Intracellular Hedgehog signaling controls Th17 polarization and pathogenicity


 T helper 17 (Th17) cells play a key role in barrier protection against fungal and bacterial pathogens but are also pathological drivers of many inflammatory diseases. Although the transcription factor networks governing Th17 differentiation are well defined, the signaling pathways that regulate the development and function of this important CD4+ T cell subset are still poorly understood. Hedgehog (Hh) signaling plays important roles in regulating cell fate decisions during embryogenesis and adult tissue patterning. Using novel CD4-specific Hh knockout mice, we find that intracellular Hh signaling, independently of exogenous Hh ligands, selectively drives Th17 lineage differentiation but not the development of Th1, Th2, or iTreg CD4+ Th cells. We show that the endogenous Indian Hh (Ihh) ligand signals via the signal transducer Smoothened to activate both canonical and non-canonical Hh pathways, through the Gli3 transcription factor and AMPK phosphorylation, respectively. Using two models of intestinal inflammation, we demonstrate that inhibition of the Hh pathway with either the clinically approved small molecule inhibitor vismodegib or genetic ablation of Ihh in CD4+ T cells greatly diminishes disease severity. Taken together, we have uncovered Hh as a novel signaling pathway controlling Th17 differentiation and Gli3 as a crucial transcription factor in this process. Our work paves the way for a potential use of Hh inhibitors in the treatment of inflammatory bowel disease and other autoimmune diseases.


INTRODUCTION
60 The adaptive immune system is able to initiate highly tailored immune responses. After antigenic stimulation, naïve CD4 + T cells can differentiate into specialized T helper (Th) cell lineages (Th1, Th2, Th17 and Tregs), characterized by the production of key effector cytokines that dominate subsequent immune responses. Th lineages are each able to respond to different classes of immune challenges. Th1 cells secrete IFNg to promote 65 immunity against intracellular pathogens, while Th2 cells secrete IL-4, IL-5 and IL-13 to respond to parasitic and helminth infection. Th17 cells produce IL-17 and IL-22 required for clearance of extracellular pathogens, while Tregs suppress immune responses (reviewed in 1 ).

70
This lineage-specific fate decision is hence of great importance and is governed by environmental signals received by the CD4 + T cell. These include the mechanism through which the T cell is activated by antigen presenting cells (APCs)  Th17 differentiation is initiated by TGFb and inflammatory cytokines, which signal through SMAD2 and STAT3, respectively, to induce the Th17 differentiation program; this is 80 guided by a number of transcription factors including the 'master' transcription factor RORgt 3 . However, although the transcription factor network governing Th17 identity has been well described 4 , the intracellular signaling pathways regulating this complex differentiation program are less clear.

85
Th17 cells are critical for maintaining the integrity of intestinal barrier surfaces and coordinating the immune response against pathogenic extracellular bacteria and fungi.
However, Th17 cells are also key drivers of autoimmune diseases including inflammatory bowel disease (IBD), rheumatoid arthritis and multiple sclerosis 5 . There is therefore strong clinical interest to specifically inhibit differentiation and pathogenicity of these cells. In the immune system, the Hh signaling pathway has been implicated in T cell development in the thymus 8 . Given the structural and morphological similarities between the primary cilium and immune synapses 9, 10 , we were prompted to study Hh signaling in 105 mature T cells 11 . TCR signaling at the immune synapse upregulates Hh components in CD8 + T cells independently of exogenous Hh ligands, which we have shown to be important for CD8 + T cell killing. Our data raised the possibility that Hh signaling in CD8 + T cells may occur intracellularly 11 . By contrast, less is known about Hh signaling in mature CD4 + T cells. One study has shown that transgenic expression of Gli2 activator and 110 repressor forms in CD4 + T cells can influence Th2 differentiation 12 , but the role of endogenous Hh signaling in Th polarization is not known.
Here, we find that Hh signaling is functionally important for Th17 differentiation but not Th1, Th2, and iTreg CD4 + T cell development. While conventional Hh signaling involves paracrine signaling -in which Hh ligand is secreted from one cell and signals to another 115 cell that expresses the Hh receptor -we discovered that in CD4 + T cells Hh signaling is intracellular and is mediated by Ihh, Smo and Gli3. Mechanistically, Smo activates Gli3 which is needed for the induction of IL-17a, and also regulates metabolic fitness in Th17 cells in a non-canonical fashion through AMPK phosphorylation. Functionally, we show that blocking Hh signaling in vivo genetically or with small molecule Hh inhibitors 120 ameliorates disease in two models of Th17-driven inflammatory bowel disease.

Core Hedgehog (Hh) signaling components are specifically upregulated in Th17 125
cells Previous work has suggested that Hh components Smo and Ptch are expressed in CD4 + T cells 13 . We extended these analyses by comprehensively profiling Hh ligands, receptors, signal transducer and transcription factors (Fig. 1A) throughout CD4 Th differentiation in an established in vitro polarization protocol (Suppl.  (Fig. 1C). Using our new monoclonal antibody, we find that Smo resides on the plasma membrane and on intracellular vesicles as has been described for CD8 + T cells previously 11 (Fig. 1D). 145 Vertebrate Hh signaling activity can be evaluated by the expression of the downstream Gli transcription factors 14 . All lineages apart from Th2 modestly upregulated Gli1 upon Th polarization. Given its potential to regulate Th2 differentiation 12 , we were interested in assessing Gli2 expression. However, we were unable to detect Gli2 transcripts in any of the conditions tested (Suppl. Fig. 2). Strikingly, Gli3 was the only transcription factor 150 showing lineage-specific expression in Th17 and iTreg T cells peaking at day 3 of culture ( Fig. 1E). Taken together, while all lineages expressed core Hh components, only Th17 cells   express high levels of Smo, as well as Gli1 and Gli3 transcription factors by day 3 of   culture when lineage choice is

Th17 lineage-polarizing cytokines induce the expression of central Hh signaling components. 175
Th17 lineage polarization is initiated when naïve CD4 + T cells recognize cognate antigen on APCs in the presence of a cytokine milieu dominated by IL-6, TGFb, IL-1b and IL-23.
Given the Th17-specific upregulation of Smo and Gli1/Gli3 (Fig.1), we asked whether Th17-inducing cytokines play a role in the induction of these Hh signaling components.
First, we titrated TGFb in the presence of fixed concentrations of IL-6, IL-23, and IL-1b 180 and assessed Gli1 and Gli3 levels by qRT-PCR and Smo by western blot. Strikingly, Gli3 but not Gli1 expression was upregulated by TGFb in a dose-dependent manner ( Fig. 2A, upper panel), which is critically dependent on the presence of IL-6. 1 µM TGFb led to optimal IL-17a expression (Suppl. Fig 4) and highest Smo levels ( Fig. 2A, lower panel).
Interestingly, IL-6 alone was necessary and sufficient to induce Smo protein and stepwise 185 addition of TGFb, IL-23, and IL-1b increased not only IL-17a production (Suppl. Fig 4A) but also Gli3 RNA expression levels (  In the canonical pathway, Hh signaling is initiated upon binding of the exogenous Hh ligand, generated by Hh-producing cells, to the receptor Ptch on the plasma membrane of Hh-responsive cells 7 . We wondered whether exogenous Hh ligand could promote Th17 polarization under Th0 stimulation conditions or in the presence of low (IL-6 alone), 195 intermediate (IL-6 and TGFb), or strong (IL-6, TGFb, IL-23, and IL-1b) Th17 polarizing conditions (Fig. 2C, Suppl. Fig 4B). Surprisingly, even high concentrations of active Ihh N-peptide were unable to promote Th17 differentiation in any condition tested. This points towards a novel intracellular form of Hh signaling in CD4 + T cells.

210
Naïve CD4 + T cells were stimulated with the indicated polarizing cytokines in the presence or absence of recombinant N-terminal murine Ihh fragment at the indicated concentrations. All cells were polarized in the presence of anti-IFNg and anti-IL4 blocking antibodies and were harvested for analysis by FACS on day 5. Representative flow cytometry plots of n = 3 independent experiments are shown with a summary on the right. Data are means +/-SD. p-values were 215 calculated using an unpaired two-tailed Student's t test. * p<0.05, ** p<0.01, *** p<0.001.

Small molecule Hedgehog inhibitors selectively impair Th17 polarization in vitro.
Hh signaling can be efficiently blocked by small molecule inhibitors of the key signal transducer Smo. We investigated the effect of Hh inhibitor cyclopamine on CD4 + Th 220 lineage polarization by treating T cell cultures with inhibitor for the first 3 days (when lineage identity is determined in vitro) and assessing lineage identity on day 5 (Fig. 3A).
Strikingly, only Th17 (IL-17a, top row) but not Treg (FoxP3, bottom row) polarization was compromized in a dose-dependent manner (Fig. 3B). Viability and cell number of Th17 cells were not significantly affected at the inhibitor concentrations used (Fig. 3B, last two 225 columns). Similar results were obtained with clinically approved Smo inhibitor vismodegib 15 (Fig. 3C). Cyclopamine treatment also downregulated surface CCR6 expression, a hallmark protein of Th17 cells, in a dose-dependent manner ( Fig. 3D) but did not affect Th1/Th2 differentiation (Fig. 3E). The same dose-dependent downregulation of CCR6 was also seen in vismodegib-treated Th17 cells (Suppl. Fig 5). 230 Next, we investigated the window of active Hh signaling during in vitro CD4 + T cell differentiation. For this, we limited the Hh inhibitor treatment to the last 24h of the 5-day Th17 and Treg differentiation culture before assaying them by flow cytometry on day 5 ( Fig. 3F). Interestingly, we found that inhibition of Hh signaling post day 4 did not affect 235 IL-17a production or FoxP3 levels indicating that Hh signaling is active in the first 3 days of culture. These results fit nicely with the high expression of Hh signaling components during this time. Taken together, we found that Hh signaling is crucial for Th17 polarization but not Th1, Th2, or iTreg differentiation.

Conditional ablation of Hh components in CD4 + T cells impairs Th17 polarization 265 in vitro.
To confirm our results genetically, we generated two independent conditional Hh

Hh signaling controls Th17 polarization in vivo. 290
Th17 cells are critical drivers of inflammatory bowel disease including pathological inflammation of the small intestine. A model of small intestinal inflammation driven by Th17 cells was developed by the Flavell lab 16 . Here, the intraperitoneal injection of anti-CD3 antibodies leads to polarization and accumulation of Th17 cells in the small intestine within 3 days. Using this model system we treated wildtype C57BL/6 mice by oral gavage 295 with either carrier or Hh inhibitor vismodegib, which has been clinically approved for oral use (Fig. 5A). Strikingly, intestinal inflammation was ameliorated in vismodegib-treated mice. Mice, in which Hh signaling had been inhibited, showed reduced weight loss ( Fig.   5B) and less thickening and shortening (lower Weight/Length ratio) of the small intestine ( Fig. 5C). Importantly, vismodegib-treated mice had significantly fewer Th17 300 intraepithelial lymphocytes (IELs) while the number of IFNg + IL17 neg IELs was less affected (Fig. 5D). This was accompanied by a 50% reduction of IL-17a levels in the serum (Fig. 5E).
In order to prove genetically that Hh signaling is instrumental to induce Th17-mediated  Fig. 9). 320 Taken together, we demonstrate for the first time that intracellular Hh signaling is a critical driver of Th17 polarization in chronic intestinal inflammation models of human inflammatory bowel disease (IBD). Importantly, we show that the small molecule Hh inhibitor vismodegib, which is clinically approved by the EMA/FDA, effectively protected 325 against IBD.   Having established that Hh signaling is functionally important for Th17 polarization, we sought to determine the mechanism for this effect. Downstream of the IL-6 and IL-23 355 cytokine receptors, Th17 differentiation is initiated by STAT3 phosphorylation which in turn translocates to the nucleus and initiates the transcription of key Th17 polarizing transcription factors (TFs). We found that phosphoSTAT3 (pSTAT3) levels are unaltered upon IL-6 stimulation when we inhibit the Hh pathway with cyclopamine in Th17 cells, indicating that Th17-initiating cytokine signaling is not altered by the Hh pathway (Fig.  360 6A).
Next, we profiled the expression of master Th17 TFs 18 in Th17 cells treated with Hh inhibitor cyclopamine or carrier control (Fig. 6B). Although Il17a mRNA was markedly reduced in our treatment conditions, the Th17-regulating TFs Rorct, Rora, Irf4, Runx1, and Batf were not majorly affected and neither was a novel Th17-associated TF Vax2 19 . 365 To interrogate the mechanism of Hh-mediated Th17 polarization more globally we performed RNA-Seq analysis of Th17 cells polarized in the presence of carrier control or Hh inhibitor cyclopamine. As expected, expression of known Hh target genes 20 like Smo, Jag2, Prdm1, and Fst was significantly reduced in Hh inhibitor-treated cells (Fig. 6C). The expression level of master Th17 TFs in Hh inhibitor-treated Th17 cells was ambiguous: 370 while for example Rora and Batf were significantly reduced upon inhibitor treatment, Junb and Runx1 were increased (Fig. 6C). Th17 cells have stem-cell like features: plasticity and self-renewal 21 . Interestingly, we find that Tcf7, the key stem cell-associated gene in Th17 cells 22 , as well as other Wnt target genes were downregulated in Hh inhibitortreated cells (Fig. 6C) suggesting that Hh signaling may be critical to endow Th17 cells 375 with stem cell properties.
Overall, GSEA (Gene Set Enrichment Analysis) certified that Hh inhibitor-treated cells moved away from a Th17 transcriptional profile (Fig. 6D, Suppl. Fig. 10A). However, we show that this effect of Hh signaling blockade is not due to any effects on proximal cytokine signaling or expression levels of key Th17 TFs. 380

Gli3 is a novel Th17-promoting transcription factor that acts in concert with noncanonical Hh signaling via pAMPK in Th17 polarization.
Smo is the key signal transducer of the Hh pathway and can function in both a canonical and non-canonical fashion (Fig. 7A). Smo activates Gli transcription factors as part of the 400 canonical pathway, but has also been shown to activate AMPK 23 or act through its GTPase activity when associated with non-canonical Hh signaling 24 .
Three transcription factors are associated with Hh signaling, but only Gli1 and Gli3 are expressed in CD4 + T cells (Fig. 1E, Suppl. Fig. 2). To dissect their relevance, we first investigated the role of Gli1 in CD4 + T cell polarization. We isolated naïve CD4 + T cells 405 from Gli1 knockout mice as well as heterozygous and wildtype littermates (validation, Suppl. Fig. 6C, Fig. 7B left panel) and polarized the cells into Th0, Th1, Th2, Th17, and iTreg cells. Surprisingly, no differences in Th polarization were observed between Gli1 WT, HET, and KO mice showing that Gli1 is functionally not important for Th17 differentiation (Fig. 7B right panel). The balance between GliR and GliA in the nucleus shapes the Hh response. We found that upon Hh inhibitor treatment the majority of the predicted Gli3 target genes increase or decrease in a dose-dependent manner (Fig. 7D, Suppl. Fig. 10B)

indicating that indeed 430
Hh-induced Gli3 regulates the transcription of these target genes to control Th17 polarisation.
In order to interrogate the functional relevance of Gli3 we used CRISPR to delete Gli3 in primary CD4 + T cells and achieved a 52% reduction of Gli3 RNA by day 3 (Fig. 7E).
Importantly, knockout of Gli3 in a proportion of primary Th17 cells led to a significant 435 reduction in IL-17a producing cells. Taken together, we have shown that Gli3 is the only Hh transcription factor that is expressed and functionally important in Th17 cells.  (Fig. 7F). Furthermore, the expression of the upstream AMPK regulating kinases CaMKK2 and LKB1 was also reduced as had been previously described in adipose tissue 23 .
Another non-canonical signaling mode of Smo is its direct function as a GPCR 24 , which is highly sensitive to pertussis toxin inhibition. To investigate whether this signaling mode 450 is important for Th17 polarization we treated naïve CD4 + T cells for 5 days with pertussis toxin under Th17 polarizing conditions. Pertussis toxin treatment did not affect viability or proliferation and had no effect on IL-17a production (Fig. 7G).
Taken together, we have uncovered Gli3 as a fundamental new TF implicated in Th17 polarization that works together with non-canonical Hh signaling via pAMPK. 455

DISCUSSION 485
Our work is the first to describe that Hh signaling selectively controls Th17 polarization, but not polarization into Th1, Th2, and iTreg lineages. We show that the critical Hh  We identify Gli3 as a novel transcription factor that controls Th17 lineage polarization.
Our observation functionally validates an ATACSeq-based bioinformatic model that suggested Gli3 is part of the top 30 "core" TFs of the Th17 transcriptional regulatory network 19 . We show that Hh-induced Gli3 regulates the transcription of previously 505 predicted Gli3 target genes (Suppl. Fig. 10) known to control Th17 polarisation 19 . This crucial role for Gli3 seems to be restricted to CD4 + T cells, since Gli3 expression has not been reported in CD8 + T cells 11 . Furthermore, we show that Hh signaling influences the metabolic regulation of Th17 cells through a known non-canonical Hh signaling axis via CaMKK2/LKB1 and pAMPK in 510 adipose tissue 23 . This is in line with published data indicating a role of pAMPK in human and murine Th17 polarization 27,28 and in the pathogenicity of Th17 cells in a model of adoptive T cell transfer colitis 27 .
Our model is summarized in Suppl. Fig. 11. Hh signaling has been implicated in CD4 + T cell polarization of Th2 but not Th1 cells 12 .
Furmanski et al used mice expressing transgenic Gli2A (activator) or Gli2R (repressor) and showed an effect of exogenous Shh on Th2 differentiation. Our findings that Gli2 transcripts are absent in murine Th cultures and that Th17 cells do not respond to exogenous Ihh ligand might resolve these differences. In addition, it has been shown that 520 exogenous recombinant Shh is unable to induce Th17 polarization in human CD4 + T cells 30  Th17 cells with stem cell potential in addition to being crucial for Th17 polarization.
Intriguingly, we see that cells treated with Hh inhibitors lose expression of stem-cellassociated Wnt target genes Lef1, Tcf4, Tcf7l1, Ascl2, as well as Tcf7 which was identified as the main marker of Th17 stemness (Fig. 6C) 22 .

535
Due to the central role that pathogenic Th17 cells play in numerous autoimmune diseases, including inflammatory bowel disease, rheumatoid arthritis, multiple sclerosis and psoriasis 37 , there has been great interest in developing therapeutic interventions that specifically target Th17 polarization and pathogenicity. The importance of Th17 cells in COVID-19 pathogenesis has further increased this interest 38,39 . 540 There is a shortage of small molecule inhibitors that target Th17 polarization and effector function. A seminal drug screen for Rorgt antagonists identified digoxin as a potent inhibitor able to suppress Th17 responses in vitro and in vivo 40 . More recently, bromodomain inhibitors JQ1 and MS402 were found to block Th17 differentiation in preclinical models of intestinal inflammation 41,42 . However, the extremely narrow therapeutic 545 window of digoxin and the broad-ranging effects of targeting epigenetic regulators with bromodomain inhibitors make their use in the clinic difficult. Here, we present data showing that clinically approved Hh inhibitor vismodegib specifically and potently inhibits Th17 polarization in vivo, opening up the possibility of treating autoimmune diseases with Hh inhibitors. 550

MATERIALS AND METHODS 555
Mice RAG2KO were a generous gift from Suzanne Turner (University of Cambridge) and OTI mice were purchased from the Jackson Laboratory (C57BL/6-Tg(TcraTcrb)1100Mjb/j, Stock no. 003831). OTI RAG2KO mice were generated from these. Gli1-eGFP mice were 560 a generous gift from Alexandra Joyner (Sloan Kettering Institute) 43   CD4 + T cells were plated onto a 96-well plate pre-coated overnight at 4°C with 2.5 µg/ml anti-CD3e and 2 µg/ml anti-CD28 antibody (

CRISPR of primary naïve CD4 + T cells
Alt-R ® CRISPR-Cas9 crRNA (IDT) and tracrRNA (IDT, cat no. 1072534) were reconstituted at 100 µM in nuclease-free duplex buffer (IDT). 1.9 µl of each stock solution was mixed with 8.7 µl duplex buffer. Samples were heated in a thermal cycler at 95°C for 5 min and left for 10 min at room temperature. 10.5 µl Buffer T (Thermo, cat no. 615 MPK10096) was added as well as 2 µl TrueCut™ Cas9 Protein v2 (5mg/ml, Thermo, cat no. A36499) which was pipetted very slowly in a circular motion to ensure optimal solubility. The mixture was incubated at 37°C for 10 min to assemble the ribonucleoprotein (RNP) complex.
Naïve CD4+ T cells were isolated from C57BL/6 spleens and stimulated with plate bound 620 anti-CD3e/CD28 antibodies in the presence of Th17 polarizing cytokines as described previously. After 24 h of stimulation, cells were washed twice in pre-warmed PBS (Gibco) prior to resuspension in 80 µl Buffer T (1 million cells/electroporation reaction). The suspension was briefly mixed with the RNP complex solution prior to electroporation with the Neon™ electroporation system in 100 µl electroporation tips (Thermo, cat no. 625 MPK10096) with three pulses of 1600 V each with a pulse width of 10 ms. Cells were left to recover in complete IMDM in the absence of antibiotics for 20 min. Cells were then centrifuged at 1500 rpm for 5 min and returned into culture to an anti-CD3ε/CD28 antibody coated plate in complete IMDM (without antibiotics) supplemented with polarizing cytokines at the concentrations mentioned previously. Cells were centrifuged at 1000 rpm 630 for 1 min for quick adherence to the coated plate. Antibiotics were re-added after six hours at the indicated concentrations. TruStain fcX anti-mouse CD16/32, cat no. 101320) for 5 min at room temperature 655 protected from light. Next, cells were incubated with 50 μl of fluorophore-conjugated antibodies at the appropriate dilution (Table 1) for 20 min at 4 o C protected from light. Cells were washed twice with FACS buffer and moved to 5 ml polystyrene round bottom tubes prior to immediate analysis (Fisher Scientific/Falcon) or fixation for intracellular staining.

Intracellular staining 660
For the staining of intracellular cytokines, cells were incubated at 37°C for 4h in the presence of 1 μg/ml Ionomycin (Sigma, cat no. I9657) and 50 ng/ml PMA (Sigma, cat no.  (Table 1)      Taqman probe (Thermo Fisher, see

Image acquisition and analysis 765
Confocal spinning disc microscopy was performed on an Andor Dragonfly 500 (Oxford Instruments). Images were processed using Imaris software (Bitplane/Oxford Instruments).

Western Blot 770
Cells were harvested at 4°C, washed twice in ice-cold PBS and lysed in ice-cold RIPA

RNA Sequencing 795
Samples were generated and RNA was extracted as previously described. Six biological replicates were used per condition. RNA quality was assessed using a capillary electrophoresis system (4200 Tapestation, Agilent) using RNA ScreenTape (Agilent, cat no. 5067-5576) as per the manufacturer's instructions. RNA concentrations were quantified using the Qubit™ RNA BR Assay kit (Thermo, cat no. 10210) as instructed by 800 the manufacturer. Libraries were generated using the TruSeq stranded mRNA kit (Illumina) following the manufacturer's instructions and sequenced using single-read sequencing with the HiSeq4000 platform (Illumina).
Reads were aligned to the mouse genome version GRCm38 using STAR v2.5.3a 45 . Read 805 counts were obtained using feature Counts function in Subread v1.5.267 46 and read counts were normalized and tested for differential gene expression using the DESeq2 workflow 47 . Multiple testing correction was applied using the Benjamini-Hochberg method.

CD3 injection model of small intestinal inflammation
8-10-week-old female C57BL/6 mice were injected with 20µg anti-CD3 monoclonal antibody (Clone: 145-2C11) at 0h and 48h as previously described 16 . Mice were dosed every 12h by oral gavage with 100mg/kg vismodegib (LC laboratories), prepared in MCT 820 from a fresh vial for each experiment (0.5% methylcellulose, 0.2% Tween 80), or carrier control, with four to five mice per group. Mice were harvested at 52h, at which point serum was collected and small intestines were harvested in ice-cold PBS. Small intestines were cleaned using ice-cold PBS, cut into roughly 5 mm pieces and collected in complete IMDM. Tissue pieces were rotated for 30 min at room temperature to release 825 intraepithelial lymphocytes (IELs). Cells were strained on a 70 µm filter and IELs were collected from the interface of a 40%-80% Percoll gradient (GE Healthcare, cat no.

835
Adoptive T cell transfer colitis model CD4 + T cells were isolated from spleen and peripheral lymph nodes of donor mice as described previously using MACS isolation. The CD4 + T cells were stained in MACS Buffer for 15min at 4°C for CD4, CD25, CD45RB and DAPI was added prior to FACS 840 sorting which was performed as described above. Pure CD4 + CD25 -CD45RB hi tdTomato + cells were sorted to high purity. Cells were washed twice in PBS prior to intraperitoneal injection into Rag2 -/recipient mice. A cell suspension of 200 μl at 2x10 6 cells/ml PBS were injected intraperitoneally per mouse. Mice were weighed at least twice weekly thereafter and harvested between 5-6 weeks after injection. 845 Colons, mesenteric lymph nodes and spleens were harvested in sterile ice-cold PBS.
Colons were flushed through with ice-cold PBS to remove faeces and a 5-10 mm representative sample was collected in 10% Neutral Buffered Formalin (NBF). These samples were incubated overnight in NBF prior to being moved to a 70% Ethanol solution 850 for H&E staining (in-house histopathology core). After collection of histology samples, the colon was cut into roughly 5mm pieces collected in a 50 ml falcon with complete IMDM + 58 µg/ml DNAse (Stem Cell Technologies, cat no. 7469) + 58 µg/ml Liberase, Thermolysin low (Roche, cat no. 5401020001) and incubated in a shaker at 37°C for 25 min at 225 rpm to release colonic lamina propria lymphocytes (LPLs). Cells were strained 855 twice through a 70 µm filter and LPLs were collected from the interface of a 40%-80% Percoll gradient (GE Healthcare, cat no. 17089101). Spleens were mashed through a 70 µm filter with sterile PBS and collected from the interface of a Lympholyte gradient (Cedarlane Laboratories, cat no. CL5035). Gradient centrifugations were set up in 15 ml falcons and centrifuged for 1800 rpm for 20 min (acceleration 1/9, deceleration 0/9). 860 Mesenteric lymph nodes were strained twice on a 40 µm filter.
LPLs, splenocytes and lymph node suspensions were restimulated for 4 h with PMA and Ionomycin in the presence of Monensin and subjected to viability, surface and intracellular flow cytometric staining (for details see section on flow cytometric methods). Statistical 865 analysis was performed using Prism 7 software (GraphPad Inc.). Statistical analysis was performed using an unpaired two-tailed Student's t test or Mann Whitney U test, respectively. For each data set, the presence of a statistically significant outlier was assessed using Grubbs' test (α = 0.05) and where present was excluded from subsequent analysis. 870 Semi-quantitative analysis of colitis severity was determined on formalin-fixed, paraffinembedded and hematoxylin & eosin stained sections as similarly described 49 . Sections were analyzed in a blinded fashion by an independent expert. The sum of each subscore of mononuclear infiltration (0-3), crypt hyperplasia (0-3), epithelial injury (0-3), neutrophil 875 infiltration (0-3) and inflammatory penetration (0-2) depicted colitis severity.