C57BL/6, Tbx21 ^-/- (C57BL/6 background), Rag2 ^-/- (BALB/c background), and Il27RA ^-/- (C57BL/6 background) mice were sourced commercially (all Charles River). Ifng ^-/- (C57BL/6 background) mice were a gift from Dr Anne O’Garra (The Francis Crick Institute, London). A colony of colitis-free TRnUC mice was generated from a descendant of the TRUC colony described previously (31). Stat1 ^-/- (B6.129P2-Stat1tm1Dlv) (32) and Stat4 ^-/- (C57BL/6J-Stat4em3Adiuj/J, purchased from The Jackson Laboratory) mice were housed under specific pathogen-free conditions according to Federation of European Laboratory Animal Science Associations (FELASA) guidelines. C57BL/6N and C57BL/6J mice were purchased from Janvier Labs and used as control mice for Stat1 ^-/- and Stat4 ^-/- mice respectively. Rosa26 ^YFP/+ (Jackson labs) mice were sourced commercially and bred with T-bet^cre/+ mice to generate the T-bet^Cre/+xRosa26 ^YFP/+ (T-bet^FM) mice.

cLP and Peyer’s patch-free SI LP leukocytes were isolated using a published method (33). Briefly, the epithelium was removed by incubation in HBSS lacking Mg²⁺ or Ca²⁺ (Invitrogen) supplemented with EDTA and HEPES. The tissue was further digested in HBSS lacking Mg²⁺ or Ca²⁺ supplemented with 2% foetal calf serum (FCS Gold, PAA Laboratories), 0.5 mg/ml collagenase D, 10 μg/ml DNase I, and 1.5 mg/ml dispase II (all Roche). The LP lymphocyte-enriched population was harvested from a 40–80% Percoll (GE Healthcare) gradient interface. For neutrophil analyses leukocytes were not purified by Percoll gradient centrifugation.

Flow cytometry was performed using a standard protocol. For ILC analyses a lineage cocktail of antibodies specific for CD3, CD45R, CD19, CD11b, TER-119, Gr-1, CD5, and FcϵRI was used. For a complete list of the antibodies used see Table 1 . LIVE/DEAD^TM stain (ThermoFisher Scientific Inc.) was used to determine cell viability. A FoxP3 staining kit (ebioscience) was used for intracellular staining of transcription factors and cytokines. In case of cytokine analysis, cells were pre-stimulated with 100 ng/ml PMA and 2 µM ionomycin in the presence of 6 µM monensin for 3–4 h prior to flow cytometry analysis. Samples were acquired using an LSRFortessa™ cell analyser (Becton Dickinson, USA) or a Cytoflex LX™ for the data on Stat1 ^-/- and Stat4 ^-/- mice. All the data were analyzed using FlowJo software (Tree Star, USA). Cell counts were determined using a fixed amount of inclusion beads (Spherotec, Inc.) as a reference in the flow cytometry tubes.

Faecal content extracted from the caecum of TRUC mice (31, 34) was reconstituted in sterile PBS 25% glycerol prior to storage at -80°C. Mice were orally gavaged with 200 μl aliquots of this faecal solution and sacrificed after 3 weeks.

Results are expressed as mean ± SEM. Data were analyzed using Student’s t-test or Mann-Whitney U test, as appropriate, using GraphPad Prism 5.0 (GraphPad Inc., USA). ns: non-significant; *p < 0.05; **p < 0.01; ***p <0.001; ****p <0.0001.

Animal experiments were performed in accredited facilities in accordance with the UK Animals (Scientific Procedures) Act 1986 (Home Office Licence Numbers PPL: 70/6792, 70/8127, and 70/7869). Mice for studies on STAT1 and STAT4 were bred at the animal facility of the Institute of Animal Breeding and Genetics, University of Veterinary Medicine Vienna according to the guidelines of the Federal Ministry of Science, Research and Economy section 8ff of the Animal Science and Experiments Act, Tierversuchsgesetz [TVG], BMWF-68.205/0068-WF/V/3b/2015.

We have reported that immunocompetent Tbx21-deficient mice do not develop spontaneous colitis (30). Tbx21-deficient mice lack ILC1 and NKp46⁺ ILC3, hence, we addressed the functionality of NKp46^- ILC3, which may also have a pathogenic role in colitis. For these analyses we define ILC as live CD45⁺ Lin^- CD127⁺ leukocytes ( Figure 1A ). Surprisingly, we detected an approximate 3-fold greater population of cLP NKp46-negative ILC3 in Tbx21 ^-/- C57BL/6 mice in comparison to wild type (WT) mice ( Figures 1B, C ). Within this NKp46^- ILC3 population, there is a greater cellularity of NKp46^-CCR6^- (double-negative) ILC3, but no significant difference in the cellularity of CCR6⁺ ILC3. Non-colitic Rag2 ^-/-xTbx21 ^-/- (TRnUC) mice also showed an enhanced cellularity of cLP NKp46^- ILC3 with a greater abundance of both NKp46^- CCR6^- and CCR6⁺ ILC3 ( Figures 1D, E ). Similar to cLP ILC3, there was a greater cellularity of small intestinal (SI) LP NKp46^- ILC3 in TRnUC mice and among these, the CCR6⁺ ILC3 population size was increased significantly ( Supplementary Figures 1A, B ). For these analyses, TRnUC mice on a BALB/c background were chosen because Rag2 ^-/-xTbx21 ^-/- mice on this background appear to be more prone to colitis than C57BL/6 background Rag2 ^-/-xTbx21 ^-/- mice (35). Overall, these observations support the notion that T-bet controls the cellularity of NKp46^- ILC3.

Enhanced cellularity of NKp46^- CCR6^- ILC3 in Tbx21-deficient mice supported a previous report indicating that these cells are the precursors of NKp46⁺ ILC3 (2). Hence, the lack of T-bet could result to a developmental blockade and the accumulation of NKp46^- CCR6^- ILC3. An interlinkage of NKp46⁺ ILC3 and NKp46^- CCR6^- ILC3 was also highlighted in another study reporting that NKp46⁺ ILC3 can lose NKp46 expression (36). In order to detect NKp46^- CCR6^- ILC3 with a history of T-bet expression, we generated a mouse model that expresses Cre-recombinase under the expression of the Tbx21 endogenous promoter by inserting an IRES-Cre cassette downstream of the Tbx21 stop codon ( Figure 2A ). This T-bet^Cre mouse was then bred to the Rosa26-lox-stop-lox YFP mouse (Rosa26^YFP/+) (37) to generate the T-bet^Cre/+xRosa26 ^YFP/+ fate-mapper mouse (T-bet^FM). As expected, cLP ILC1 defined as NKp46⁺NK1.1⁺T-bet⁺ were found to be T-bet fate mapper positive (T-bet^FM+) ( Figure 2B ), confirming the functionality of the model as these cells have previously been shown to express T-bet. Interestingly, T-bet^FM+ NKp46^- SI LP ILC3 were detected supporting the notion that these cells have potential to express this transcription factor ( Figure 2C ).

To further evaluate the ILC3 phenotype, we analyzed the cytokine profiles of CD127⁺ ILC in Tbx21 ^-/- and WT mice. As expected CD127⁺ ILC from Tbx21 ^-/- mice produced very low amounts of IFNγ, but surprisingly there was no altered expression of IL-17A on a per cell basis in the same mice ( Figures 3A, B ). However, due to the greater cellularity of CD127⁺ ILC in the intestine of mice lacking T-bet (13) we anticipate more ILC expressing IL-17A. Hence, we aimed to investigate whether T-bet also controls the cellularity of cLP neutrophils ( Figures 3C, D ). In line with the greater cellularity of NKp46^- ILC3 in Tbx21 ^-/- mice, there was indeed a greater neutrophilia in these mice. These neutrophils had an unaltered level of CD11b expression and granularity measured as SSC-A ( Figure 3E ) indicating that these neutrophils were not activated. We aimed to determine whether these neutrophils can be activated with a pathogenic microbiota. In order to test this Tbx21 ^-/- mice received an oral gauvage injection of fecal microbes derived from colitic Rag2 ^-/-xTbx21 ^-/- (TRUC) mice (31, 34). These Tbx21 ^-/- mice did not show weight abnormalities during the course of 3 weeks upon fecal microbial transplant (FMT) in comparison to Tbx21 ^-/- control mice ( Figure 3F ). Furthermore, the mass of colon and spleen did not show a significant difference due to FMT 3 weeks after the treatment ( Figure 3G ). FMT did also not result in altered cLP neutrophilia or cLP neutrophil activation detected by CD11b expression and granularity at this time point ( Figures 3H, I ). Hence, Tbx21 ^-/- mice appeared to be resilient to spontaneous colitis driven by the pathogenic microbes used.

We have shown that untreated Tbx21 ^-/- mice have a greater cellularity and activation of ILC2 (13). Hence, IL-5-producing ILC2 may promote an immune response counteracting to a more pathogenic response driven by ILC3 in naïve Tbx21-deficient mice (13). To explore this further we analyzed cLP eosinophilia in Tbx21 ^-/- mice, but detected no significant difference to WT mice ( Supplementary Figure 2 ).

Strikingly, the finding of enhanced cellularity of NKp46^- ILC3 in Tbx21 ^-/- mice correlated with enhanced RORγt expression in total NKp46^- CCR6^- ILC3, but not in CCR6⁺ ILC3 ( Figure 4A ). In contrast, in TRnUC mice RORγt expression was only enhanced in SI LP NKp46^- CCR6^- ILC3, but not SI LP CCR6⁺ ILC3 or cLP ILC3 ( Supplementary Figure 3 , Figure 4B ).

Previously, we have also reported that Tbx21 deficiency causes greater expression of the α chain of IL-7R (CD127) on total CD127⁺ ILC in the intestine (13). In the current study we can now pinpoint that within this Tbx21-deficient population cLP NKp46^- CCR6^- ILC3, but not CCR6⁺ ILC3 or ILC2 express more CD127 ( Figures 4C, D ). This was again in contrast to TRnUC mice as cLP and SI LP ILC3 and ILC2 in these mice did not display an altered CD127 expression in comparison to Rag2 ^-/- mice ( Figure 4E , Supplementary Figures 3B, C ). Mechanistically, we have reported previously that T-bet binds to the Cd127 locus in Th1 cells polarized in vitro (31, 38). Hence, it appears that T-bet is a regulator of Cd127 and Rorc (encoding RORγt) expression at the transcriptional level which may be factors limiting the cellularity of NKp46^- CCR6^- ILC3 in Rag-sufficient mice.

We further aimed to identify mediators upstream of T-bet that may be involved in limiting the cellularity of NKp46^- ILC3. Surprisingly, there was no alteration in the cellularity of NKp46^- CCR6^- and CCR6⁺ cLP ILC3 and no change in RORγt and CD127 expression in ILC3 derived from STAT1 (Stat1 ^-/-) or STAT4 (Stat4 ^-/-) deficient mice ( Figure 5A , Supplementary Figures 4C, D ). This observation was confirmed by the analysis of the same parameters in those ILC3 from mice deficient of IFNγ (Ifng ^-/-) or IL-27Rα (Il27RA ^-/-), both of which cause signaling events through STAT1 ( Supplementary Figures 4A–D ). Furthermore, the absence of STAT4 but not STAT1 in the germline did result in significantly altered IL-17A expression in CD127⁺ CD90.2⁺ cLP ILC ( Supplementary Figures 4E, F ).

T-bet-deficient mice have a greater cellularity of cLP ILC2 (13). In order to correlate with these data, we analyzed ILC2 abundance in Stat1 ^-/- and Stat4 ^-/- mice. As observed with NKp46-negative ILC3, the cellularity of cLP ILC2 did not alter in mice deficient of either STAT1 or STAT4 and CD127 expression levels in these cells did also not change ( Figure 5B , Supplementary Figure 5A ). Furthermore, CD127 expression was not altered in cLP ILC2 from Il27RA ^-/- mice ( Supplementary Figure 5A ). Interestingly, the potency of CD127⁺ CD90.2⁺ cLP ILC to co-produce IL-13 and IL-5 was reduced in Stat1 ^-/- but not in Stat4 ^-/- ( Supplementary Figure 5B ). Overall, STAT1 and STAT4 did not appear to play a critical role in in controlling the cellularity of the NKp46-negative ILC3 populations and ILC2 in the cLP ( Figure 5 ).