key: cord-0875904-nci1e61a
authors: Goldberg, Rimma; Clough, Jennie N; Roberts, Luke B; Sanchez, Jenifer; Kordasti, Shahram; Petrov, Nedyalko; Hertweck, Arnulf; Lorenc, Anna; Jackson, Ian; Tasker, Scott; Appios, Anna; Omer, Omer; Parkes, Miles; Prescott, Natalie; Jenner, Richard G; Irving, Peter M; Lord, Graham M
title: A Crohn’s Disease-associated IL2RA Enhancer Variant Determines the Balance of T Cell Immunity by Regulating Responsiveness to IL-2 Signalling
date: 2021-06-12
journal: J Crohns Colitis
DOI: 10.1093/ecco-jcc/jjab103
sha: 9184639bc57479162c143954f0fa87fc24202803
doc_id: 875904
cord_uid: nci1e61a

BACKGROUND AND AIMS: Differential responsiveness to interleukin [IL]-2 between effector CD4(+) T cells [T(eff)] and regulatory T cells [T(reg)] is a fundamental mechanism of immunoregulation. The single nucleotide polymorphism [SNP] rs61839660, located within IL2RA [CD25], has been associated with the development of Crohn’s disease [CD]. We sought to identify the T cell immune phenotype of IBD patients who carry this SNP. METHODS: T(eff) and T(reg) were isolated from individuals homozygous [TT], heterozygous [CT], or wild-type [CC] for the minor allele at rs61839660, and used for phenotyping [flow cytometry, Cytometry Time Of Flight] functional assays or T cell receptor [TCR] sequencing. Phosphorylation of signal transducer and activator of transcription 5 [STAT5] was assessed in response to IL-2, IL-7, and in the presence of basiliximab, a monoclonal antibody directed against CD25. T(eff) pro-inflammatory cytokine expression levels were assessed by reverse transcription quantitative polymerase chain reaction after IL-2 and/or TCR stimulation. RESULTS: Presence of the minor T allele enhances CD25 expression, leading to increased STAT5 phosphorylation and pro-inflammatory cytokine transcript expression by T(eff) in response to IL-2 stimulation in vitro. T(eff) from TT individuals demonstrate a more activated gut homing phenotype. TCR sequencing analysis suggests that TT patients may have a reduced clonal capacity to mount an optimal regulatory T cell response. CONCLUSIONS: rs61839660 regulates the responsiveness of T cells to IL-2 signalling by modulating CD25 expression. As low-dose IL-2 is being trialled as a selective T(reg) modulator in CD, these findings highlight the potential for adverse effects in patients with this genotype.

Conclusions: rs61839660 regulates the responsiveness of T cells to IL-2 signalling by modulating CD25 expression. As low-dose IL-2 is being trialled as a selective T reg modulator in CD, these findings highlight the potential for adverse effects in patients with this genotype. 

C C rs61839660 IL2RA Chr10 T T IL2RA Chr10 C T IL2RA

Crohn's disease [CD] is a complex immune-mediated disorder with polygenic inheritance, in which inappropriate activation of the intestinal immune system in a genetically susceptible individual triggers chronic inflammation of the gastrointestinal tract. Despite the identification of multiple associated genetic polymorphisms, 1 few have been mapped to a mechanistic pathway and none have yet yielded a tractable therapeutic approach for CD patients. Recent years have seen an increasing array of treatment options with the advent of biologic therapy, including anti-tumour necrosis factor [TNF]-α and anti-integrin agents. However, rates of surgery and hospitalisation for CD patients have not significantly changed in the biologic era. 2 Therefore, there is a pressing need to explore novel therapeutic pathways. The value of a personalised approach to treatment is increasingly recognised, with the aim of targeting a treatment at the dominant aberrant pathway in a given patient to maximise efficacy and minimise side effects. 3 The IL-2 pathway has been identified as potentially therapeutically tractable in numerous autoimmune conditions, including inflammatory bowel disease [IBD] , with recruitment to a trial of low-dose IL-2 in CD under way [ClinicalTrials.gov NCT01988506]. Regulatory [T reg ] and effector [T eff ] CD4 + T cells are important mediators of the immune response in the gut of IBD patients, with defects in T reg number and suppressive function noted in the lamina propria and peripheral blood of patients with active IBD. 4, 5 Additionally, many identified IBD genetic risk loci map to immune cell enhancer regions, with particular enrichment in CD4 + T cell enhancers. 6 T reg are characterised by high constitutive expression of the IL-2 receptor alpha chain [IL2RA, CD25], a component of the highaffinity IL-2 receptor heterotrimer [IL2Rα/β/γ]. The lineage-defining transcription factor Forkhead box P-3 [FOXP3] is essential for their suppressive phenotype and stability. 7, 8 An additional defining feature of T reg is their low expression, or absence, of the IL-7 receptor alpha chain [CD127]. 9 In homeostatic conditions, IL-2 is mainly produced by activated CD4 + T cells in secondary lymphoid organs and is consumed at the same site by cells that express CD25, of which the majority are T reg. 10 By contrast, CD4 + T eff express CD25 only after activation, and at relatively lower levels compared with T reg , suggesting that T reg can preferentially respond to low concentrations of environmental IL-2. This differential responsiveness to IL-2 signalling is an important mechanism of immunoregulation.

Over 200 risk loci for IBD have been identified through genomewide association studies [GWAS] , with many in key regulatory pathways. 1, 11 The single nucleotide polymorphism [SNP] rs61839660 denotes a cytosine [C] to thymidine [T] base change on chromosome 10, located within a putative intronic enhancer region [intron 7] of IL2RA. rs61839660 was initially identified through GWAS, and is one of a minority of SNPs to be confidently resolved through finemapping as a variant highly associated with CD; 9.4% of CD patients carry the risk allele, 12 which equates to almost 18 000 CD patients in the UK. 13 rs61839660 is also associated with the development of other autoimmune conditions including ankylosing spondylitis, psoriasis, and primary sclerosing cholangitis, 6,14 demonstrating its importance for core immunoregulatory processes.

Previous experimental work has generated conflicting results on the effect of rs61839660 on CD25 expression in CD4 + T cells. A study in healthy human volunteers found that carriage of the rs61839660 minor allele was associated with increased levels of IL2RA messenger mRNA [mRNA] and increased CD25 surface expression on CD4 + memory T cells. 14 However, contemporaneous studies in which rs61839660 was delivered into murine CD4 + T cells via CRISPR-based methods, in addition to studies of healthy human carriers, reported that the minor variant of rs61839660 was associated with reduced CD25 transcript levels in CD4 + T cells stimulated with anti-CD3/CD28. 15 This group hypothesised, therefore, that rs61839660 impairs the function of the intronic enhancer that regulates CD25 expression in response to T cell receptor [TCR] stimulation of T eff. We sought to define the immune phenotype associated with rs61839660 in CD patients, with the aim of understanding potential therapeutic targets in this patient cohort.

Due to the Covid-19 pandemic and the closure of many patientfacing services in the UK, we were unable to source TT homozygote patient samples for certain experiments included in our study. Specifically, at certain times during which our work was carried out, the UK-wide IBD BioResource, upon which this study is based, was unable to continue recruitment based on UK Government legislation. Therefore, we were obliged to focus certain experiments on patients who are heterozygous for the minor risk allele at rs61839660, which represents nearly 10% of all patients with Crohn's disease worldwide. The data from these experiments are presented in Figures 

To ascertain the functional relevance of rs61839660, pSTAT5 assays were performed on sorted T reg and T eff of subjects of each genotype. PBMCs were defrosted the day before the experiment and rested overnight in XVIVO- 15 [Lonza] , at 37°C, 5% CO 2 . On the day of the experiment, cells were counted and 4 x 10 6 were kept aside for cytometry time of flight [CyTOF] staining. The remainder were washed and stained for flow sorting as described above. T reg and T eff from each sample were sorted into XVIVO-15 and taken forward into the pSTAT5 assay.

In the first experimental phase, the assay was optimised on T reg and T eff sorted from a healthy control [HC] [blood bank cone]. Three concentrations of recombinant human [rh]IL-2 were used [1, 10, or 100 IU/ml per 1 x 10 5 cells] and each concentration was assessed at three time points [15, 20, and 30 min ]. An unstimulated sample was run in parallel. 10 IU/ml IL-2 for 15 min yielded a robust pSTAT5 response, but not the maximal response. This was selected as the most appropriate initial concentration to assess pSTAT5 response without saturating the pathway, to permit discrimination of differences in T reg and T eff responses.

As a control, 1 x 10 5 sorted T reg and T eff from each subject were stimulated with recombinant human IL-7 across a logarithmically increasing dose range of concentrations [0, 1, 10, 100 IU/ml] under the same conditions. We hypothesised that, as IL-7 signals through CD127 and not through CD25, there should be no difference in the pSTAT5 response seen across genotypes.

For STAT5 phosphorylation assays, the BD Phosphoflow Protocol for human PBMCs was followed. Cells were washed, re-suspended in X-VIVO15, and rested for 1 h at 37°C, 5% CO 2 , before stimulation with IL-2 for 15 min in a 37°C water bath. Cells were fixed with warmed Cytofix buffer and incubated at 37°C for 10 min, then centrifuged at 600 x g for 5 min, supernatant discarded, and re-suspended in 100 μl Perm Buffer III. Cells were transferred to a 96-well v bottom plate and incubated on ice for 30 min, before washing and staining with 1 µl pSTAT5 AF488 (BD, clone 47/ Stat5[pY694]) in 10 μl Stain Buffer, and incubated at room temperature for 60 min. Finally, cells were washed and re-suspended in 100 μl of PBS for acquisition on a Fortessa [BD] . Standardised acquisition settings were put in place and used in all subsequent experiments. Gates for the pSTAT5-positive population were set against an fluorescence minus one [FMO] control and the unstimulated sample.

2.4. Assessment of pro-inflammatory cytokine transcript expression by CD4+ T eff in response to IL-2 and acute T cell receptor stimulation. T eff were isolated by fluorescence-activated cell sorting [FACS] as described. Cells were cultured at 37 o C/5% CO 2 at a density of 4 x 10 5 -1 x 10 6 

Optimisation work was performed to identify basiliximab conditions that would permit discrimination of CD25 blockade and subsequent pSTAT5 suppression between samples [HC blood bank cone, two replicates performed]. Basiliximab doses of between 0.001 and 10 µg/ml have been reported to cause marked CD25 blockade and thus reduction of lymphocyte proliferation when applied for up to 48 h to cell culture. 16, 17 Trials were performed using 0.01, 0.1, and 1.0 µg/ml basiliximab to sorted T reg or T eff in culture. Following the defined treatment period, cells were stimulated with 0-100 IU/ml IL-2 and the functional pSTAT5 assay was performed as described above. An untreated sample was used as a control. A 6-h application of 0.1 and 1.0 µg/ml basiliximab caused marked, but not complete, pSTAT5 suppression, and were taken forward into assays on patient samples. Cells for CyTOF staining were defrosted the day before the experiment and rested overnight in complete RPMI supplemented with 10% FCS at 37°C, 5% CO 2 . Cells were washed and dispersed into FACS staining tubes at 2 x 10 6 cells per panel. A cisplatin viability stain was performed (cells re-suspended in 1 ml PBS, with 1 μl of cisplatin for 5 min, then reaction quenched with 3 ml MACS buffer, which was used as the cell staining medium [CSM] for all CyTOF experiments). Cells were then centrifuged at 400 x g for 5 min at room temperature, supernatant discarded, pellet dispersed, and 4 μl of FC block [MACS, Miltenyi Biotec] applied per sample. Conjugated antibodies were purchased from Fluidigm. Unconjugated antibodies were purchased from Biolegend and conjugated to the requisite metal isotopes using MaxPar labelling kits [Fluidigm] . A master mix of the extracellular CyTOF antibodies was made up for each panel and the samples were stained in a total volume of 100 μl for 30 min at room temperature. Following staining, cells were washed with CSM and re-suspended in fix/perm buffer [eBioscience FOXP3/ Transcription factor staining buffer set] followed by a 1-h incubation at room temperature. Cells were washed with 1x permeabilisation buffer and stained with intracellular/transcription factor antibody master mix for 30 min at room temperature. Cells were then washed again with 1x permeabilisation buffer and re-suspended in 500 μl of 2% paraformaldehyde [PFA] , left at room temperature for 1 h, and refrigerated overnight. On the day of acquisition, the cells were pelleted, washed, and resuspended in 1x Ir [Iridium] intercalator solution for 30 min; this acted as a DNA/final viability stain. After two further PBS washes, cells were re-suspended in water at a maximal concentration of 5 x 10 5 with an internal bead standard [DVS EQ beads, Fluidigm].

Raw FCS files were normalised using Fluidigm CyTOF Software [https://www.fluidigm.com/software]. Data were uploaded to Cytobank [https://mrc.cytobank.org] and standard pre-processing performed to remove debris, doublets, beads, and dead cells. CD4 + T Cells were manually gated and used as a starting population for the automated unsupervised analysis.

Cytobank's implementation of the t-Distributed Stochastic Neighbour Embedding [t-SNE] algorithm [known as viSNE] 18 was used to transform the data to two dimensions, while still conserving the high dimensional structure of the data. The resulting t-SNE1 and t-SNE2 dimensions were then inputted into a clustering algorithmspanning tree-progression analysis of density normalised events [SPADE; available on Cytobank], 19 which extracts a cellular hierarchy from high-dimensional cytometry data and presents the populations in a branched tree diagram.

Cluster frequencies were quantified based on SPADE population clusters, and different cell populations were identified based on the median expression of known markers using our in-house developed pipeline [CytoClustR; https://github.com/kordastilab/cytoClustR]. For example, T reg and T eff nodes were identified in the SPADE tree based on the expression of CD127, CD25, and FOXP3. Further populations were identified within the T reg and T eff nodes, based on differential expression of one or more of the rest of the markers within those populations. Both marker expressions and cell abundances were analysed with one-way analysis of variance [ANOVA] for differences between the genotype groups for each identified population.

To quantify similarity/difference between the populations as well as expression of different markers, marker enrichment modelling [MEM] scores were calculated for each identified population and the output was visualised using heatmaps. The MEM algorithm objectively describes characteristics of cell populations using large numbers of markers, rather than a few most differentially expressed, and provides each marker with a score which denotes its expression level. 20 2.9. TCR sequencing 2.9.1. RNA extraction 1 x 10 5 T reg and T eff were FACS-purified [ There were no significant differences between genotype groups in terms of age, sex, or disease distribution. There was no difference in rates of medication use or need for surgery, suggesting that minor allele carriers do not have a more severe disease phenotype.

CD4 + T reg and T eff were sorted by flow cytometry [Supplementary Figure 1 ]. We demonstrated increased CD25 cell surface expression in CD patients and healthy controls, specifically on T eff of TT patients as measured by mean fluorescence intensity [MFI] of flow cytometry staining [ Figure 1a ], compared with CC subjects. There was a non-significant trend for increased CD25 expression on T reg in TT patients [ Figure 1b ] as well as a non-significant trend towards enhanced CD25 expression on T eff and T reg from CT heterozygote carriers of the minor allele at rs61839660 [ Figure 1a , b, and d]. The ratio of T reg :T eff CD25 expression was significantly reduced in TT patients [ Figure 1c ] compared with CC patients. No difference was seen in T reg or T eff CD25 expression between CD patients and healthy controls with the same genotype [ Figure 1d -f], suggesting that differential CD25 expression between genotypes is conferred by the risk allele, rather than by the presence of inflammatory disease.

To determine whether rs61839660 affected the CD4 + T cell response to IL-2 signalling, the extent of STAT5 phosphorylation [pSTAT5] was measured in T eff and T reg following short-term exposure to 10 IU IL-2 [ Figure 2 ]. In optimisation work, this concentration of IL-2 was found to be sufficient to induce pSTAT5 expression in approximately 50% of T reg, with no pSTAT5 expression in T eff in CC patients. T eff and T reg from TT patients had an enhanced pSTAT5 response as measured by mean fluorescence intensity [MFI] of pSTAT5, compared with CT and CC patients [ Figure 2a and b]. T eff from TT subjects also exhibited a higher proportion of pSTAT5 + cells compared with T eff from CC patients [ Figure 2c ]. There was a nonsignificant increase in the proportion of pSTAT5 + T reg in TT patients [ Figure 2d ]. There was no difference in T eff or T reg pSTAT5 response to IL-2 signalling between healthy controls and CD patients of the same genotype [ Figure 2e , f, g, and h]. As noted for CD25 expression, a non-significant trend for enhanced pSTAT5 staining was observed in CT heterozygote carriers of the minor allele, and this was more apparent in T eff than T reg [ Figure 2a -h].

It was noted that among TT subjects, there was a subgroup that exhibited a more pronounced pSTAT5 response to IL-2 in both T eff and T regs [ Figure 2a -d]. We performed a subanalysis on this 'hyper-responder' cohort, defining subjects with a T eff MFI pSTAT5 greater than 395 and a T reg MFI pSTAT5 greater than 1100 in response to 10 IU/ml IL-2, as 'hyper-responders' [ Table 2 ]. The same five subjects with hyperresponsive T eff also exhibited an exaggerated T reg pSTAT5 response. There was no significant difference in age, sex, disease distribution, disease severity, or proportion of patients exposed to biologic medication between the TT 'hyper-responders' and the remaining TT subjects. There was no significant difference in T eff or T reg CD25 expression between the TT 'hyper-responder' subgroup and the remaining TT subjects. As a control, the pSTAT5 response to IL-7 was examined across a range of doses [0-1000 IU]. rs61839660 did not affect CD127 surface expression across the genotypes in T eff or T reg [ Supplementary Figure 2a 3.3. The rs61839660 minor allele enhances transcription of pro-inflammatory cytokines following exogenous IL-2 and acute TCR stimulation of T effs Together, our findings suggested that the rs61839660 minor allele may confer a functional 'hyper-responsive' state to IL-2 signalling in T eff . To further investigate this proposal, we sought to identify whether presence of the minor allele influenced the functional response to IL-2 by T eff , by measuring cytokine expression [IFNG and TNF] responses by the cells after TCR stimulation.

T eff from cryopreserved PBMCs of CD patients, genotyped as CC and CT at rs61839660, were isolated by FACS and used to conduct the assays. T eff were stimulated for 24 h with a range of IL-2 doses only, or with IL-2 followed by a brief stimulation [6 h] with different concentrations of plate-bound αCD3/αCD28 [ Figure 3 ].

We observed that for both expression of IFNG [ Figure 3a and b] and TNF [ Figure 3c and d], compared with CC samples, the presence of the minor T allele at rs61839660 enhanced the relative transcript abundance in response to IL-2, together with TCR stimulation. This was most clearly observed at the highest concentration of IL-2 [100 IU/ml] for each of the TCR cross-linking conditions [ Figure 3a and c]. Importantly, greater pro-inflammatory cytokine relative transcript abundance was also observed in CT samples treated with 100 IU/ml of rhIL-2 in the absence of TCR crosslinking, despite similar baseline [0 IU IL-2, no TCR cross-linking] transcript abundance levels between the genotypes [ Figure 3b and d]. These data support the notion that within the context of CD, presence of the rs61839660 minor allele confers a functional capacity for hyper-responsiveness of CD4 + T eff in response to IL-2 signalling. [d] There was no difference in CD25 expression in T eff or [e] T reg or [f] in the MFI CD25 ratio T reg /T eff between CD patients and healthy controls within each genotype group. Mean ± SEM plotted. Statistical analyses performed using one-way ANOVA, Tukey post hoc test, for comparisons involving more than two groups, and unpaired t-ests for comparisons involving two groups. TT: n = 16, CT: n = 19, CC: n = 19. **p <0.01, ***p <0.001. HC, healthy control; CD, Crohn's disease; ns, not significant; SEM, standard error of the mean; ANOVA, analysis of variance.

Basiliximab, a mouse-human chimeric monoclonal antibody which antagonises CD25, is currently licensed for the prevention of renal allograft rejection. We examined whether basiliximab added to in vitro culture of sorted T reg and T eff could block pSTAT5 activation in a cohort of CD patients with CC and CT rs61839660 genotypes, recruited from our local patient population. Basiliximab was effective in blocking T eff and T reg pSTAT5 activation in both genotypes [ Figure  4a -f], confirming that the enhanced pSTAT5 response is dependent on signalling through CD25. Of note, CT T eff also demonstrated elevated baseline pSTAT5 activation in the absence of basiliximab, when compared with CC T eff [ Figure 4d ], as previously seen in the other sample cohorts used in our study. There was no significant difference in the response to basiliximab between the genotype groups.

To further elucidate the phenotype of T reg and T eff from patients harbouring the rs61839660 minor variant, we performed an in-depth phenotypic analysis using Cytometry Time of Flight [CyToF] [Panels Figure 5a ]. However, this analysis also revealed a comparable increase in CD25 expression on TT T reg , providing an explanation as to why homozygous minor allele T regs also exhibited hyper-responsiveness during IL-2/STAT5 phosphorylation assays [ Figure 2b ]. Furthermore, this analysis demonstrated that both T reg and T eff from minor allele homozygotes have a more activated, gut-homing phenotype compared with cells from CT and CC subjects [ Figure 5c ]. Key markers that were differentially upregulated in TT patients included the chemokine receptor CXCR3, which is typically associated with inflammatory Th1 cell infiltration, and CXCR4, associated with IL-2-mediated immune activation. GPR15, which specifically traffics cells to the colon, was also preferentially expressed on both T reg and T eff of TT patients, as were other markers associated with T reg tissue trafficking such as CCR6 and integrin alpha 7 [ Supplementary Figures 4a-c and 5a- 

To assess whether the difference in phenotypes observed between the three patient groups was driven by their ability to respond to antigens, we performed T cell receptor [TCR] sequencing of T reg and T eff from three CD patients from each SNP genotype. We assessed the variability displayed in both the V and J segment of the CDR3 region. We found that the J segments in TT patient cells consistently displayed the least variability, those from CC patient cells displayed the greatest variability, and the CT patients had an intermediate phenotype [ Figure   αCD3 : 0 µg/ml αCD28: 0 µg/ml αCD3: 0 µg/ml αCD28: 0 µg/ml 6a]. J segments in T eff showed higher variability than T reg across the three genotypically defined groups. Further analysis of individual J segments revealed that TT cells preferentially use two TRBJ segments, segments TRBJ2-1 and TRBJ2-7 [ Figure 6b and c]. Principal component analysis of J segment usage revealed distinct clustering by genotype [ Figure 6d ]. Such differential clustering was not observed on analysis of the V segment usage. We then examined whether any of the individuals cluster together based on shared clonotypes on either an amino acid or a nucleotide level, and did not observe clustering based on genotypes [ Figure 6e ]. Finally, to assess whether TCR diversity was different between T reg and T eff subsets, we examined their diversity ratio across the three genotypes. We found that this ratio was lower in TT and CT patients compared with CC patients. Whereas T regs were more diverse than T eff [measured by Gini index] in CC patients, in TT and CT patients, T reg diversity was comparable to or lower than T eff [ Figure 6f ]. The reduced T reg diversity when compared with T eff in TT patients, coupled with T eff hyper-responsiveness to IL-2, may suggest that TT patients are less efficient at mounting a regulatory response to counter T eff -driven inflammation.

In summary, we show that the rs61839660 minor allele enhances sensitivity to IL-2 signalling in CD4 + T cells by increasing CD25 expression. We propose that the reduced ratio of T reg :T eff CD25 expression in individuals homozygous [TT] for this SNP may increase the ability of T eff to respond to low environmental IL-2, permitting activation of T eff at levels of IL-2 which would usually only activate T reg . This T eff hyper-responsiveness to IL-2 could therefore be a major driver of inflammation in these subjects. Mass cytometric analysis of immune cells from these patients confirmed that CD25 is elevated in patients homozygous for the risk allele at rs61839660, and that effector T cells from these patients are primed to home to the intestine by virtue of upregulation of gut trafficking molecules such as alpha 4 beta 7 integrin and GPR15.

Here we show that rs61839660 enhances IL-2 signalling in CD4 + T cells from CD patients by regulating the expression of CD25. Effector T cells from TT patients display the ability to respond to doses of IL-2 that normally only activate regulatory T cells. Additionally, T eff from carriers of the minor allele express elevated transcript levels of pro-inflammatory cytokines in response to IL-2 and/or TCR signalling-supportive of our conclusion that rs61839660 regulates of pSTAT5 stained T reg after stimulation with 10 IU/ml IL-2 following basiliximab treatment. Mean ± SEM plotted. Statistical analyses performed using two-way ANOVA, Tukey's multiple comparisons test for comparisons involving more than two groups, and paired t tests for comparisons involving two treatments within the same genotype. *p <0.05, **p <0.01, ****p <0.0001. BX, basiliximab; Ns, not significant; SEM, standard error of the mean; ANOVA, analysis of variance.

the functional responsiveness of T cells in CD patients via the regulation of sensitivity to IL-2 signalling. Our findings are in contrast to previous work involving CRISPR inactivation of immune enhancers in mice, which proposed that rs61839660 impaired induction of CD25 and therefore disrupted IL-2 signalling in T cells, and found that carriage of the risk allele was associated with reduced CD25 transcript levels in stimulated undifferentiated CD4 + T cells from healthy human subjects. 15 However, our findings are supported by a previous study of 323 healthy human subjects, which found that the presence of the rs61839660 minor allele was associated with increased IL2RA transcript levels and surface expression of CD25 on CD4 + memory T cells. 14 Basiliximab was found to effectively block T eff hyperresponsiveness in minor allele heterozygote [CT] patients. Previous clinical trials of basiliximab in small cohorts of unselected patients with ulcerative colitis, in whom rs61839660 had not been identified as a risk locus, showed safety but did not demonstrate efficacy compared with conventional treatment. 22, 23 However, these data suggest that basiliximab could have a role in blocking the hyper-responsive T eff population seen in rs61839660 carriers. Further in vitro experimental work is required to assess the potential therapeutic value of CD25 blockade in this cohort.

Mass cytometric analysis of immune cells from these patients was also performed, which confirmed that CD25 expression is elevated in patients homozygous for the TT allele at rs61839660 and that effector T cells from these patients are primed to home to the intestine by virtue of upregulation of gut-trafficking molecules such as alpha 4 beta 7 integrin and GPR15. Analysis of the TCR repertoire further reveals that diversity in the T reg subset does not mirror diversity in the T eff subset in TT patients. Lack of ability in the T reg compartment to respond to aberrant T eff responses in TT patients, may be another underlying pathogenic mechanism. Unlike in type 1 diabetes mellitus, where rs61839660 is protective, 12, 14 in CD rs61839660 is linked to pathology. Based on our observations, the mechanism of this likely involves the capacity of rs61839660 to render T eff hyper-responsive to doses of IL-2 which typically activate only T regs , putatively resulting in an enhanced activation profile of the cells. This, alongside an activated gut-homing profile of TT T eff , potentially explains the association between rs61839660 and CD. However, it will be necessary to follow up our observational analyses with further mechanistic studies to fully confirm or refute these hypotheses.

This study provides a unique perspective on the link between genetic risk and the development of disease, by studying genotyped CD and HC patients recruited across the UK. Recruiting genotyped human subjects across the country presents a logistically complex process; thus a limitation of our study is that insufficient patient numbers could be recruited to adequately study the link between disease severity and the risks allele. Further studies are also needed to define the interplay between the rs61839660 risk allele and other genotypic and environmental factors in patients with CD, to allow for a precise definition of the mechanism of pathogenicity of the rs61839660 risk allele. This is one of the first works in IBD to link a pathogenic mechanism to a CD-associated genetic variant, advancing the goal of a personalised therapeutic approach. Accurate phenotypic and genetic profiling of CD patients would allow targeted therapeutic intervention, permitting selection of a treatment targeted to correct the aberrant pathway in a given patient. We highlight the vital role of open-access resources, such as the NIHR IBD BioResource, for sourcing genetically selected material for functional analysis of rare variants. We provide a novel mechanistic link between genotype and function relevant to multiple human autoimmune diseases. Over 40 phase 2/3 trials of low-dose IL-2 in the treatment of autoimmune conditions are currently registered, including in CD [ClinicalTrials.

gov NCT01988506]. These trials typically use doses of 0.5-3 x 10 6 IU/m 2 body area, which are delivered subcutaneously under trialspecific regimens. The premise of this treatment is that low doses of IL-2 will preferentially expand the tolerising T reg population, correcting the dysregulated autoimmune response. However, given the exaggerated T eff response to low doses of IL-2 in carriers of rs61839660, these subjects could be at increased risk of deleterious clinical outcomes due to theoretical T eff hyperactivation. With 9.4% CD patients carrying the risk allele, this represents a significant subset of potential trial participants. Furthermore, our findings suggest there is merit in further assessing the role of basiliximab in genetically stratified patients, as an adjunct to existing treatments. 

International IBD Genetics Consortium [IIBDGC]. Host-microbe interactions have shaped the genetic architecture of inflammatory bowel disease

Introduction of anti-TNF therapy has not yielded expected declines in hospitalisation and intestinal resection rates in inflammatory bowel diseases: a population-based interrupted time series study

Personalised medicine in Crohn's disease

Peripheral and intestinal regulatory CD4+ CD25[high] T cells in inflammatory bowel disease

Common immunologic mechanisms in inflammatory bowel disease and spondylarthropathies

Genetic and epigenetic fine mapping of causal autoimmune disease variants

Pillars article: Foxp3 programs the development and function of CD4+CD25+ regulatory T cells

Pillars article: control of regulatory T cell development by the transcription factor Foxp3

CD127 expression inversely correlates with FoxP3 and suppressive function of human CD4+ T reg cells

The role of interleukin-2 during homeostasis and activation of the immune system

International Multiple Sclerosis Genetics Consortium; International IBD Genetics Consortium. Association analyses identify 38 susceptibility loci for inflammatory bowel disease and highlight shared genetic risk across populations

Fine-mapping inflammatory bowel disease loci to single-variant resolution

IBD prevalence in Lothian, Scotland, derived by capture-recapture methodology

A rare IL2RA haplotype identifies SNP rs61839660 as causal for autoimmunity

Discovery of stimulationresponsive immune enhancers with CRISPR activation

Comparative in vitro study of the immunomodulatory activity of humanized and chimeric anti-CD25 monoclonal antibodies

Short-term anti-CD25 monoclonal antibody administration down-regulated CD25 expression without eliminating the neogenetic functional regulatory T cells in kidney transplantation

viSNE enables visualization of high dimensional single-cell data and reveals phenotypic heterogeneity of leukemia

Extracting a cellular hierarchy from highdimensional cytometry data with SPADE

Characterizing cell subsets using marker enrichment modeling

IBD BioResource: an openaccess platform of 25 000 patients to accelerate research in Crohn's and Colitis

BASBUC Investigators. Basiliximab for the treatment of steroid-resistant ulcerative colitis: further experience in moderate and severe disease

Basiliximab does not increase efficacy of corticosteroids in patients with steroid-refractory ulcerative colitis

We acknowledge the help of the IBD Bioresource, supported by the NIHR, Medical Research Council, Crohn's and Colitis UK, Open Targets, and the Wellcome Trust Sanger Institute.

Supplementary data are available at ECCO-JCC online.