Introduction

The important role of the homeostatic cytokine interleukin-7 (IL-7) in the development of autoimmune diseases is well established [1]. Especially T cells depend on IL-7 availability; in the presence of IL-7, self-reactive CD4+ T cells license dendritic cells to induce type 1 diabetes in NOD animals [2]. IL-7 receptor blockade has been shown to revert immune pathology by inhibiting effector T-cell functions in the same type 1 diabetes model [3]. Regulation of IL-7 availability and target cell sensitivity is exerted on the IL-7 receptor level [4]. In accordance, autoimmune disease-associated genetic polymorphisms of this pathway are located in the IL7RA gene locus (coding for the alpha-chain of the IL-7 receptor, IL-7Rα) [5, 6]. Different IL-7 receptor single nucleotide polymorphisms (SNPs) were identified, which affect IL-7 receptor protein expression and alternative splicing (reviewed in [7]). The best characterized IL-7 receptor SNP (i.e., rs6897932) affects exon 6 splicing and expression of the soluble IL-7 receptor (sIL-7R) [5]. The rs6897932 T-allele is associated with protection against autoimmune diseases and causes diminished sIL-7R expression [5]. As a consequence, protective allele carriers generally have lower sIL-7R serum levels with potential effects on sIL-7R reservoir function and IL-7 availability [8, 9]. We have previously found using cell lines that the rs6897932 T-allele also affected membrane IL-7R (mIL-7R) expression [10]; however, differences of mIL-7R expression on adaptive immune cell subsets between different IL7RA allele carriers were not found [11]. Recently, Al-Mossawi et al. showed that mIL-7R expression is increased in human monocytes on LPS-induced activation indicating potential functional effects of IL-7 for monocytes [12]. Notably, the induced mIL-7R expression level was higher on monocytes from protection-associated rs6897932 T-allele carriers [12]. These results suggested a potential role of IL-7 mediated monocyte functions for autoimmune susceptibility.

The relevance of myeloid antigen-presenting cells (APCs) (including monocytes, macrophages, dendritic cells (DC)) for type 1 diabetes pathogenesis is well established [13]. Type 1 diabetes animal models showed that especially DC subsets located in the pancreatic islets contribute to immune pathogenesis of type 1 diabetes [14], but also resident macrophages and macrophages derived from peripheral blood monocytes may play a role [15]. Differences in the phenotype of APC subsets were described and differential polarization of APCs (i.e., classical monocytes; alternative monocytes) was identified in distinct pancreatic islets regions [16]. The role of APCs in autoimmunity resembles a double-edged sword. On the one hand, APCs are central inducers of T-cell tolerance during T-cell development as well as in the periphery. On the other hand, self-reactive T cells critically depend on effective antigen-presentation by APCs. The mechanisms underlying tolerance failure in autoimmunity are multifaceted and intensive previous research revealed relevant factors. APC-secreted factors, which affect T-cell function such as proinflammatory cytokines or type 1 IFNs, were shown to play a role in human type 1 diabetes [17-22]. In addition, differentially expressed cytokines, like GM-CSF, may also affect APC functions in type 1 diabetes [23, 24]. Direct APC T-cell interaction is potentially involved since T-cell activation requires MHC/epitope recognition and co-receptor binding on APCs. The important coreceptor CD40, which is a target of the CD40 ligand (CD40L), is upregulated by T cells on activation. CD40/CD40L interaction is mandatory for the development of an effective CD8 T cell response [25, 26]. Notably, interference with CD40/CD40L binding ameliorated type 1 diabetes in the NOD animal model [27].

The expression of the IL-7R on monocytes (expressed as a heterodimer composed of the alpha-chain and the common-gamma chain) has been largely neglected, likely because the IL-7Rα is also part of the Thymic Stromal Lymphopoietin (TLSP) receptor expressed on monocytes [28]. Thus, IL-7 effects on APCs are not well characterized. Early studies investigated IL-7-mediated tumoricidal effects and found increased IL-7-induced monocyte cytokine expression [29]. Others showed IL-7 promoted antimicrobial monocyte effects [30, 31]. Recently, the IL-7R was found to affect the development of tissue-resident macrophages during fetal myelopoiesis [32]. Evidence for a potential role of IL-7 effects on monocytes in autoimmune diseases comes from studies on patients with rheumatoid arthritis [reviewed in 33]. However, there is still a lack of studies about the relevance of IL-7 mediated effects on monocytes for type 1 diabetes disease pathology.

To address this topic, we characterized IL-7 mediated effects on different monocyte subpopulations as well as the effects of IL-7 on monocyte phenotype and function in children with type 1 diabetes. Ex vivo phenotyping and in vitro culture experiments using peripheral blood immune cells from type 1 diabetes patients and matched controls were performed.

Results Lower IL-7R expression on type 1 diabetes patients’ monocytes in the presence of IL-7

Moderate ex vivo IL-7R expression on monocytes and stimulation-dependent upregulation in vitro have been described [12]. Against this background, we determined IL-7R expression of monocytes ex vivo and after stimulation with LPS or IL-7 for 24h in type 1 diabetes patients and matched controls (Fig. 1). In the absence of stimulation, in vitro culture increased IL-7R expression on monocytes (median MFI: 346, IQR: 317–396) and this effect was strongly enhanced by LPS stimulation (median MFI: 702, IQR: 582–838) (Fig. 1A and B). Slightly lower IL-7R expression was found in the presence of IL-7 (median MFI: 334, IQR: 289–390) as compared to non-stimulated samples (p = 0.03) (Fig. 1B). Interestingly, IL-7 mediated effects on monocyte IL-7R expression differed between the study groups. Type 1 diabetes patients (median MFI: 310, IQR: 256–336) had significantly lower IL-7R levels on monocytes exclusively in the presence of IL-7 as compared to controls (median MFI: 368, IQR: 324–402; p = 0.021) (Fig. 1C). Since IL-7R expression varied between individuals, we performed pairwise comparisons of IL-7R expression between IL-7- and non-stimulated monocytes. Whereas we did not find differences in controls, significantly lower IL-7R expression was detected for type 1 diabetes patients in the presence of IL-7 (p = 0.004; Fig. 1D). To estimate the influence of ex vivo IL-7R expression, we calculated IL-7R differences between IL-7 treated monocytes and monocytes ex vivo for individuals from both study groups. No differences between the study groups were detected without stimulation (Fig. 1E). In contrast, IL-7 induced differences of IL-7R expression in monocytes were lower in type 1 diabetes patients, suggesting decreased sensitivity to IL-7 stimulation in type 1 diabetes patients (Fig. 1E). These results confirmed specific IL-7 effects on monocyte IL-7R expression and differences between type 1 diabetes patients and controls.

IL-7Rα expression on monocytes ex vivo and after in vitro PBMC culture with LPS, IL-7, or without. (Geometric) MFI values determined by flow cytometry are given. Each symbol represents the mean value of duplicates for an individual donor. (A) Representative plots and histograms depict the gating strategy. (B, C) IL-7Rα expression ex vivo (“0h”) and after culture with (“24h LPS,” “24h IL-7”) or without (“24h”) stimulation is compared between all individuals (B) and the study groups (C). (D) IL-7Rα expression after 24h culture with or without IL-7 is shown as connected symbol plots. (E) Calculated delta values of IL-7Rα expression (ex vivo subtracted from 24h culture with or without IL-7) are shown for individual donors and compared between the study groups. (B, C, E) Violin plots indicate median, 25- and 75-percentile values as dotted lines. Nominal p-values of the Wilcoxon matched-pairs signed-rank test (B, D) and the two-tailed Mann–Whitney U-test (C, E) are given. Twenty type 1 diabetes patients and 20 controls were included. Four independent experiments, including 10 concomitantly thawed patient (n = 5) and control (n = 5) samples each time, were performed.

Mechanisms and clinical associations of IL-7 induced IL-7R regulation in monocytes

To further evaluate the mechanisms of IL-7 induced IL-7R regulation in monocytes, we compared soluble and membrane IL-7R mRNA after in vitro stimulation of healthy donors´ monocytes. While LPS strongly increased sIL-7R and membrane IL-7R mRNA, we did not see IL-7 effects on the IL-7R isoforms (Fig. 2A), not suggesting IL-7 effects on alternative splicing. In accordance, patients´ and controls´ sIL-7R serum levels were not associated with IL-7R expression after IL-7 stimulation (Fig. 2B). Moreover, we analyzed potential associations between type 1 diabetes patients’ clinical characteristics with monocyte IL-7 responses. We did not find associations with disease duration, age at disease onset, and HbA1c (Supporting information Fig. 1). Finally, we assessed the impact of the protection-associated single nucleotide polymorphism rs6897932 on IL-7 induced IL-7R expression on monocytes. rs6897932 allele frequencies did not significantly differ between type 1 diabetes patients and controls (Supporting information Table 1). Moreover, we did not detect differences in constitutive or IL-7 induced mIL-7R expression between study participants with different rs6897932 genotypes (Fig. 2C).

IL-7 induced membrane (m)IL-7R and soluble (s)IL-7R and association with IL7RA genotypes. (A) mRNA expression of full-length IL-7R ( = mIL-7Rα) and IL-7R lacking exon 6 ( = sIL-7Rα) from healthy controls´ isolated monocytes is shown as fold change. Fold change was calculated using the 2–∆∆ct method with GAPDH and ex vivo measured samples as references. (B) Correlation between monocyte mIL-7Rα expression after 24h IL-7 stimulation and serum sIL-7R is presented. A trend line was fitted by linear regression analysis. Spearman rank correlation coefficient r and p-value are given. (C) mIL-7Rα expression on monocytes is compared between different rs6897932 genotypes from both study groups. Each symbol represents the mean value of duplicates for an individual donor. Violin plots indicate median, 25- and 75-percentile values as dotted lines. (A) Four healthy donors were included. Two independent experiments with two donors (n = 2) were performed. (B) Eighteen controls and 20 type 1 diabetes patients were included. Four independent experiments, including 9–10 concomitantly thawed patient (n = 5) and control (n = 4-5) samples, were performed. (C) Twenty controls and 19 type 1 diabetes patients were included. Four independent experiments, including 9–10 concomitantly thawed patient (n = 4-5) and control (n = 5) samples, were performed.

Polarized monocyte subpopulations differ in IL-7R expression and IL-7 response between the study groups

Next, we analyzed differentially polarized monocyte subpopulations for IL-7R expression and IL-7 in vitro response. Classical (CD14+; termed cMono), nonclassical (CD16+; termed ncMono), and intermediate (CD14+/CD16+; termed intMono) monocyte subpopulations were compared (Fig. 3A). cMono (median MFI: 132, IQR: 116–151) and intMono (median MFI: 132, IQR: 118–151) subpopulations showed higher ex vivo IL-7R median expression levels as compared to ncMono (median MFI: 108, IQR: 96–124; p < 0.001) (Fig. 3B). No differences between the study groups were detected (data not shown). In vitro culture increased IL-7R expression on all subpopulations (p < 0.001; Fig. 3C). Study group comparisons revealed comparable IL-7R expression of monocyte subpopulations without stimulation (Fig. 3D). In the presence of IL-7, however, we detected lower median IL-7R expression for monocyte subpopulations from type 1 diabetes patients and this was significant for ncMono (p = 0.041; Fig. 3E). These results indicated IL-7 specific ncMono population differences for type 1 diabetes patients.

Monocyte subpopulation IL-7Rα expression ex vivo and after stimulation. Each symbol represents the mean value of duplicates for an individual donor. (A) Representative zebra plots indicate the gating strategy for monocyte subpopulations (i.e., cMono: CD14+/CD16–; intMono: CD14+/CD16+; ncMono: CD14–/CD16+). (B) Ex vivo IL-7Rα expression on monocyte subpopulations is shown. (C) Connected symbol plots of IL-7Rα expression on monocyte subpopulations compared between ex vivo (“0h”) and 24h cell culture are presented. (D, E) Study group comparisons of IL-7Rα expression on monocyte subpopulations after 24h cell culture without IL-7 (D) and with IL-7 (E) are shown (controls, squares; type 1 diabetes patients, circles). (B, D, E) Violin plots indicate median, 25- and 75-percentile values as dotted lines. Nominal p-values of the two-tailed Mann–Whitney U-test (B, D, E) and the Wilcoxon matched-pairs signed-rank test (C) are given. Nineteen type 1 diabetes patients and 16 controls were included. Four independent experiments, including 6–10 concomitantly thawed patient (n = 2-5) and control (n = 4-5) samples each time, were performed.

IL-7 promotes ncMono in vitro

To further characterize monocyte subpopulations and IL-7 function, we compared ex vivo and in vitro stimulated cMono, ncMono, and intMono proportions between the study groups. Type 1 diabetes patients had ex vivo significantly higher cMono proportions (median: 88.4%, IQR: 85.4–91.3%) and lower intMono proportions (median: 6.8%, IQR: 6.0–8.7%,) as compared to controls (cMono: median: 85.5%, IQR: 77.0–87.9%, p = 0.046; intMono: median: 10.1%, IQR: 7.7-12.1%, p = 0.02; Fig. 4A). ncMono proportions were comparable between the study groups (Fig. 4A). In vitro culture markedly changed monocyte subset distributions: Especially intMono proportions increased, whereas cMono and ncMono proportions decreased within 24h (Supporting Information Fig. 2 and Fig. 4B). No differences between the study groups were detected after 24h (Supporting Information Fig. 2). IL-7 stimulation had only marginal effects on cMono and intMono proportions (Fig. 4B; left and middle graph) but resulted in increased ncMono proportions as compared to non-stimulated samples (Fig. 4B; right graph). Notably, IL-7 effects were different between the study groups. ncMono proportions were significantly lower in IL-7 treated type 1 diabetes patients’ samples (median: 2.8%, IQR = 1.7–4.4%) as compared to controls (median: 3.9%, IQR: 2.8–6.3%; p = 0.044) (Fig. 4C; right graph), whereas no differences for cMono and intMono proportions (Fig. 4C; left and middle graph) or in the absence of IL-7 were seen (Supporting Information Fig. 3). Calculation of individual differences between IL-7- and non-treated samples confirmed IL-7 specific increase of ncMono proportions in controls (median: 1.4%, IQR: 0.45–2.2%), whereas type 1 diabetes patients had similar ncMono proportions with or without IL-7 (median: 0.28%, IQR: −0.16–1.09%; Fig. 4D).

Monocyte subpopulation changes during in vitro culture and IL-7 effects on subpopulation distribution and counts. (A) Monocyte subpopulation distribution for both study groups ex vivo. (B) Individual cMono, intMono, ncMono proportions after 24h cell culture with and without IL-7 are shown as connected symbol plots. (C) Study group-specific IL-7 effects on monocyte subpopulation proportions. (D) Calculated delta values (24h subtracted from 24h culture with IL-7) of ncMono proportions are shown for both study groups. (E) Individual monocyte counts after 72h in vitro culture with or without IL-7 are presented. (F, G) Individual monocyte subpopulation counts after 72h in vitro culture with or without IL-7 are shown as calculated fold-change values (F) and connected symbol plots (G). (A) Median and IQR are given. Significant differences between study groups calculated by the two-tailed Mann–Whitney U-test are indicated as * for p < 0.05. (C, D) Violin plots indicate median, 25- and 75-percentile values as dotted lines. Nominal p-values of the two-tailed Mann–Whitney U-test (C, D, F) and the Wilcoxon matched-pairs signed-rank test (B, E, G) are given. (A–D) Nineteen type 1 diabetes patients and 16 controls were included. Four independent experiments, including 6–10 concomitantly thawed patient (n = 2-5) and control (n = 4-5) samples each time, were performed. (E–G) Fourteen type 1 diabetes patients and 11 controls were included for count experiments. Three independent experiments, including 6–10 concomitantly thawed patient (n = 2-5) and control (n = 4-5) samples each time, were performed.

To analyze potential IL-7 effects on monocyte survival, we determined monocyte subpopulation counts after 72h of in vitro culture. IL-7 stimulated samples showed no general differences in monocyte numbers as compared to untreated samples (Fig. 4E). However, subpopulation comparisons revealed increased ncMono and cMono numbers (p < 0.001 and p < 0.001) whereas intMono numbers decreased in the presence of IL-7 (Fig. 4F and G). The relative increase was strongest in ncMono (median: 4.6-fold) as compared to cMono (median: 1.6-fold) (Fig. 4F). No differences between the study groups were detected for cell counts (data not shown).

IL-7 induced monocyte subsets are characterized by distinct marker expression pattern

Next, we characterized IL-7 stimulated monocytes by multi-color flow cytometry for described activation and maturation markers. LPS stimulation has been included in these experiments as a strong activator of monocytes. To determine the level of similarity of IL-7-sensitive monocytes and influential factors, we performed t-SNE analyses including several monocyte markers (i.e. CD40, CD80, CD86, IL-7R) (Fig. 5A). t-SNE plots depict similarity of cells on the basis of included markers using two calculated components. A two-dimensional graph was generated, and each dot represents a monocyte (from one respective individual). The smaller the distance between two monocytes, the greater their similarity. Initially, we compared IL-7Rhigh (gated on top 5% of unstimulated cells) and IL-7Rlow (lowest 5%) monocytes for the different stimulations. Whereas IL-7Rlow cells were dispersed and did not show high similarities, IL-7Rhigh monocytes were found in delimitable regions (Fig. 5A). When stimulated with IL-7, three main IL-7Rhigh populations were identified, which showed high similarities, but were clearly distinct from each other (Fig. 5A). In contrast, LPS and non-stimulated monocytes had no distinct subsets (Fig. 5A).

Monocyte activation marker differences after in vitro stimulation of PBMC from type 1 diabetes patients and controls. Stimulations with IL-7, LPS, or without for 24h are shown. (A) t-distributed Stochastic Neighbor Embedding (t-SNE) analysis of CD11b positive monocytes after in vitro culture (left plot) and LPS or IL-7 stimulation (middle and right plot). t-SNE analysis was based on flow cytometry measurements of IL-7Rα, CD80, CD86, and CD40. Each dot represents a single cell and the distance between dots correlates with similarity. IL-7Rαhigh and IL-7Rαlow cells (gated on top or bottom 5% of IL-7Rα MFI of unstimulated cells) are marked by dark and bright blue color, respectively. Histograms indicate IL-7Rα, CD80, CD86, and CD40 expression for IL-7Rαhigh and IL-7Rαlow subsets after IL-7 stimulation. (B) HLA-DR, CD40, CD80, and CD86 expression ex vivo and after stimulation. (C) Study group comparisons of CD40 expression ex vivo and after in vitro culture with and without stimulation. (B, C) Violin plots indicate median, 25- and 75-percentile values as dotted lines and connected symbol plots visualize individual expression with and without IL-7. Nominal p-values of the Wilcoxon matched-pairs signed-rank test (B) and the two-tailed Mann–Whitney U-test (C) are given. Twenty type 1 diabetes patients and 20 controls were included. (B: CD40, CD80, CD86) Four independent experiments, including 10 concomitantly thawed patient (n = 5) and control (n = 5) samples each time, were performed. (B: HLA-DR) Nineteen type 1 diabetes patients and 16 controls were included. Four independent experiments, including 6–10 concomitantly thawed patient (n = 2-5) and control (n = 4-5) samples each time, were performed.

Comparison between IL-7 induced IL-7Rhigh and IL-7Rlow monocytes suggested differences in monocyte marker CD40, CD80, and CD86 expression (Fig. 5A, histograms). Hence, we compared these markers (plus HLA-DR included from a separate panel) on monocytes after different stimulations. Like for the IL-7R, all markers showed increased expression after 24h of culture and different stimuli affected expression pattern (Fig. 5B). LPS strongly enhanced HLA-DR and CD80 but diminished CD86 expression (all p < 0.001) (Fig. 5B). IL-7 also enhanced HLA-DR (although not as strong as LPS; p < 0.001) and, to a small extent but significantly diminished CD80 (p < 0.001) and showed a trend toward increased CD86 expression (p = 0.048) (Fig. 5B). Notably, CD40 expression was markedly enhanced by IL-7, whereas LPS had no detectable effects on CD40 expression (Fig. 5B). Comparison between the study groups revealed differences for CD40 in type 1 diabetes patients showing lower CD40 expression levels ex vivo and after culture (Fig. 5C). Only in the presence of IL-7, study group differences were not significant and concomitantly showed high variance in CD40 expression (Fig. 5C).

IL-7-induced CD40high monocytes show phenotype differences between type 1 diabetes patients and controls

Since CD40 was specifically induced by IL-7 and showed variability as well as differences between study groups, we next characterized CD40high and CD40low monocytes after IL-7 stimulation. Both study groups were compared for CD40high and CD40low monocytes. CD40high monocytes had generally higher CD80, CD86, and IL-7R expression as compared to CD40low monocytes (p < 0.001 for all antigens and both study groups) (Fig. 6A). For CD80 and CD86, no differences between the study groups were found in CD40high and CD40low monocytes (Fig. 6A, left and middle graph). Notably, IL-7R expression was higher on CD40high monocytes from the control group (median: 480, IQR: 399–534) as compared to type 1 diabetes patients (median: 372, IQR: 320–433; p = 0.007) in the presence of IL-7 (Fig. 6A, right graph). No IL-7R differences were found for CD40low monocytes between the study groups (Fig. 6A, right graph). Comparison of non-stimulated monocytes confirmed IL-7 specificity of study group differences (Supporting Information Fig. 4). To characterize these differences on the individual donor level, we determined the correlation of IL-7R and CD40 expression after IL-7 stimulation for each individual from both study groups. Notably, IL-7R and CD40 were positively correlated in controls (r = 0.62, p = 0.004; Fig. 6B, left graph), whereas a negative correlation was found for type 1 diabetes patients (r = −0.60, p = 0.005; Fig. 6B, right graph). These results suggested differences in IL-7R regulation of CD40high monocytes on IL-7 stimulation between type 1 diabetes patients and controls.

CD40 expressing monocyte subpopulations after IL-7 stimulation and IL-7Rα expression differences between CD40high subsets. (A) CD80, CD86, and IL-7Rα MFI expression on CD40high and CD40low (top and lowest 10% based on MFI, respectively) monocytes treated with IL-7 is shown. Each symbol represents the mean value of duplicates for an individual donor. Violin plots indicate median, 25- and 75-percentile values as dotted lines. Nominal p-values of the two-tailed Mann–Whitney U-test are given. (B) Correlations between IL-7Rα and CD40 expression on monocytes after IL-7 stimulation are shown. A trend line was fitted by linear regression analysis. Spearman rank correlation coefficients r and p-values are given. (C) t-distributed Stochastic Neighbor Embedding (t-SNE) analysis of CD11b positive monocytes from healthy controls (upper plots) and type 1 diabetes patients (bottom plots). t-SNE analysis was based on flow cytometry measurements of IL-7Rα, CD80, CD86, and CD40. Three distinct subsets of CD40high (top 10% of CD40 MFI) monocytes were gated and marked by different colors (left panel: subset 1, blue; subset 2, orange; subset 3, green). In the middle and right t-SNE plots, CD86 as well as IL-7Rα high and low expressing monocytes are marked. Black lines indicate the position of CD40high subsets. Histograms show CD40, IL-7Rα, and CD86 expression of CD40high subsets. Samples from type 1 diabetes patients (n = 20) and controls (n = 20) were included. Four independent experiments, including 10 concomitantly thawed patient (n = 5) and control (n = 5) samples each time, were performed.

A subset of CD40high monocytes from type 1 diabetes patients is impaired for IL-7R expression

These results prompted us to characterize CD40 expressing monocytes from both study groups using t-SNE analysis. Both, type 1 diabetes patients and controls, had three distinguishable monocyte subpopulations characterized by high CD40 expression (top 10% CD40 MFI) (Fig. 6C, left plots). These subpopulations showed marked differences in CD86 expression (Fig. 6C, histograms). We identified three CD40high monocyte subsets, which were characterized by CD86high (green color, CD40high1), CD86intermediate (orange, CD40high2), and CD86low (blue, CD40high3) expression. Interestingly, overlap between CD40high and IL-7Rhigh monocytes was only detectable for CD86high but not for the CD86intermediate/CD86low monocytes from controls (Fig. 6C, upper plots). Strikingly, this IL-7Rhigh/CD86high monocyte population was absent in type 1 diabetes patients, where none of the CD40high subsets showed detectable IL-7Rhigh expression levels (Fig. 6C, lower plots). We concluded that IL-7 induced CD40high monocytes characterized by high CD86 expression were impaired for IL-7R expression in type 1 diabetes patients.

Discussion

The present study shows IL-7 specific in vitro effects on monocyte subpopulations and phenotype. Differences of IL-7R expression predominantly on CD40high monocyte subsets between type 1 diabetes patients and controls suggest a role of IL-7 dependent monocyte functions in autoimmune pathogenesis.

Classical polarized (cMono) and non-classical activated (ncMono) monocytes are hypothesized to play different roles in autoimmunity. ncMono are viewed as anti-inflammatory, whereas cMono are involved in initiation of inflammation as a potential key factor in autoimmune pathogenesis. There is strong evidence for increased inflammation and cMono activity in type 1 diabetes patients [17, 22, 34, 35]. Increased pro-inflammatory cytokine serum concentrations were found [22] and cMono were shown to spontaneously express higher IL-6 and IL-1ß levels in type 1 diabetes patients [17, 34, 35]. Here we detected differences in the monocyte subset distribution between the study groups. cMono were the predominant subset in both study groups but were significantly higher in type 1 diabetes patients. This was accompanied by decreased intermediate (CD14+/CD16+) monocyte proportions, whereas ncMono were comparable between the study groups. These monocyte proportion differences may reflect polarizing effects of the described pro-inflammatory serum environment in type 1 diabetes patients that promotes cMono [22]. We could also show that monocyte subpopulations differed in IL-7R expression. cMono and intermediate monocytes had significantly higher IL-7R expression as compared to ncMono. In vitro culture generally increased IL-7R expression of all monocyte subpopulations and this was similar between the study groups. In the presence of IL-7, however, monocyte subpopulations showed significant differences and ncMono (although expressing the lowest IL-7R levels) increased in proportion and absolute numbers. These results suggested functional effects of IL-7R expression on monocytes and indicated polarizing efficacy of IL-7 towards anti-inflammatory ncMono. This is seemingly controversial, since IL-7 availability has been associated with increased susceptibility for autoimmunity [8, 36]. However, most studies focused on IL-7 effects on T-cell response and neglected IL-7 mediated monocyte function. First evidence for protective IL-7 monocyte function against autoimmunity comes from the study of Al-Mossawi et al. [12]. They showed that carriers of the IL7RA haplotype associated with partial protection against autoimmunity have higher IL-7R expression on monocytes when stimulated with LPS [12] and this was also predicted by our cell-line based studies [10]. In addition, an IL7RA haplotype associated with increased susceptibility (rs1494555 and rs1494558) showed lower IL-7R expression [10]. On this basis, a discussion on the role of the IL-7R and its genetic polymorphisms in different diseases has been initiated [7]. In the present study, we show that type 1 diabetes patients´ monocytes had lower IL-7R expression in the presence of IL-7 and this was mainly due to lower ncMono IL-7R expression of type 1 diabetes patients. IL-7 is well known for its capacity to promote homeostatic proliferation of peripheral blood T cells [37]. The effects of IL-7 on the proliferation and polarization of monocytes and macrophages are so far unknown. Our data suggest that IL-7 is enhancing the pre-existing polarization of classical and, to the greatest extent, non-classical monocytes at the cost of reducing those with an intermediate phenotype. Against this background, we hypothesize that lower IL-7R expression on ncMono from type 1 diabetes patients negatively influences the anti-inflammatory ncMono population in type 1 diabetes. Since proportions of in vitro IL-7 treated ncMono but not of blood ncMono differed between the study groups, we assumed potential functional effects dependent on differences in IL-7R expression.

To initially address this question, we performed comprehensive phenotyping of monocytes treated with IL-7 in vitro. We identified IL-7 dependent monocyte phenotype differences as compared to the background in vitro and LPS effects. Increased HLA-DR expression and, to a minor extent, lower CD80 were identified in the presence of IL-7. Whereas LPS induced strong up-regulation of HLA-DR and CD86, CD40 expression was exclusively increased by IL-7. CD40 is well known for its role in T-cell/APC interaction and T-cell activation [38]. After T-cell receptor/HLA binding, several co-receptors form pairs in the T-cell/APC interface and affect T-cell activation. Stimulatory as well as inhibitory signals are provided (e.g., via CD80/CD86) and CD40 is bound by the ligand CD40L (CD154) expressed on activated T cells. T cells activated in the absence of CD40 show impaired recall responses and memory especially of CD8+ T cells [39]. APCs are also affected by CD40/CD40L signaling and induced IL-7 secretion by dendritic cells has been shown to depend on CD40 ligation [40]. CD40 mediated induction of IL-7 secretion by dendritic cells was important for CD8 memory generation [40] and IL-7 produced by tolerogenic dendritic cells may play a role in type 1 diabetes [41]. A potential role of CD40 in autoimmune pathogenesis has been hypothesized and increased soluble CD40L serum levels were found in type 1 diabetes patients [42]. In addition, CD40 expressing T cells were identified as a relevant factor in the NOD animal model of type 1 diabetes [43]. In this context, CD40 was shown to promote inflammation and autoimmune T cells [44].

Here, we found lower CD40 expression on monocytes from type 1 diabetes patients ex vivo and after in vitro culture. Notably, only in the presence of IL-7, which induced marked variation in CD40 expression, no difference between the study groups was detected. In accordance, IL-7 induced CD40high monocytes showed significant differences in IL-7R expression for type 1 diabetes patients having significantly lower IL-7R expression on CD40high monocytes. These results were strengthened by the characterization of three main subsets on IL-7 treated CD40high monocytes by t-SNE analyses. These subsets were marked by different CD86 expression and the CD86high/CD40high monocyte subset showed IL-7Rhigh expression only in controls but not in type 1 diabetes patients. We were not able to concomitantly include markers for cMono and ncMono in these analyses but like for ncMono, IL-7 induced IL-7Rhigh monocytes were a feature predominantly found in controls, suggesting protective effects of IL-7 mediated monocyte functions against type 1 diabetes.

Further investigations are needed to characterize IL-7 effects on monocyte subsets and to elucidate underlying mechanisms and the relevance of differential IL-7R expression. First analyses indicated no association of IL-7 induced IL-7R expression with the protective IL7RA SNP rs6897932 and no induction of the sIL-7R by IL-7. The functionality of the monocyte IL-7R receptor was shown in our earlier studies on monocytes in tuberculosis, where a decrease in mIL-7R expression from tuberculosis patients was associated with decreased ex vivo and IL-7 induced STAT5 phosphorylation [45]. Future studies should include further functional analyses to assess the effects of differential IL-7R expression in type 1 diabetes patients as well as cell-line or animal-based models. In the current study, we cannot dissect the extent of direct and indirect effects of IL-7 on monocytes, e.g. via T cells as shown by others [46]. However, there is initial evidence that IL-7 directly influences GM-CSF-dependent generation of dendritic cell-like monocytes in vitro [47]. This study showed that IL-7 (in comparison with IL-4) induced a distinct population of dendritic-like cells [47].

Taken together, our study suggests a potential role of differential IL-7R expression and IL-7 function of monocytes in type 1 diabetes patients. This provides indications of how autoimmunity-associated genetic polymorphisms of the IL-7 receptor functionally affect disease susceptibility. It remains to be investigated how IL-7 is specifically involved in relevant monocyte and APC function in type 1 diabetes pathogenesis.

Material and methods Donor characteristics

Children with type 1 diabetes (n = 20) and healthy children (n = 20) were recruited at University Children´s Hospital, Duesseldorf, Germany. All type 1 diabetes patients received insulin therapy at recruitment. Controls had a negative history of autoimmune and systemic autoinflammatory diseases. Study group characteristics are given in Table 1.

Table 1. Study group characteristics of type 1 diabetes patients and healthy controls Characteristic Type 1 diabetes patients “T1D” Healthy controls “Contr” p-Value Number (n) 20 20 Age (years) 12.34 (9.59-13.69) 12.55 (7.61-16.79) 0.84 Sex distribution (m|f) 10|10 11|9 0.75 Disease duration (years) 5.17 (3.04-8.14) na Age at disease onset (years) 6.21 (2.79-10.27) na Abbreviations: m, male; f, female; na, not applicable. Median (IQR) and p-values from the two-tailed Mann-Whitney U-test (for age) and the chi-square test (for sex distribution) are given. Phenotyping of PBMC

PBMC were isolated from heparinized whole blood by using density centrifugation (Biocoll, Biochrom, Berlin, Germany) according to manufacturer´s instructions and stored in liquid nitrogen. 10 samples (5 from each study group) were concomitantly thawed and processed for phenotyping and stimulation experiments. Directly after thawing (“0h”), PBMC were stained without fixation using two different antibody panels. In the first panel we included CD11b-PE-Cy-7 (ICRF44, Biolegend), CD14-BV650 (M5E2, BD Biosciences), CD80-BV510 (2D10, Biolegend), CD40-AF488 (5C3, Biolegend), CD86-PerCp-Cy5.5 (IT2.2, Biolegend), and IL-7Rα-PE (A019D5, Biolegend). In the second panel, CD11b-PE-Cy7 (ICRF44, Biolegend), CD14-BV650 (M5E2, BD Biosciences), CD16-APC (3G8, Biolegend), HLAD-DR-BV750 (L243, Biolegend), and IL-7Rα-PE (A019D5, Biolegend) were used. Fixable viability dye eFlour780 (Thermo Fisher Scientific) was added in both panels to exclude dead cells. Staining was performed in duplicates. Measurement was performed on an LSR Fortessa flow cytometer (BD Biosciences). FlowJo software (Miltenyi Biotech) was used to analyze the data. The investigator was blinded to the group allocation when analyzing the data. For flow cytometry analysis, we adhered to the “Guidelines for the use of flow cytometry and cell sorting in immunological studies” [48]. Representative plots of our gating strategies are shown in Figs. 1A and 3A.

Stimulation experiments

A total of 1.5 × 105 PBMC per well were cultured in 96-F-Bottom plates (Greiner Bio-One) in RPMI 1640 (Thermo Fisher Scientific) supplemented with l-glutamine (2 mM, Sigma–Aldrich), HEPES (10 mM, Thermo Fisher Scientific), and 10% FCS (Thermo Fisher Scientific).

For 24h stimulation experiments, PBMC were incubated with and without 1ng/ml LPS (Sigma–Aldrich) or 10ng/ml IL-7 (Sigma–Aldrich). After 24h of culture at 37°C and 5% CO2, cells were collected and stained without fixation using the two antibody panels.

For 72h stimulation experiments and cell count analyses, PBMC were incubated with and without IL-7 (10ng/ml). After 72h, cells were stained with antibodies from panel 2. Prior to measurements at the LSR Fortessa flow cytometer, counting beads (123count eBeads, Thermo Fisher Scientific) were added to the samples. For each sample, 7000 beads were recorded.

t-SNE analysis

t-SNE analysis calculates a two-dimensional depiction of multi-factorial similarity [49]. The proximity of included cells is illustrated by their distances in the t-SNE map. t-SNE analysis was done by using a plugin in FlowJo v10.4. t-SNE calculations were performed with 1000 iterations, a perplexity of 20, an Eta (learning rate) of 200, and a Theta of 0.5.

For t-SNE analysis after LPS and IL-7 stimulation (Fig. 5A), CD11b positive cells (for gating strategy see Fig. 1A) from all study participants (n = 40) were downsampled for every donor (1000 events) and concatenated for every condition (24h/ 24hLPS/ 24hIL-7). t-SNE analysis was performed including the parameters IL-7Rα, CD40, CD80, CD86. IL-7Rα-high and -low cells (gated on top or bottom 5% of IL-7Rα MFI of unstimulated cells) were visualized.

In a second step, t-SNE analysis was performed separately for CD11b positive cells from type 1 diabetes patients (n = 20) and healthy controls (n = 20) after downsampling to 1000 cells for every donor (Fig. 6C). The parameters IL-7Rα, CD40, CD80, and CD86 were included for t-SNE-analysis. CD40 high cells (gated on top10% of CD40 MFI) were gated on the t-SNE plot and compared for their CD40, CD80, CD86, and IL-7Rα expression.

Real-time qPCR analysis

Monocytes were isolated from PBMCs of four healthy donors with the EasySep human monocyte isolation kit (Stemcell) according to manufacturer's protocol. RNA was then isolated using the NucleoSpin kit (Macherey-Nagel) following manufacturer's instructions. cDNA was synthesized with the Maxima H Minus First Strand cDNA Synthesis kit (Thermo Scientific) according to manufacturer's protocol. Real-time quantitative PCR (qPCR) was performed using the QuantiTect SYBR Green PCR ki