Relationship between Multiple Sclerosis-Associated IL2RA Risk Allele Variants and Circulating T Cell Phenotypes in Healthy Genotype-Selected Controls

Single nucleotide polymorphisms (SNPs) in or near the IL2RA gene, that encodes the interleukin-2 (IL-2) receptor α (CD25), are associated with increased risk of immune-mediated diseases including multiple sclerosis (MS). We investigated how the MS-associated IL2RA SNPs rs2104286 and rs11256593 are associated with CD25 expression on T cells ex vivo by multiparameter flow cytometry in paired genotype-selected healthy controls. We observed that MS-associated IL2RA SNPs rs2104286 and rs11256593 are associated with expression of CD25 in CD4+ but not CD8+ T cells. In CD4+ T cells, carriers of the risk genotype had a reduced frequency of CD25+ TFH1 cells (p = 0.001) and an increased frequency of CD25+ recent thymic emigrant cells (p = 0.006). Furthermore, carriers of the risk genotype had a reduced surface expression of CD25 in post-thymic expanded CD4+ T cells (CD31−CD45RA+), CD39+ TReg cells and in several non-follicular memory subsets. Our study found novel associations of MS-associated IL2RA SNPs on expression of CD25 in CD4+ T cell subsets. Insight into the associations of MS-associated IL2RA SNPs, as these new findings provide, offers a better understanding of CD25 variation in the immune system and can lead to new insights into how MS-associated SNPs contribute to development of MS.


Introduction
The cytokine interleukin-2 (IL-2) has an extensive array of effects that are central for the balance between immune tolerance and immunity [1,2]. The IL-2 receptor α chain (IL-2Rα), also known as CD25, specifically binds IL-2 and is a central component in the trimeric IL-2 receptor complex [1,3]. CD25 has a short cytoplasmic domain and is not involved in the downstream signal transduction [4]; instead CD25 delivers IL-2 to IL-2Rβ and γc in a stepwise assembly [5,6] and creates a high-affinity receptor complex for the low levels of IL-2 in vivo [3]. Responsiveness to IL-2 directly depends on the cell's expression of CD25 as IL-2 signaling increases with its expression and correlates with the capacity to bind IL-2 in a stable IL-2/IL-2R complex [7,8].
Among T cells, there is heterogeneity in CD25 expression [7,8], with some T cells such as T regulatory (T Reg ) cells constitutively expressing high levels [2,9]. However, on most cells expression of CD25 is inducible by T cell receptor activation and cytokine signals [3,7], including a positive IL-2 feed-back loop [10]. Heterogeneity in CD25 expression implies variability in the T cell's ability to benefit from the shared pool of IL-2, resulting in a tug of war with T cells with the highest expression of CD25 as winners [8,11].
IL-2 signaling is important for the survival and suppressive capacity of T Reg cells [1]. In addition, IL-2 modulates polarization of effector functions by inducing IFN-γ as well as receptors and transcription factors associated with a T helper 1 (T H 1) committed lineage [12,13], and expression of CD25 is required for the optimal development of T H 1 cells [14]. IL-2 is also crucial for the expression of GM-CSF in human T cells [15]. Conversely, expression of receptors and transcriptions factors associated with T H 17 lineage commitment is inhibited by IL-2 signaling [12,13]. In a competing environment, T Reg cells augment T H 17 differentiation, possibly by depletion of IL-2 [16,17]. Additionally, decreased IL-2 signaling increases the differentiation of T follicular helper (T FH ) cells [18,19], a CD4 + T cell subset able to provide specialized B cell help [20].
In CD8 + T cells, IL-2 signaling impacts the T cell's fate decision between memory and effector differentiation [1]. Cells receiving prolonged and strong IL-2 signaling by maintaining high CD25 expression expand rapidly, show enhanced effector functions, and are driven towards a terminally differentiated cell fate and apoptosis [7,21,22]. Conversely, cells receiving less IL-2 signaling are not cytolytic and differentiate into long-lived functional memory cells associated with a central memory phenotype [7,[22][23][24]. However, transient high-affinity IL-2 receptor signaling early after priming promotes development of highly functional effector memory cells that later give rise to long-lived central memory cells [25]. Furthermore, high-affinity IL-2 signaling is required for the ability of memory T cells to expand in a competing environment with other T cells and induce a potent secondary expansion [23,24].
The SNPs rs2104286 and rs11256593 in or near the IL2RA gene are associated with increased risk of developing MS [28][29][30]. The association between the SNP rs11256593 near the IL2RA gene and risk of MS has only recently been established in the MS replication chip study [29]. Previous studies of the MS-associated IL2RA SNP rs2104286 effects on immune cells have focused on a limited number of CD4 + T cell phenotypes. Carriers of the risk allele (T) for SNP rs2104286 were reported to have reduced IL-2 receptor signaling as measured by STAT5 phosphorylation [40], increased frequency of GM-CSF producing memory CD4 + T cells [15], increased frequency of CD25 + naïve T cells [41], and increased concentration of soluble CD25 [42]. Furthermore, studies in cell line models for helper and regulatory T cells have found that rs2104286 polymorphisms influence the activity of enhancer elements from the first intron in the IL2RA gene and the binding affinity of the transcription factor TFAP4 [43,44].
We aimed to investigate how CD25 expression is associated with MS-associated SNPs rs2104286 and rs11256593 in or near the IL2RA gene in human CD4 + and CD8 + T cell subsets ex vivo. We analyzed this in freshly isolated peripheral blood mononuclear cells (PBMC) from genotype-selected healthy controls by multiparameter flow cytometry using a paired experimental design allowing for the quantitative assessment of CD25 expression on a wide range of T cell subtypes. We confirm that homozygous carriers of the MS-associated IL2RA risk alleles have an increased frequency of CD25 + naïve CD4 + T cells and find that this difference is mainly observed in recent thymic emigrant cells. In addition, we report that homozygous carriers of the MS-associated IL2RA risk alleles have reduced CD25 expression on a wide range of memory CD4 + T cells and decreased frequency of CD25 + T FH 1 cells.

Study Participants
Study participants were recruited among 1000 healthy subjects in the Danish Blood Donor Cohort [45] who previously donated blood to the Danish Multiple Sclerosis Centers' (DMSC) contribution to the International Multiple Sclerosis Genetics Consortium (IMSGC) replication chip study [29]. The study was conducted in accordance with the Declaration of Helsinki and the protocol was approved by the scientific Ethics Committee in the Capital Region of Denmark (H-15008896). All participants gave written informed consent for inclusion before they participated in the study. Participants were selected based on a comprehensive lifestyle questionnaire (translated from Swedish and used with permission from Karolinska Institute, Sweden) [46] and the SNPs rs2104286 and rs11256593. SNP rs11256593 is the strongest associated SNP in the IL2RA gene region and in linkage disequilibrium (LD) with the previously associated lead SNP rs2104286 [29].
Selected study participants were recruited using a paired study design where each pair consisted of a homozygous carrier of the risk allele (T) for both MS associated IL2RA SNPs rs11256593 and rs2104286, defined as the risk genotype, and a homozygous carrier of the protective allele (C) of both SNPs, defined as a protective genotype. In addition, each pair was of the same sex and had a maximum age difference of 5 years. All participants were of self-reported European ancestry in two generations and not regular smokers for one year prior to inclusion. They reported no family history of MS and did not have any known autoimmune diseases (including type 1 diabetes, rheumatoid arthritis, inflammatory bowel disease, thyroid disease, or systemic lupus erythematosus). Furthermore, pregnant healthy subjects and healthy subjects with previous cancer were not recruited. Study participants had no self-reported intake of systemic anti-inflammatory, immunosuppressive, or immunomodulatory medications.
By recruiting 25 pairs, each pair consisting of a homozygous carrier of the risk genotype and a homozygous carrier of the protective genotype, we had 80% power to detect a standardized difference of 0.55 between subjects carrying the protective and risk genotype in a paired comparison. In a previous study we found that the association of the IL2RA SNP rs2104286 on the plasma concentration of soluble CD25 showed a standardized difference of 1.5 [47]. By including 25 genotype matched pairs we had sufficient statistical power to detect IL2RA SNP associations less than 40% of the magnitude of the known association with soluble CD25 concentrations.

Genotyping
Genotyping was performed on Illumina's MS replication custom bead chip and conducted by the International MS Genetics Consortium (IMSGC) [29].

Sample Collection and PBMC Isolation
Blood sampling and experimental procedures for each pair were performed on freshly prepared samples on the day of sample collection. Venous blood was collected in BD Vacutainer EDTA tubes (BD Bioscience, Plymouth, UK) from 1-2 fasting pairs between 7-9 a.m. Within 1 h the blood was diluted with phosphate buffered saline (PBS) (Life Technologies, Paisley, UK) containing 2mM EDTA (Thermo Fisher Scientific Baltics, Vilnius, Lithuania) and transferred to Leucosep tubes (Greiner Bio One; Frickenhausen, Germany), and the peripheral blood mononuclear cells (PBMCs) were isolated by density gradient centrifugation on Lymphoprep (Axis-Shield, Oslo, Norway). PBMCs were washed twice in PBS/EDTA.

Gating Strategy
Data analysis was performed blinded to genotype status using FlowJo software (Tree Star, Ashland, OR, USA). Gating strategies are shown in Figures 1, 3, 4 and 6. Matched isotype controls were applied where appropriate. Median fluorescence intensity (MFI) was used as a measure for surface expression of CD25 on CD25 + T cells ( Figure 3D,E). Samples with a sub-optimal staining were excluded pair-wise before unblinding of genotype status. Thereby, we excluded 2 pairs that had poor staining of CD45RA; 3 pairs with poor staining of CD39 and CD31; 4 pairs with no usable sample for CD39/CD31.

Statistical Analysis
Statistical analysis was conducted with IBM SPSS Statistics 22 and GraphPad Prism 7. Normal distribution was determined by a Shapiro-Wilk test. Depending on normal distribution, all statistical analyses regarding genotype-phenotype associations were performed using a paired-sample t-test or a Wilcoxon Signed Rank test of samples collected the same day from matched pairs carrying either the risk or protective IL2RA alleles. The significance level was set at p ≤ 0.005 and p-values ≤ 0.05 were considered suggestive in accordance with recent recommendations [48].

Results
To identify genotype-phenotype associations between MS-associated gene variants in the IL2RA gene and changes in CD25 expression in T cells, healthy study participants were recruited based on genotype status for the MS associated IL2RA SNPs rs2104286 and rs11256593 and lifestyle information from a large cohort of healthy blood donors. Only healthy subjects were recruited as previous studies have suggested that immunological alterations in established MS may compromise the ability to demonstrate the biological associations of SNPs associated with MS risk [39,47]. Recruited participants were either homozygous carriers of the risk alleles (T) for both SNPs rs2104286 and rs11256593, defined as a risk genotype (TT) or homozygous carriers of the protective alleles (C) for both SNPs, defined as a protective genotype (CC). Two-hundred subjects met our recruitment criteria and were contacted. Approximately 100 responded and in total we recruited 50 healthy subjects who were paired with respect to sex, age and genotype status for experimental procedures and statistical analysis of 25 pairs (Table 1).

The MS-Associated IL2RA Risk Genotype Has Few Associations with CD25 Expression on CD8 + T Cells
In CD8 + T cells, expression of CD127 and CD25 defined the same two populations as in CD4 + T cells ( Figure 1J). The CD127 + CD25 + population represented 8.4% (IQR: 6.0-12.1%). The CD127 + CD25 + gate was applied to the six CD8 + T cell differentiation subsets ( Figure 1H). We found that the T-CM population had the highest frequency of CD127 + CD25 + cells among CD8 + T cells ( Figure 1L) and that carriers of the risk genotype (TT) had a suggestively lower frequency of CD127 + CD25 + T-CM cells ( Table 2: p = 0.03). We observed no associations of the MS-associated IL2RA genotype with frequency or surface expression (MFI) of CD25 on the other CD127 + CD25 + CD8 + T cell subsets (Table 2). In CD8 + T cells, the CD127 lo CD25 hi population only represented 0.49% (IQR: 0.24-1.2%) and therefore excluded from statistical analysis.

Lower Frequency of CD25 + T FH 1 Cells in Carriers of the IL2RA Risk Genotype
Blood T follicular helper (T FH ) cells can be identified by surface expression of CXCR5 and PD-1 [20,53]. The PD-1 + CXCR5 + (T FH ) cells were gated on CD127 + cells to exclude contamination from T Reg cells ( Figure 4B) and comprised 3.1% (IQR: 2.0-3.8%) of CD4 + T cells. Carriers of the risk genotype (TT) had a suggestively lower frequency of CD25 + T FH cells (p = 0.02) compared to carriers of the protective genotype (CC) ( Figure 4G). Differential surface expression of chemokine receptors CXCR3 and CCR6 separates T FH 1 (CCR6 − CXCR3 + ), T FH 2 (CCR6 − CXCR3 − ), and T FH 17 (CCR6 + CXCR3 − ) cells [54] ( Figure 4E). T FH 1 cells made up the largest group of T FH cells ( Figure 4I) and we found that carriers of the risk genotype had a significantly lower frequency of CD25 + T FH 1 cells (p = 0.001) while we observed no IL2RA genotype associations on the T FH 2 and T FH 17 cells ( Figure 4J). We observed no IL2RA genotype associations with surface expression levels of CD25 on the CD25 + T FH subsets ( Figure 4K).  protective genotype (CC) ( Figure 4G). Differential surface expression of chemokine receptors CXCR3 and CCR6 separates TFH1 (CCR6 − CXCR3 + ), TFH2 (CCR6 − CXCR3 − ), and TFH17 (CCR6 + CXCR3 − ) cells [54] ( Figure 4E). TFH1 cells made up the largest group of TFH cells ( Figure 4I) and we found that carriers of the risk genotype had a significantly lower frequency of CD25 + TFH1 cells (p = 0.001) while we observed no IL2RA genotype associations on the TFH2 and TFH17 cells ( Figure 4J). We observed no IL2RA genotype associations with surface expression levels of CD25 on the CD25 + TFH subsets ( Figure  4K).
Finally, we applied the T Reg (CD127 lo CD25 hi ) gate to each of the CD45RA/CCR7 subsets defined in Section 3.1 on CD4 + T cells ( Figure 1B). We found that the highest frequency of T Reg cells was in the EM (CD45RA − CCR7 − ) population in relation to CD4 + T cells ( Figure 6F). Risk genotype (TT) carriers had a suggestively lower surface expression of CD25 on T Reg cells with an EM phenotype (p = 0.01) ( Figure 6G). Approximately no T EMRA cells had a T Reg phenotype and therefore this subset was excluded from statistical analysis. = 0.01) ( Figure 6G). Approximately no TEMRA cells had a TReg phenotype and therefore this subset was excluded from statistical analysis.

Discussion
Expression of CD25 is central in T cells responsiveness to IL-2 [7,8]. There is great heterogeneity in CD25 expression on T cell subtypes [7,8], however, the contributions from MS-associated IL2RA genetic variants are poorly understood. We have systematically explored how the MS-associated SNPs rs2104286 and rs11256593, in or near the IL2RA gene, that encode CD25, are associated with CD25 expression on T cells by multiparameter flow cytometry in genotype-selected healthy subjects.
We found that carriers of the MS-associated IL2RA risk genotype (TT) had reduced surface expression of CD25 on CD25 + CD127 + CD4 + T cells that had undergone post-thymic peripheral expansion (CD31 − CD45RA + ) or had a memory phenotype, such as CXCR5 − T H subsets or CD45RA −/lo-int cells including EM-T Reg and CD39 + CD45RA − T Reg cells. Additionally, carriers of the risk genotype (TT) had a significantly lower frequency of CD25 + T FH 1 cells and a suggestively lower frequency of CXCR5 − CD25 + T H 1 and CD25 + Th17.1 cells. Finally, we confirmed that carriers of the risk genotype (TT) had an increased frequency of CD25 + CD45RA + CD4 + T cells [15,41] and we found that this was due to an increased frequency of CD25 + CD45RA + CD4 + T cells that were recent thymic emigrants (CD31 + ).
Contrary to preceding studies [15,40,41], we only compared healthy subjects homozygous for the risk (T) or protective (C) allele, for both MS-associated SNPs rs11256593 and rs2104286, to eliminate gene-dosage effects on phenotypes previously observed for the rs2104286 SNP [15]. Furthermore, compared to preceding studies [15,40,41], we only included SNPs in or near the IL2RA gene associated with MS and did not include IL2RA SNPs associated with other autoimmune diseases which resulted in inclusion of the well-established SNP rs2104286 and the newly discovered SNP rs11256593 near the IL2RA gene from the MS-replication chip study [29]. In our cohort of a 1000 HCs only 8.3% were homozygous for the protective allele (C) for SNP rs2104286 and therefore we were not able to reproduce findings in a secondary cohort and it was only possible to determine the combined association of the MS-associated SNPs rs2104286 and rs11256593 and not the independent associations of the rs2104286 and rs11256593 SNPs on expression of CD25 without compromising the size and pairing of the genotype groups. A study determining the independent associations of rs2104286 and rs11256593 on CD25 expression are warranted. However, the selection cohort of study participants should exceed 1000 HCs as some IL2RA genotypes are rare.
Compared to previous studies, we applied a more in-depth phenotyping of T cell subsets including both CD8 + and CD4 + T cell subsets not previously studied [15,40,41]. In CD8 + T cells, we only observed a single suggestive difference although the number of individuals in each genotype group in CD8 + T cells was the same as for CD4 + T cells suggesting that MS-associated IL2RA SNPs rs2104286 and rs11256593 are associated with CD25 expression in CD4 + , but not in CD8 + T cells. This is consistent with a previous study that observed that the rs2104286 SNP controls GM-CSF production in CD4 + but not in CD8 + T cells [15]. In addition, we found MS-associated IL2RA gene variants rs2104286 and rs11256593 are associated differently with CD25 expression within CD4 + T cell subgroups. In CD4 + T cells positive for either CD31 or CXCR5, MS-associated IL2RA variants were associated with the frequency of CD25 + cells but not the MFI of CD25. Contrary, in CD4 + T cells negative for either CD31 or CXCR5 we observed several associations between MFI of CD25 and MS-associated IL2RA gene variants as well as frequency of CD25 + cells. Our findings indicate that CD25 heterogeneity within CD4 + T cell subgroups are dependent on MS-associated IL2RA gene variants and emphasizes that cell subsets need to be clearly defined by several markers in studies of MS-associated IL2RA gene variants. In CD4 + T cells, use of too few cell markers for definition of cell subsets can lead to study of mixed effects and lead to inconclusive results.
IL-2 signaling increases with expression of CD25 [7,8] and a previous study by Cerosaletti et al. reported that homozygous carriers of MS-associated IL2RA risk allele for SNP rs2104286 had reduced IL-2 receptor signaling as assessed by STAT5 phosphorylation in CD25 hi and memory CD4 + T cells [40]. Thus, these results are consistent with our observation that homozygous carriers of MS-associated IL2RA risk allele rs2104286 and rs11256593 have decreased MFI of CD25 on T Reg cells with an EM phenotype or CD39 + and CD4 + CD127 + CD25 + T cell subsets being CD45RA −/lo or expressing CXCR3 and/or CCR6, markers preferentially expressed on memory CD4 + T cells [49]. Cerosaletti et al. did examine if the observed lower STAT5 phosphorylation was due to decreased surface expression of CD25 but contrary to our results found no difference between carriers of the risk and protective allele for SNP rs2104286 [40]. However, Cerosaletti et al. performed stainings on cells from not paired HCs that had undergone a freeze-thaw process while we stained freshly isolated PBMCs from paired HCs. A previous study found that CD4 + memory cells undergone a freeze-thaw process due to cryopreservation had between 28-62% lower expression of CD25 compared to CD4 + memory cells analyzed in fresh whole blood [41] which can explain the discrepancy between the studies. In addition, pairing of the samples as performed in our study reduces day-to-day variation in staining and is more able to identify associations between MS-associated IL2RA gene variants and CD25 expression in cell subsets where cell frequency is age dependent. Finally, the discrepancy between Cerosaletti et al. and our study can be due to difference in cell markers applied to identifying cell subsets and subsequent gating strategy [40].
Analyzing freshly isolated PBMCs limits our study to a steady-state setting and cannot address the extent to which MS-associated IL2RA gene variants rs2104286 and rs11256593 influence T cells IL-2 signaling and effector responses upon activation without compromising our studies ability to identify genotype associations on ex vivo CD25 surface expression and frequency. CD25 expression correlate with T cells responsiveness to IL-2 [7,8], a central cytokine for T Reg cells ability to sustain suppressive function and modulates effector responses and T helper lineage commitment [1]. It is a possibility that in a competing immune environment the reduced expression of CD25 in carriers of the IL2RA risk genotype may inhibit the CD4 + T cells' ability to benefit from the shared pool of IL-2 and as a result promote CD4 + T cells normally inhibited by IL-2 such as T H 17 and T FH cells [12,13,18,19], and inhibit cells normally promoted by IL-2 such as T H 1 cells and T Reg cells [1,12,13]. Furthermore, it is well-established that risk allele carriers of the MS-associated IL2RA SNP rs2104286 have increased concentration of soluble CD25 in sera [42]. Studies have found that soluble CD25 inhibit the action of IL-2 on T cells [39] and may have a functional implication for modulation of T cell responses in MS [60]. In relation to our observation of decreased surface bound CD25, it might be a possibility that carriers of the IL2RA risk genotype have more unstable CD25 that leads to the lower levels of surface bound CD25 and higher levels of soluble CD25 and this exerts a dual effect on IL-2 responsiveness in CD4 + T cells. As mentioned, the present study is not able to address these questions, however future studies should address how the reduced CD25 expression observed in carriers of the MS-associated IL2RA risk allele affects CD4 + T cells IL-2 responsiveness and whether this changes T helper cells effector functions and T Reg cells suppressive capacity and if there is a dual effect with soluble CD25. Interestingly, in carriers of the IL2RA MS-associated risk allele we observed a significant reduction in the frequency of CD25 + T FH 1 cells and a suggestive reduction in the frequency of CD25 + T H 1 and Th17.1 cells indicating a possible shift in effector functions. Furthermore, such studies could also provide new insight to if and how MS-associated IL2RA SNPs contribute to the impaired suppressive capacity of CD39 + T Reg cells [33], the reduced frequency of CXCR5 + T FH 1 cells [37] and the promotion of pro-inflammatory CD4 + T cells observed in MS patients [31].
In summary, we found that MS-associated IL2RA SNPs rs2104286 and rs11256593 are associated with CD25 expression on CD4 + , but not in CD8 + T cells. In CD4 + T cells, MS-associated IL2RA gene variants are associated differently within subsets where carriers of the MS-associated risk allele (TT) show decreased surface expression of CD25 on memory CD127 + CD25 + and T Reg subsets, increased frequency of CD25 + RTE T cells and decreased frequency of CD25 + T FH 1 cells. These changes in CD25 expression may influence the cells IL-2 signaling thus affecting CD4 + T cell differentiation and T Reg cells suppressive activity.