Alpha kinase 1 controls intestinal inflammation by suppressing the IL-12/Th1 axis

Inflammatory bowel disease (IBD) are heterogenous disorders of the gastrointestinal tract caused by a spectrum of genetic and environmental factors. In mice, overlapping regions of chromosome 3 have been associated with susceptibility to IBD-like pathology, including a locus called Hiccs. However, the specific gene that controls disease susceptibility remains unknown. Here we identify a Hiccs locus gene, Alpk1 (encoding alpha kinase 1), as a potent regulator of intestinal inflammation. In response to infection with the commensal pathobiont Helicobacter hepaticus (Hh), Alpk1-deficient mice display exacerbated interleukin (IL)-12/IL-23 dependent colitis characterized by an enhanced Th1/interferon(IFN)-γ response. Alpk1 controls intestinal immunity via the hematopoietic system and is highly expressed by mononuclear phagocytes. In response to Hh, Alpk1−/− macrophages produce abnormally high amounts of IL-12, but not IL-23. This study demonstrates that Alpk1 promotes intestinal homoeostasis by regulating the balance of type 1/type 17 immunity following microbial challenge.

I nflammatory bowel disease (IBD) pathogenesis is mechanistically complex and includes elements of genetic susceptibility, immune dysregulation, environmental factors, and the microbiome. As with humans, colitis in mice is strongly influenced by host genetics, such that different inbred strains exhibit widely divergent phenotypes in models of IBD 1,2 . For example, whereas 129SvEv.Rag2 −/− mice develop aggressive colitis following infection with Hh, C57BL/6.Rag1 −/− mice do not 3 . 129SvEv mice deficient for the Wiskott-Aldrich syndrome protein (Was −/− ) are also susceptible to spontaneous colitis 4 .
In recent years, two loci on chromosome 3-termed "cytokinedependent colitis susceptibility locus" (Cdcs1) in C3H/HeJBir mice, and "Helicobacter hepaticus-induced colitis and associated cancer susceptibility" (Hiccs) in 129SvEv mice -have been identified as the critical genetic determinants of colitis susceptibility in these strains 2,3,5-8 . The Cdcs1 locus was first identified in the context of interleukin-10 deficient (Il10 −/− ) mice over 10 years ago 5,8 , and also controls spontaneous colitis in C57BL/6. Tbx21 −/− Rag2 −/− (TRUC) mice 9 . The Hiccs locus, located in a similar region of chromosome 3, controls susceptibility to Hhinduced colitis and eventual onset of colitis-associated colorectal cancer 3 . The ability of the Cdcs1 and Hiccs loci to control colitis susceptibility in several mechanistically distinct models suggests that they include one or more critical immunoregulatory genes. However, individual Cdcs1/Hiccs-related genes have not been directly studied in the context of intestinal inflammation.
A Hiccs locus gene, Alpha kinase 1 (Alpk1), has multiple nonsynonymous single nucleotide differences in its coding region between the colitis-susceptible 129SvEv and the resistant C57BL/6 mouse strains 3 . We previously showed that Alpk1 expression is upregulated in response to inflammatory stimuli in myeloid cells 3 . Also, single nucleotide polymorphisms (SNPs) in the human ALPK1 gene have been linked to a variety of inflammatory conditions, including gout and chronic kidney disease 10,11 . More recently, in vitro studies suggested that Alpk1 mediates pathogen-induced IL-8 expression in gastric epithelial cells 12,13 , making it a relevant gene to explore in the context of gut inflammation.
In this study, we further refine the Cdcs1 locus and identify a core colitis-determining region that is essentially identical to the Hiccs locus. To address a potential role for Alpk1 in regulation of intestinal homoeostasis, we have generated Alpk1-deficient mice. We show that Alpk1 deficiency leads to severe colitis and an exaggerated Th1 immune response in mice infected with the intestinal pathobiont Helicobacter hepaticus. We further demonstrate that Alpk1 exerts its anti-inflammatory function within the hematopoietic compartment, in which it restrains production of IL-12 by myeloid cells in response to Helicobacter challenge.

Results
A genetic locus controlling colitis susceptibility in mice. In addition to the Il10 −/− and TRUC models, we now show that the risk-conferring genotype of Cdcs1 also confers susceptibility to spontaneous colitis in C57BL/6.Was −/− mice ( Supplementary  Fig. 1a-c). In both Was −/− and TRUC mice, colitis susceptibility requires homozygosity for the C3H-derived Cdcs1 allele in hematopoietic cells ( Supplementary Fig. 1d-e) 9 . Previously, we mapped the critical region of the Hiccs locus to a 1.71-Mb interval that contains five microRNAs and eight protein-coding genes 3 . We have similarly fine-mapped the susceptibility-controlling region of Cdcs1 in TRUC mice to a congenic interval flanked by the genetic markers D3Mit348 and D3Mit319 ( Supplementary Fig. 2). This Cdcs1 core region is essentially identical with Hiccs (Fig. 1a). Therefore, we hypothesized that this locus contains a previously unidentified gene that controls colitis susceptibility in multiple mouse models of IBD.
The role of Alpk1 in a lymphocyte-replete model of colitis. Having found a critical role for Alpk1 in regulating inflammation in a Rag1-deficient setting, we next assessed its impact on lymphocyte-replete colitis. Hh infection does not induce colitis in wild-type B6 mice unless IL-10 signalling is blocked 22 . We therefore treated Alpk1 +/− and Alpk1 −/− mice with Hh and an IL-10R (IL-10 receptor alpha) blocking antibody to induce colitis (Fig. 3a). In agreement with results obtained using Rag1 −/− mice, Alpk1 −/− animals developed exacerbated intestinal inflammation based on histological scoring (Fig. 3b, c), colonoscopy (Fig. 3d), and total CD45 + lamina propria leucocytes (Fig. 3e) Fig. 4b).
The increased disease burden in Alpk1 −/− mice was not due to increased intestinal colonization by Hh ( Supplementary Fig. 4c). The Hh + anti-IL10R protocol normally elicits a CD4 + helper T-cell response with mixed Th1 and Th17 characteristics. However, compared to WT or heterozygous controls, intestinal CD4 + T cells of Alpk1 −/− mice were highly skewed towards a Th1 phenotype (IFN-γ + IL-17A -) following induction of colitis, as measured by cell frequency and IFN-γ expression per cell (Fig. 3f, g, Supplementary Fig. 4d, e). In keeping with their similar degree of colitis, wild type and Alpk1 +/− mice showed comparable T cell phenotypes; only Alpk1 −/− mice developed a polarized Th1 response ( Supplementary Fig. 4d). Although IFN-γ was largely coexpressed with TNF-α, frequencies of TNF-α expression were not significantly elevated in Alpk1 −/− T cells ( Supplementary Fig. 4f). In contrast, the frequency of Th17 cells (IFN-γ -IL-17A + ) and their intensity of IL-17A expression was diminished in Alpk1 −/− mice ( Fig. 3f, g, Supplementary Fig. 4e). IL-22 producing CD4 + T cells were also less frequent in Alpk1 −/− mice ( Supplementary  Fig. 4g). Boolean gating analysis of cytokine-producing CD4 + T cells confirmed a profound shift in effector phenotype in the absence of Alpk1, with IFN-γ + IL-17A − cells vastly outnumbering other populations ( Supplementary Fig. 4h). Although the total number of colonic FOXP3 + regulatory T cells (Treg) was comparable between Alpk1 −/− and control mice, their frequency among CD4 + T cells was reduced in Alpk1 −/− mice, particularly the FOXP3 + RORγt + subset that is induced in response to Hh (Supplementary Fig. 4i) 23 . While we observed a similarly Th1skewed phenotype in the mesenteric lymph nodes of Alpk1 −/− animals, this was not apparent in their peripheral lymph nodes (inguinal/axillary) or spleens ( Supplementary Fig. 4j), suggesting that Alpk1 exerts local control over Th1 immunity.
Gene expression analysis of whole-colon tissue confirmed elevated amounts of pro-inflammatory and Th1 cytokines in colitic Alpk1 −/− mice, while expression of Th2 cytokines and IL-10 were reduced relative to WT/heterozygous animals ( Supplementary Fig. 4k). Interestingly, Alpk1 −/− mice displayed a robust intestinal Th1 response and reduced frequency of FOXP3 + RORγt + Treg following Hh infection in the absence of IL-10R blockade ( Fig. 3h-k). This alteration in Th1/Treg balance did not cause marked pathology or recruitment of myeloid leucocytes to the colon, however (Fig. 3h, Supplementary Fig. 4l). In contrast, Alpk1 −/− mice treated with anti-IL-10R antibody alone did not show any signs of pathology or activation of a Th1 response (Fig. 3h, i, k). Therefore, Alpk1 deficiency augments the IL-12/Th1 axis and reduces the Treg/Th1 ratio following Hh infection, causing mild immunopathology that is exacerbated in the absence of IL-10 signalling.
Alpk1 controls colitis severity via hematopoietic cells. Analysis of the Immgen database 24 and FACS-sorted intestinal populations revealed a broad spectrum of Alpk1-expressing cell types, including antigen presenting cells (APCs), epithelial cells, and intestinal stromal cells ( Supplementary Fig. 5). To determine whether Alpk1 impacts intestinal inflammation via the hematopoietic compartment, we conducted reciprocal bone marrow chimera experiments. Consistent with prior observations of the Hiccs and Cdcs1 loci 3 (Supplementary Fig. 1-2), irradiated WT mice reconstituted with Alpk1 −/− bone marrow developed more severe Hh + anti-IL-10R-induced colitis in comparison to those that received WT bone marrow (Fig. 4a-c). Conversely, irradiated WT or Alpk1 −/− animals reconstituted with wild-type bone marrow developed comparable levels of colitis ( Fig. 4d-f), demonstrating that Alpk1 regulates colitis via the hematopoietic compartment.
Because CD4 + T-cell responses are largely programmed by APCs, we employed the naive T-cell transfer model of colitis (i.e., adoptive transfer of naive CD4 + T cells to Rag1 −/− hosts) to test whether, in the absence of acute bacterial challenge, deficiency of Alpk1 in APCs could influence the phenotype of expanding T cells. However, transfer of WT T cells to Alpk1 −/− Rag1 −/− or Alpk1 +/− Rag1 −/− hosts resulted in equivalent T-cell differentiation and severity of colitis (Supplementary Fig. 6a-d). Similarly, no differences were observed between Alpk1 +/+ Rag1 −/− recipients following transfer of Alpk1 −/− or WT T cells (Supplementary Fig. 6e-h). Thus, Alpk1 deficiency in the absence of a strong microbial driver does not appear to significantly impact CD4 + Tcell differentiation or susceptibility to T-cell-driven colitis. We also tested the role of Alpk1 in another model of colitis that bypasses bacterial stimuli, in which treatment of Rag1 −/− mice with an agonistic anti-CD40 antibody directly activates APCs and induces IL-12/IL-23-mediated pathology 21 . Alpk1 +/− Rag1 −/− and Alpk1 −/− Rag1 −/− mice developed comparable levels of inflammation following anti-CD40 challenge (Supplementary Fig. 6i-n), providing further evidence that Alpk1 is functionally linked to microbial sensing.
Alpk1 regulates Hh-driven IL-12 production by phagocytes. Alpk1 has been proposed to affect several distinct cellular pathways 13,[25][26][27][28][29] , yet none explain the pathological effect of its deficiency in the gut. However, our in vivo data suggest that the defect may be related to bacterially induced IL-12 production. Because APCs are the major producers of IL-12, we generated bone marrow-derived macrophages (BMDMs) via culture of bone marrow with GM-CSF (granulocyte-macrophage colony stimulating factor) and examined their response to Hh (Fig. 5a). After 8 days of culture, flow cytometry analysis revealed no clear differences in differentiation between WT and Alpk1 −/− cells ( Supplementary Fig. 7a-c). Compared to WT cells, Alpk1 −/− BMDMs secreted more IL-12 following exposure to Hh, whereas IL-23 production was similar (Fig. 5b). At the mRNA level, expression of Il12a and Il12b, but not Il23a, was significantly increased in Alpk1 −/− BMDMs (Fig. 5c). IL-12 secretion by Hh was blocked in the presence of anti-TLR2 antibody (Fig. 5b), suggesting that Alpk1 mediates signal transduction downstream of TLR2 and upstream of Il12a/Il12b transcription. Interestingly, pure TLR2 and TLR4 agonists, such as Pam3CSK4 and LPS-induced comparable levels of IL-12 in both wild type and Alpk1 −/− cells (Fig. 5b).
Transcriptomic analysis of Alpk1-regulated genes. To examine the role of Alpk1 in an unbiased manner, we compared whole transcriptomes of Hh-treated wild type and Alpk1 −/− BMDMs using RNA sequencing. Under resting conditions, the transcriptomes of Alpk1 −/− and Alpk1 +/− cells were almost indistinguishable by tSNE (t-stochastic neighbour embedding) analysis ( Supplementary Fig. 8a), with 32 genes (including Alpk1 itself) Steady state Alpk1 Alpk1 showing a statistically significant difference, and only 9 > 2-fold ( Supplementary Fig. 8b, Supplementary Data 2). This confirms that Alpk1 did not dramatically affect the differentiation of BMDMs in vitro. In contrast, 599 genes were significantly differentially expressed between the genotypes after Hh stimulation (adjusted p-value <0.05; 67 genes with >2-fold change), including Il12a and Il12b (Fig. 5d, Supplementary Fig. 8b, Supplementary Data 2). Pathway enrichment and clustering analyses of the top upregulated genes in Hh-treated Alpk1 −/− cells revealed enrichment of processes related to IL-12 signalling, regulatory interactions between immune cells (e.g., components of CD200R and CD300 signalling pathways), chemokine signalling, and the extracellular matrix (Fig. 5e, Supplementary Fig. 8c, Supplementary Data 2). Expression of selected genes at the mRNA level was confirmed using qPCR (Fig. 5f).

Steady state Alpk1
Precisely how Hh triggers TLR2 and downstream IL-12 production in mouse BMDMs is unclear. We found that classical TLR and C-type lectin receptor (CLR) agonists, including LPS, CpG, zymozan, and heat-killed mycobacteria, induced comparable IL-12 expression in wild type and Alpk1 −/− BMDMs ( Supplementary Fig. 8d). Assessment of key signalling molecules in the TLR2 pathway, including MAPK kinases and NF-κB revealed no differences between wild type and Alpk1 −/− cells after Hh stimulation (Supplementary Fig. 8e).
Alpk1 expression in human inflammatory bowel disease. We previously reported that Alpk1 expression is detectable in murine macrophages under steady-state conditions, but can be further induced by stimulation with LPS or Hh 3 . IFN-γ stimulation

Il12a
Il12b Ifng   Fig. 8f). Consistent with these findings, ALPK1 mRNA is highly expressed in the inflamed intestinal mucosa of patients with IBD relative to tissue from healthy control donors in three independent cohorts (Fig. 6a). Notably, ALPK1 expression in IBD tissue was closely correlated with that of Th1-related cytokines (IFNG, IL12A, IL12B, and CXCL10) but not Th2 or Th17 cytokines (IL4, IL13, and IL17A) (Fig. 6b). Thus, in addition to regulation by pattern recognition receptors, Alpk1 expression may be induced by IFN-γ signalling as part of a negative feedback mechanism to limit the intensity of IL-12 production and downstream Th1 immunity.

Discussion
In this study, we have identified Alpk1 as a negative regulator of intestinal inflammation. We demonstrate that Alpk1-deficient mice infected with a pathobiont Helicobacter hepaticus (Hh) develop an unusually potent Th1 CD4 + T-cell response and exacerbated colitis in the absence of IL-10 signalling. Unlike IFNγ-producing Th1 cells, the frequencies of FOXP3 + regulatory and IL-17-producing (Th17) CD4 + T cells are decreased in these animals compared to the wild-type controls. Using bone marrow chimera experiments, we also establish that Alpk1 operates in hematopoietic cells to regulate Hh-induced pathology and the Th1 response. We show that Alpk1 acts as a checkpoint specifically limiting IL-12 production in mononuclear phagocytes, and that this occurs independently of the classical anti-inflammatory cytokine IL-10.
Our results suggest distinct immunoregulatory roles for Alpk1 and IL-10 in control of Hh driven inflammation (Fig. 3). Although Hh infection of Alpk1 −/− mice resulted in marked skewing of T cell differentiation towards Th1 cells in the colon compared to wild-type mice this was not sufficient to result in severe colitis. By contrast, combined Hh infection and IL-10R blockade caused severe disease with high numbers of colonic Th1 and myeloid cells in Alpk1-deficient mice, as compared to wildtype littermates. These results suggest that Alpk1 and IL-10 represent complimentary checkpoints in the Th1 inflammatory response. Alpk1 may be a critical modulator of CD4 + T cell differentiation through its suppressive effects on IL-12 production, whereas IL-10 is more important for limiting the magnitude of Hh-driven inflammation through controlling the myeloid response 30 .
We have shown that Hh stimulation induces elevated IL-12 expression in Alpk1 −/− BMDMs, which is blocked by a neutralising TLR2 antibody (Fig. 5b). However, stimulation of BMDMs with a TLR2 ligand Pam3CSK4 or other purified pattern recognition receptor ligands induced comparable IL-12 expression in wild type and Alpk1 −/− deficient cells. We speculate that there may be a second, unknown factor under the control of Alpk1 that partners with TLR2 to mediate Hh recognition. Another possibility is that Alpk1 regulates phagocytic clearance of Hh and signalling pathways linked to this process. Interestingly, Alpk1 was initially shown to regulate apical transport in epithelial cells via interaction with myosin Ia 26 . It is possible that a similar machinery is recruited to aid in Hh processing by phagocytes. Deciphering the Alpk1 interactome will be required to investigate this further.
Contrary to the data obtained from in vitro cultured human gastric epithelial cells, in which Alpk1 is essential for IL-8 expression induced by certain bacterial pathogens 12,13 , we show that Alpk1 deficiency resulted in elevated cytokine expression in BMDMs treated with Hh. This may be explained by a possible cell-type-specific function of Alpk1, as observed for some other inflammatory regulators. For instance, IL-18 signalling in intestinal epithelial cells and lamina propria leucocytes has seemingly opposite effects on intestinal inflammation 31 . Indeed, our bone marrow chimera studies indicate that Alpk1 expression by hematopoietic, but not epithelial cells, is important for control of Hh-induced colitis in vivo (Fig. 4). Consistent with this, in vitro analysis showed that Alpk1 can function in macrophages to control Hh-driven IL-12 production. However cell type-specific deletion of Alpk1 will be required to address the context-specific function of Alpk1 in various cell types.
Our results identify Alpk1 as the first gene in the Hiccs/Cdcs1 locus with a definitive role in regulating intestinal inflammation. While our data collectively imply that the risk-conferring alleles of Hiccs/Cdcs1 contain a hypomorphic variant of Alpk1, generation of mice with specific point-mutations in Alpk1 will be required to assess Alpk1 as the causative gene in the Hiccs/Cdcs1 locus. The same genomic interval contains Tifa (TRAF Interacting Protein With Forkhead Associated Domain), which encodes an adaptor protein involved in TNFR/TLR and NLRP3   inflammasome signalling 32,33 . We did not focus on Tifa in our studies as there are no non-synonymous nucleotide differences in the Tifa coding region and Tifa mRNA expression levels are similar in myeloid cells between the colitis susceptible and resistant mouse strains 3 . However, TIFA has been recently implicated in sensing of the bacterial metabolite HBP by human epithelial cells, a process that also requires Alpk1 12,13. Therefore, involvement of TIFA in Alpk1-dependent pathways in the gut cannot be excluded. Considering the distinct roles of Alpk1 in human epithelial and mouse myeloid cells, generation of TIFA knockout mice is essential to explore this axis further. SNPs in the ALPK1 gene have been shown to be associated with inflammatory disease risk in humans 10,11 , yet no link has been established between ALPK1 and inflammatory bowel disease. Taking this into account, we examined ALPK1 expression in tissue biopsies from patients with IBD and observed elevated expression compared to healthy individuals. Moreover, Alpk1 expression correlated with the expression of Th1 but not Th2 axis genes in IBD (Fig. 6). These data are in line with our findings that Alpk1 expression in BMDMs is induced by both TLR stimulation 3 and IFN-γ, and therefore may represent a negative feedback loop that constrains intestinal Th1 responses. This concept is consistent with the well-known induction of the negative regulator IL-10 by TLR activation, and elevated IL-10 expression found in inflamed tissues. These human findings necessitate additional work to determine the mechanisms that regulate Alpk1 expression, as well as the spatiotemporal pattern of intestinal Alpk1 expression in healthy, inflamed and dysbiotic settings.
In summary, we have identified Alpk1 as a critical determinant of murine colitis susceptibility encoded by the Hiccs and Cdcs1 loci. By selectively regulating production of IL-12 downstream of bacterial stimulation, Alpk1 restricts the activation of Th1 immunity in favour of Treg and Th17 responses that maintain homoeostasis. Intriguingly, Alpk1 and IL-10 appear to regulate intestinal homoeostasis via distinct mechanisms, in that Alpk1 controls the nature of intestinal immune responses, while IL-10 controls their magnitude. Thus, loss of Alpk1 function permits aberrant Th1 immunity that, in combination with IL-10 deficiency, causes aggressive and highly destructive colitis (Fig. 7). How Alpk1 influences human pathology, and the precise biochemical mechanism by which it regulates expression of inflammatory factors, are key outstanding questions that should be addressed in future studies. While this paper was in press, data showing a functional role for Alpk1 in anti-microbial responses were published 34 .

Methods
CRISPR-assisted generation of ALPK1 knockout mice. Using the web based tool 35 four sgRNA, two sgRNA flanking either side of exon 10 of mouse Alpk1 gene (ENSMUSE00001278253) were identified. Deletion of exon 10 generates a frameshift mutation upstream of the alpha-type protein kinase domain. The guide sequences (sgRNAs) were ordered from Sigma Genosys as sense and antisense oligonucleotides and annealed before individually cloning into the T7 expression vector (kind gift from Sebastian Gerety). Following plasmid linearization, RNA was transcribed using a MEGAshortscript T7 Transcription Kit (Thermo Fisher Scientific) and column purified (RNeasy, Qiagen). The humanised Cas9 protein developed by Mali et al. 36 was modified by replacing the CMV promotor with a dual CAG-T7 promotor cassette (kind gift from Katharina Boroviak), thereby allowing expression in the mouse zygote and in vitro transcription of Cas9 mRNA. The vector was linearised and subjected to mMESSAGE mMACHINE® T7 ULTRA Transcription (Thermo Fisher Scientific) before column purification (RNeasy, Qiagen). 75 ngs of Cas9 mRNA, 9.3ngs each of 4 guides was injected into the cytoplasm of fertilised C57BL6n oocytes and transferred into CBAB6F1/J recipients. All applicable European, national, and institutional guidelines for the care and use of animals were followed. All procedures performed in studies involving animals were in accordance with the ethical standards of the Sanger Institute.
Both male and female mice were used in approximately equal proportions for all experiments. For antibody blockade experiments, mice were randomized to allow for at least two treatments represented in each cage of animals. Minimum sample size of six animals per experimental group (3 for steady state) was determined based on experience with colitis models.
Bone marrow chimeras. For BM chimera experiments with CD45.1 and Alpk1 knockout mice, recipient mice were lethally irradiated (2 × 550 rads/5.5 Gys) before reconstitution with BM from the indicated donor line. BM was aseptically collected from tibia and femur of the respective donor strain, and 5 × 10 6 cells were injected into the tail vein of the irradiated recipients. Mice were allowed to reconstitute for 8 wk before being infected with H. hepaticus.
T-cell transfer model of colitis. B6.Rag1 −/− mice were injected intraperitoneally (i.p.) with 0.2 × 10 6 FACS-sorted naive CD45RB high CD4 + T cells derived from mouse spleens. Mice were monitored bi-weekly for weight loss. 6-8 weeks after the injections, after the weight loss was approaching 20%, mice were killed and the extent of colitis was analysis by FACS and histology.
Hh-driven colitis models and in vivo treatments. Scoring of mouse colitis. Colonoscopy to assess colitis severity was performed and scored according to the methods of Becker et al. 37 . Histological assessment of colitis severity was performed following the established procedures 38 . Briefly, formalin-fixed paraffin-embedded cross-sections of proximal, middle, and distal colon were stained with haematoxylin and eosin and graded on a scale of 0-3 for four parameters: epithelial hyperplasia and goblet cell depletion, leucocyte infiltration, area affected, and features of severe disease activity. Common severity features include crypt abscess formation, submucosal leucocyte infiltration, and interstitial oedema. Scores for each criterion are added to give an overall score of 0-12 per colon section. Data from the three colon regions are then averaged to give an overall score. Scoring was conducted in a blinded fashion and confirmed by an independent blinded observer. Interobserver Pearson correlation coefficients ranged from 0.8 to 0.9. Photomicrographs of H&E stained colon sections were taken with a Coolscope Slide Scanner (Nikon). For TRUC mice scoring, the histological parameters were mononuclear cell infiltration, polymorphonuclear cell infiltration, epithelial hyperplasia, and epithelial injury were scored as absent (0), mild (1), moderate (2), or severe (3), giving a total score of 0-12 9 .
Mouse colon tissue preparation and cell isolation. LPLs were isolated following established procedures 39 . Briefly, mouse colons were washed with EDTA to remove epithelium and digested with collagenase VIII to liberate cell populations. Tissue digests were separated by centrifugation on a 40%/80% Percoll (Sigma) gradient. Cells at the 40%/80% interface were collected as the lamina propria leucocyte enriched fraction.
Colon explant cultures. 3 mm segments isolated from a middle part of mouse colon were cultured overnight in RPMI media supplemented with Pen-strep antibiotics (Sigma), 10% FCS (Gibco) and 50 μM β-mercaptoethanol (Gibco). IFNγ was quantified in the supernatant by enzyme-linked immunosorbent assay (ELISA, R&D Systems, UK) and normalized to explant weight (mg of tissue).
Quantitation of H. hepaticus using real-time PCR. DNA was purified from caecal contents taken from H. hepaticus-infected mice using the DNA Stool kit (Qiagen). H. hepaticus DNA was determined using a Q-PCR method based on the cdtB gene 40 .
RNA extraction, cDNA synthesis, and qPCR. Tissues were disrupted using lysis beads and a homogenizer unit (Precellys, UK) in the RLT buffer (Qiagen, UK). Sorted or cultured cells were lysed directly in the RLT buffer and homogenized by pipetting. RNA was isolated using RNEasy Mini or Micro kits (Qiagen, UK) followed by reverse transcription using random primers (Applied Biosystems, UK). Quantitative PCR (qPCR) was performed using Taqman assays (Applied Biosystems) and PrecisionPlus Mastermix (Primer Design, UK) on a ViiA7 384-well realtime PCR detection system (Applied Biosystems). All expression levels were normalized to an internal house-keeping gene Hprt and calculated as 2 (CTHprt−CTgene) .
Microarray analysis. Expression profiles of colon tissues from 129.Rag2 −/− and 129. HiccsB6 .Rag2 −/− were obtained using Illumina MouseWG-6-V2 microarrays (n = 4 for each condition). Array signal intensities were background adjusted, transformed using the variance-stabilizing transformation and quantile normalized using Lumi 41 from R/Bioconductor. Probes were retained if they were expressed significantly above background levels in at least four samples. This resulted in the analysis of 20,997 probes representing 15,443 genes. Differential expression analysis was performed using the empirical Bayes method in LIMMA 42 .
Significance was defined as a Benjamini-Hochberg adjusted P-value <0.05.
RNA sequencing analysis. PolyA-selected RNA from in vitro differentiated macrophages (Fig. 5) from pooled bone marrows from three mice (split into three technical replicates per condition) was used to prepare cDNA libraries (in-house dUTP protocol) that were subject to next-generation sequencing (Illumina HiSeq4000; 75 bp; minimum of 50M read pairs per sample). Mapping was performed using HISAT2 43 (two pass strategy to incorporate novel splice-sites) together with an index built from the mouse genome (mm10) and Ensembl transcript annotations (version 88). Differential expression analysis was performed using read counts (quantitated with featureCounts 44 ) and the DESeq2 algorithm 45 (local fit). P-values were adjusted for multiple-testing using the Benjamini-Hochberg correction.
Pathways and gene set enrichment analysis. We tested for the enrichment of pathways that were differentially upregulated at day 2 of Hh infection in 129. Rag2 −/− and 129 HiccsB6 .Rag2 −/− mice. Gene set enrichment analysis (GSEA) was conducted using the GSEA desktop program; Hallmark gene sets were downloaded from the Molecular Signatures Database (http://software.broadinstitute.org/gsea/ msigdb/collections.jsp). For gene ontology enrichment analysis (focusing on GO_ImmuneSystemProcess and REACTOME terms), the top 100 most significantly differentially expressed genes in 129.Rag2 −/− vs 129 HiccsB6 .Rag2 −/− mice after 2 days of Hh stim were analysed using ClueGO v2.1.7 with a minimum kappa score threshold of 0.5 and right-sided hypergeometric test with Bonferroni stepdown correction to identify significant terms. For RNA-Seq, top significant (adj. pvalue <0.01) genes differentially expressed in Alpk1 +/− and Alpk1 −/− BMDMs were analyzed using the online tool Panther 46 and Reactome pathways 47 as a source of terms for pathway enrichment analysis. TCRβ-BV510 (1/100; H57-597). All antibodies were from eBioscience (UK), Biolegend (UK), Becton Dickinson (UK), or R&D Systems (UK). Dead cells were excluded using efluor-780 fixable viability dye (eBioscience). Samples were acquired on FACS LSR Fortessa and FACS LSRII flow cytometers (Becton Dickinson). Cell sorting was performed using a FACS ARIA III (Becton Dickinson). Data were analyzed using FlowJo (Tree Star, USA). For intracellular cytokine staining, cells were restimulated with PMA (10 ng mL −1 ; Sigma-Aldrich), ionomycin (1 μg mL −1 ; Sigma-Aldrich), and 5 µg mL −1 brefeldin A (Sigma-Aldrich). After 3 h, cells were stained with fixable viability dye and surface markers, fixed with BD FACS Lysing Solution (Becton Dickinson, UK), and stained for intracellular cytokines in permeabilization buffer containing 0.05% saponin (Sigma-Aldrich). For staining FOXP3, cells were stained with fixable viability dye and surface markers prior to fixation and permeabilization using the FOXP3 staining buffer kit (eBioscience) according to manufacturer instructions. Gating strategy employed in the study is shown on Supplementary Fig. 9. Boolean analysis of cytokine expression by CD4 + T cells was performed using SPICE with default parameters 48 .
RNAScope. Alpk1 +/− Rag1 −/− and Alpk1 −/− Rag1 −/− mice were untreated (n = 4) or infected with Helicobacter hepaticus by oral gavage with 1 × 10 8 CFU (n = 5) and killed after 2 days. Formalin-fixed paraffin-embedded tissues from proximal colons were sectioned at 5 μm and collected onto Superfrost glass slides. Tissue sections were dewaxed in xylene and rehydrated through graded alcohol to water. Antigens were retrieved by boiling the tissue sections in target-retrieval buffer provided by ACD, following the manufacturer's instructions. For detection of mouse Il12b and Alpk1 mRNA, the RNAScope Multiplex Fluorescent Reagent kit v2 (ACD Europe SRL) was used. For Alpk1 mRNA detection, the probe Alpk1 was custom made to span all exons with the exception of exon 10, which is excised in Alpk1 KO animals. Briefly, paraffin sections were freshly cut, dried for 1 h at 60°C and dewaxed before mild unmasking with Target Retrieval buffer and protease, as per the manufacturer's instructions. Pretreated sections were hybridized with specific probes to Il12b and Alpk1, as well as Ppib and Polr2a (positive control) and irrelevant probe to dapb as a negative control. After hybridization signal amplification, TSA Plus fluorophores (PerkinElmer) were reconstituted and used to develop individual channel signal (Il12b in C3, Alpk1 in C5). Nuclei were stained with DAPI and sections were cover-slipped with Prolong Gold Antifade mounting medium (Thermo Fisher). Images were acquired using the A1 Zeiss Axioscope (Carl Zeiss™) at ×40 and ×100 magnifications. For final presentation, images were enhanced using Adobe Photoshop to increase contrast between the red (Alpk1) and yellow (Il12b) signals to background DAPI signal. All evaluations were performed in a blinded manner. For Il12b mRNA quantification, ×100 magnification images were used to count Il12b (yellow) dots in Alpk1-expressing cells (red dots). A minimum of 10 individual cells per mouse and per condition was used for quantification. The total number of cells expressing Il12b and Alpk1 in proximal colon was computed using ×40 magnification images.