILC2s through the production of the neuropeptide neuromedin U (2022). Thus, the mammalian nervous system appears to have evolved dual mechanisms to rapidly activate or repress these innate immune cells to protect the host against diverse inflammatory stimuli.

Supplementary Materials

www.sciencemag.org/content/359/6379/1056/suppl/DC1

Materials and Methods

Figs. S1 to S3

Table S1

References (3045)

References and Notes

Acknowledgments: We thank the D. Artis lab members and the G. F. Sonnenberg lab members for discussion and critical reading of the manuscript. We also thank D. Mucida (The Rockefeller University) and P. A. Muller (The Rockefeller University) for advising on TH staining, S. Thomas (University of Pennsylvania), G. Karsenty (Columbia University), P. Liu (the Wellcome Trust Sanger Institute), and J. Sun (Memorial Sloan Kettering Cancer Center) for mice. We also thank D. Farber (Columbia University) and LiveOnNY for human lung samples. Funding: This work was supported by grants from The Naito Foundation (to S.M.), the Japan Society for the Promotion of Science (JSPS) Overseas Research Fellowships (to S.M.), the Novo Nordic Foundation (grant 14052 to J.B.M.), the German Research Foundation (grant KL 2963/1-1 to C.S.N.K), an Australian National Health and Medical Research Commission Early Career Fellowship (to L.C.R.), the Jill Roberts Institute (G.G.P.), a Weill Cornell Medicine Pre-Career Award (to L.A.M.), the NIH (grants F32-DK109630-01 to N.A.Y.; F32-AI134018-01 to L.A.M.; and AI061570, AI087990, AI074878, AI083480, AI095466, AI095608, AI102942, and AI097333 to D.A.), a European Research Council Advanced Grant (742883 to H.-R.R.), the Burroughs Wellcome Fund (grant to D.A.), and the Crohn’s & Colitis Foundation of America (grant to D.A.). Author contributions: S.M., J.R.B., A.-L.F., J.B.M., C.S.N.K., L.C.R., L.A.M., and D.A. designed and performed the research. S.M., C.S.N.K., G.G.P., N.A.Y., and D.A. analyzed the data. H.-R.R. provided the Il7rcre mice. S.M. and D.A. wrote the manuscript with input from the other authors. Competing interests: The authors declare no competing interests. Data and materials availability: Data presented in this manuscript are tabulated in the main paper and in the supplementary materials. RNA-seq data are deposited under accession number GSE108884 in the Gene Expression Omnibus database.
View Abstract

Stay Connected to Science

Subjects

Navigate This Article