CISH promoter polymorphism effects on T cell cytokine receptor signaling and type 1 diabetes susceptibility

Background Impaired regulatory T cell immunity plays a central role in the development of type 1 diabetes (T1D). Interleukin-2 receptor (IL-2R) signaling is essential for regulatory T cells (TREG), and cytokine-inducible SH2-containing protein (CIS) regulates IL-2R signaling as a feedback inhibitor. Previous studies identified association of CISH promoter region single nucleotide polymorphisms (SNPs) with susceptibility to infectious diseases. Methods Here we analyzed allele frequencies of three CISH SNPs (i.e., rs809451, rs414171, rs2239751) in a study of T1D patients (n = 260, onset age < 5 years, duration > 10 years). Minor allele frequencies were compared to a control cohort of the 1000 Genomes Project. Assigned haplotypes were determined for effects on T1D manifestation and severity. Finally, the CISH haplotype influence on cytokine signaling and function was explored in T cells from healthy donors. Results We detected similar minor allele frequencies between T1D patients and the control cohort. T1D onset age, residual serum C-peptide level, and insulin requirement were comparable between different haplotypes. Only minor differences between the haplotypes were found for in vitro cytokine (i.e., IL-2, IL-7)-induced CIS mRNA expression. STAT5 phosphorylation was induced by IL-2 or IL-7, but no differences were found between the haplotypes. TREG purified from healthy donors with the two most common haplotypes showed similar capacity to inhibit heterologous effector T cells. Conclusions This study provides no evidence for an association of CISH promoter SNPs with susceptibility to T1D or severity of disease. In contrast to previous studies, no influence of different haplotypes on CIS mRNA expression or T cell-mediated functions was found. Electronic supplementary material The online version of this article (10.1186/s40348-018-0080-7) contains supplementary material, which is available to authorized users.


Background
Type 1 diabetes (T1D) is a chronic autoimmune disease characterized by the destruction of pancreatic ß cells. Autoreactive effector T cells are involved in T1D pathogenesis, and impaired regulatory T cell (T REG ) functions promote self-reactive effector T cells [1,2]. T REG crucially depend on IL-2, and IL-2 receptor (IL-2R) variants are associated with increased susceptibility to T1D [3]. T1D-associated IL-2R variants affect T REG phenotype and function [4]. Therefore, decreased IL-2R signaling is assumed to be central for impaired T REG function in T1D pathogenesis [5]. IL-2R expression and signaling of T cells are tightly regulated, and cytokine-inducible SH2containing protein (CIS), a member of the suppressor of cytokine signaling (SOCS) family, contributes as a feedback inhibitor [6]. The CIS protein (encoded by the CISH gene) plays a role in T cell activation and cytokine-induced proliferation [7]. Several cytokines (including IL-2 and IL-7) induce CIS expression during receptor binding mediated by the Jak/STAT pathway [8]. CIS expression is induced within few hours and inhibits IL-2R signaling by binding of the IL-2Rβ chain and blocking of STAT5 phosphorylation [9,10].
Previous studies demonstrated important roles of CIS on T cell function during allergic, malignant, and infectious diseases [11][12][13]. These studies showed that CIS regulated T cell activation and polarization of CD4 + and CD8 + T cells [12,13]. Furthermore, genetic CISH variants were found to be associated with susceptibility to infectious diseases including tuberculosis [11,14,15]. Tuberculosis-associated single nucleotide polymorphisms (SNPs) (i.e., rs809451, rs414171, rs2239751) are located in CISH promoter regions and were accompanied with differential CIS expression [11,16]. Tuberculosis risk allele (i.e., rs414171T/rs809451C) carriers had decreased CIS mRNA expression after IL-2 in vitro stimulation and increased IL-10 serum levels [16]. This suggested increased T REG function of tuberculosis risk allele carriers due to decreased CIS expression [16].
In the present study, we addressed the question if CISH SNPs are associated with susceptibility to develop T1D and/or with age at onset and disease severity. We determined minor allele frequencies (MAFs) of three CISH SNPs in patients with early-onset and long-term T1D and compared results with controls from the 1000 Genomes database (www.internationalgenome.org) [17]. In addition, we characterized the effects of CISH SNPs on cytokine-induced T cell functions.

Patients and healthy controls
Two hundred sixty patients with T1D (mean age 16.4 years, range 10.7-20.9 years; onset age < 5 years, > 10 years diabetes duration) were recruited for the pediatric diabetes biobank within the German Center for Diabetes Research (DZD). Further information on selection and characteristics of the study cohort has been given previously [18]. Healthy adult donors were recruited from the staff working at the University Hospital Duesseldorf (n = 14).

Genotyping of CISH SNPs
DNA was isolated from peripheral blood using QIAamp DNA Mini Kit (Qiagen). TaqMan assays (Applied Biosystems) were used for genotyping of three CISH SNPs (i.e., rs809451, rs414171, rs2239751). CISH SNP inheritance was analyzed, and linkage disequilibrium was estimated using a publically available tool (https://analysistools.nci.nih.gov/LDlink). The structure of the CISH gene region and SNP linkage are depicted in Fig. 1a. Minor allele frequencies (MAFs) were calculated and compared with the European (EUR) cohort from the 1000 Genomes Project [17]. CISH SNP variants were assigned to three haplotypes based on previous reports [14,16]. Because of the predefined number of available biobank samples, we performed post hoc power calculations and included confidence intervals for MAF differences (according to [19]) (Additional file 1: Figure S1 and Table 1).

Parameters of T1D severity
Residual random serum C-peptide levels and daily insulin requirement were used as proxies for disease severity. C-peptide was measured with a high-sensitivity assay [18]; data on insulin requirement were retrospectively obtained from the German/Austrian nationwide DPV registry [18].

In vitro T cell assays
For quantification of cytokine-induced CIS mRNA expression and STAT5 phosphorylation, peripheral blood mononuclear cells (PBMCs) from healthy donors genotyped for CISH SNPs rs809451, rs414171, and rs2239751 were isolated and stored in liquid nitrogen. Samples were then concomitantly thawed for in vitro assays. For measurement of CIS mRNA expression, PBMCs were stimulated with IL-2 (100 IU/ml, Sigma-Aldrich) or IL-7 (10 ng/ml, Sigma-Aldrich) for 1 or 2 h. mRNA was then isolated and reverse transcribed. CIS . Dependency in inheritance of SNPs is indicated as linkage disequilibrium coefficients (R 2 , estimated according to https://analysistools.nci.nih.gov/LDlink). b Three respective haplotypes were assigned (H1/H2/H3) mRNA was quantified by real-time quantitative (q)PCR as described previously [20]. A non-stimulated PBMC sample was used to determine ex vivo CIS expression in PBMC. GAPDH was used as a qPCR house-keeping control [20].

Statistical analysis
MAFs for CISH SNPs were calculated according to standard methods and compared to the European (EUR) cohort from the 1000 Genomes Project [17] using Fisher's exact test; 95% confidence intervals (CI) were calculated with the Newcombe method. The Mann-Whitney U test was applied to compare disease characteristics (i.e., onset age, C-peptide level, insulin dose) between haplotypes. Further, the Friedman test was used to evaluate the effect of IL-2 and IL-7 stimulation on CIS mRNA expression and pSTAT5 induction between haplotypes as well as for the effect of different T REG /T EFF ratios on T EFF proliferation. Two-tailed p values below 0.05 were considered statistically significant. All analyses and figure preparations were performed with GraphPad Prism (Version 7.0a. GraphPad Software) and SAS for Windows version 9.4 (SAS Institute, Cary, North Carolina, USA).
Previous studies described impaired IL-2 induced CIS expression of T cells with H2 [11,16]. We applied IL-2 and IL-7 in vitro T cell stimulation for comparison of samples from healthy individuals assigned to the three genotypes (i.e., homozygous H1/H1, H2 carrier (H1/H2 and H2/H2), and H3 carrier (H1/H3); Fig. 3). Because of the low frequency of homozygous H2 carriers, only one homozygous H2 donor was included. In vitro stimulation with IL-2 or IL-7 for 1 and 2 h induced an increase of CIS mRNA expression (both p < 0.001, Fig. 3a). We detected no differences between the three genotypes 1 or 2 h after stimulation.
To determine functional effects of CISH haplotypes on cytokine receptor signaling, we measured IL-2 and IL-7 induced STAT5 phosphorylation for the same study groups (Fig. 3b). Both IL-2 and IL-7 increased pSTAT5 (all p < 0.001) without perceptible differences between the three genotypes (Fig. 3b).
Finally, we compared T REG from the two dominant genotypes H1/H1 and H1/H2 for their in vitro ability to limit T EFF proliferation. Different proportions of T REG were co-cultured with allogenic effector T cells, and effector cell proliferation was measured after 5 days (Fig. 4a). A significant reduction of T EFF proliferation was detected when T REG were added and the effects increased with higher T REG proportions (p = 0.005, Fig. 4b). However, comparable T REG effects were found for samples from H1/H1 and H1/H2 carriers. We concluded that no association of CISH SNPs with susceptibility to T1D was detectable and that haplotypes had no influence on IL-2/IL-7 signaling or T REG functions.   (H1/H1, circles), H2 haplotype carriers (H1/H2 and H2/H2, triangles), and H3 haplotype carriers (H1/ H3, squares) were characterized regarding age at manifestation (years) (a), residual C-peptide level (pmol/l) (b), and daily insulin dosage per kilogram body weight (IU/kg/day) (c). Every symbol represents the data from an individual patient, medians are indicated. No statistically significant differences were detected between study groups (Mann-Whitney U test, two-tailed)

Discussion
This study showed that CISH promoter SNP rs809451, rs414171, and rs2239751 frequencies were comparable between patients with early onset of T1D and controls from the publically available 1000 Genomes database. Therefore, no evidence for an association of CISH promoter variants and susceptibility to develop T1D was detected. Available cohort sizes restricted the sensitivity of this study and moderate effects of CISH SNPs cannot be excluded. In addition, the 1000 Genomes Project cohort comprised individuals from different European regions and, therefore, effects due to ethnic differences between study groups are possible [17].
We detected no differences in T1D manifestation age between different haplotypes. However, all T1D patients included in the present study had an early onset of T1D below 5 years of age and this preselection limited differences in onset age between haplotypes. Differences in disease course are indicated by residual ß-cell activity-measured by residual C-peptide levels-and daily insulin requirements of T1D patients. We detected no differences between the haplotypes for these parameters. Thus, these results do not suggest haplotype-dependent effects on autoimmune mechanisms and disease course in T1D patients.
Previous studies suggested a role of CISH promoter SNPs on CIS mRNA expression [11,16]. Whereas Khor b a Fig. 3 Cytokine induced CIS mRNA expression and STAT5 phosphorylation according to CISH haplotypes. a CIS mRNA expression of PBMCs from healthy adult donors with distinct haplotypes (circles: homozygous H1 carriers (H1/H1), triangles: H2 carriers (H1/H2, H2/H2), squares: H3 carriers (H1/H3)) treated with IL-2 (left graph) or IL-7 (right graph) for 1 and 2 h. CIS mRNA levels-calculated in comparison to the housekeeping gene GAPDH-are normalized against time point 0 for each donor. Median and IQR of 2 −ΔΔCT values are indicated. p value for effect of IL-2/IL-7 stimulation on CIS mRNA levels considering all genotypes (Friedman test) is indicated as *** for p < 0.001. For IL-2 stimulation, 5 H1/H1, 5 H2, and et al. detected CIS mRNA expression differences only for the homozygous minor allele carriers (H2/H2 according to the nomenclature used here) [11], Sun et al. also detected decreased CIS mRNA for heterozygous (H1/H2) patients carrying the rs809451 minor allele [16]. Only one H2/H2 carrier was found among healthy individuals recruited and, therefore, we were not able to reproduce the results for the homozygous H2 genotype. In contrast to Sun et al. [16], we detected no differences in CIS mRNA for H1/H2 carriers as compared to H1/ H1 carriers. In addition, we did not detect H1/H2 effects on IL-2 induced STAT5 phosphorylation. Differences between rs809451 and rs414171 minor allele carriers described [16] could not be reproduced in the present study because of the strong linkage disequilibrium between these SNPs.
In addition, the study by Sun et al. found differences in IL-10 serum levels of rs414171 minor allele carriers [16]. These finding suggested functional effects of CISH haplotypes on T REG , the main producers of IL-10 [21]. Our functional T REG analyses of H1/H1 and H1/H2 carriers did not indicate differences in their ability to limit T EFF proliferation. Since other immune cell populations (e.g., follicular T helper cells) produce IL-10 [22], described differences in IL-10 serum levels may be explained by non-T REG -mediated mechanisms.
Our results do not suggest a role of CISH SNPs in the susceptibility to T1D, whereas other reports showed an association with susceptibility to infectious diseases [11]. CIS is involved in the regulation of several cytokines including IL-7, which is crucial for naïve and memory T cell homeostasis and may promote autoreactive T cell responses in T1D [23]. Therefore, a possible role of CIS in T1D may be more complex and not based on impaired cytokine-induced CIS expression of different haplotypes. In accordance, multiple roles of CIS in different immune cell populations have been described [12,13,24]. Further analyses are needed to address this question.

Conclusions
This study provides no evidence for an association of CISH promoter SNPs with susceptibility to T1D or severity of the disease. In contrast to previous studies that demonstrated marked effects of CISH SNPs on T cells, no influence of different haplotypes on CIS mRNA expression or T cell-mediated functions was found. Further studies are needed to address the question, how CISH a b Fig. 4 T REG suppression assay. a Representative depiction of CFSE stainings of allogenic effector T cells (T EFF ) incubated without regulatory T cells (T REG ) (left chart), with T REG from an H1/H1 carrier (middle chart) and T REG from an H1/H2 carrier (right chart). Percentages of proliferated T EFF cells are indicated. b Proliferation rate of T EFF (%) coincubated with T REG from H1/H1 carriers (black bars, n = 2) and from H1/H2 carriers (gray bars, n = 2) in different T EFF /T REG ratios. Median and range are indicated. The dotted line indicates proliferation rate of T EFF incubated without T REG . p value for the effect of T EFF /T REG titration considering both haplotypes (Friedman test) is indicated as ** for p = 0.005 variants mechanistically carry out their role during allergic, malignant, and infectious diseases.

Funding
This study was supported in part by a grant from the German Federal Ministry of Education and Research (BMBF) to the DZD (pediatric diabetes biomaterial bank, 01GI1109B and 82DZD01801, Early-onset T1D cohort study, 01GI0802, 01GI1109A, 82DZD00201).

Availability of data and materials
The datasets generated and analyzed during the current study are available from the corresponding author on request.
Authors' contributions JS contributed to conceptualization of the study, performed experiments, analyzed and interpreted data, and wrote the manuscript. HA performed experiments, analyzed data, and reviewed and edited the manuscript. TM, JR, and CB coordinated the Pediatric Diabetes Biobank activities for patients with early-onset and long-term T1D within the German Center for Diabetes Research (DZD e. V.). RWH coordinated the German pediatric diabetes registry and the pediatric biobank. EM, TM, JR, CB, RWH, and MR discussed results and reviewed/edited manuscript. MJ designed the study, analyzed, and interpreted data and wrote the manuscript. All of the contributing authors approved the final version of the manuscript.

Ethics approval and consent to participate
The study was approved by the Ethical Committee of the Medical Faculty of the Heinrich Heine University Duesseldorf, Germany (reference number 3431 and 4844). Sample analyses were approved by the biobank consortium. Written informed consent was received from all donors (older than 14 years) and the children's legal guardians.

Consent for publication
Not applicable.