MiR‐363‐5p modulates regulatory T cells through STAT4‐HSPB1‐Notch1 axis and is associated with the immunological abnormality in Graves' disease

Abstract MiRNAs are a class of small non‐coding RNAs with ability to regulate function of Treg cells and are involved in many autoimmune diseases. Our previous study found that miR‐363‐5p expression was significantly upregulated in peripheral Treg cells of GD patients. Herein, we aimed to investigate its effect and mechanism on Treg cell dysfunction in GD patients. The results showed that miR‐363‐5p upregulation was significantly associated with the Treg cell dysfunction and inflammatory factors levels in GD patients. Transcriptome sequencing revealed that 883 genes were significantly regulated by miR‐363‐5p in Treg cells. These genes with significant differential expression were primarily involved in lymphocyte differentiation, immunity, as well as Notch1 and various interleukin signalling pathways. Moreover, miR‐363‐5p can regulate HSPB1 and Notch1 through the target gene STAT4, thereby regulating Notch1 signalling pathway and inhibiting Treg cells. The effects of miR‐363‐5p on Treg cell function and STAT4‐HSPB1‐Notch1 axis were also verified in GD patients. In conclusion, our results indicated that miR‐363 could inhibit the proliferation, differentiation and function of Treg cells by regulating the STAT4‐HSPB1‐Notch1 axis through target gene STAT4. MiR‐363‐5p may play an important role in Treg cell dysfunction and immune tolerance abnormalities in GD patients.

Nonetheless, most of the current research on Tregs in GD patients focused on the description of abnormalities, while the molecular mechanisms behind these anomalies are still not fully investigated. The reported mechanism studies have also focused on key functional proteins of Treg cells such as FOXP3 and IL-10. 8,9 However, Treg functions are regulated by many molecules, including not only proteins such as cytokines and transcription factors, but also other molecules such as prostaglandins and nucleic acids. A thorough understanding of the functional regulation of Tregs and the mechanisms by which Tregs are abnormal in GD patients requires an indepth study of these non-protein regulatory molecules.
MiRNAs are 18-25nt non-coding RNAs with regulatory functions. They play critical roles in the development and function of immune system. Dicer is the key RNase that regulates maturation of various miRNAs. Treg-specific ablation of Dicer causes severe systemic autoimmune diseases in mice, similar to that in mice whose Tregs are eliminated directly, suggesting that miRNAs are key regulators of development, proliferation and function of Tregs. 10 Several miRNAs had been reported to play vital roles in Treg development and T cell-dependent antibody responses by targeting Foxp3. 11 MiR-106b could regulate Treg differentiation and maturation by targeting CDKN1A/p21 to regulate TGFβ pathway. 12 MiR-17 could also regulate Foxp3 expression and function. 13 However, there are also conflicting results on the roles of miRNA in Treg function. Some studies reported that miR-21 negatively regulated Foxp3 expression and Treg ratios in adult peripheral blood mononuclear cells (PBMCs), but other study showed that Foxp3 expression was positively regulated by miR-21. 14,15 This implies that the roles and mechanisms of miRNAs in Tregs function are complex and more intensive efforts are needed to elucidate them.
Accumulating evidences demonstrated that miRNAs play key roles in the development various autoimmune diseases, such as systemic lupus erythematosus and multiple sclerosis. MiR-409-3p and miR-1896 could synergistically exacerbate experimental autoimmune encephalomyelitis in mice. 16 The serum levels of several miR-NAs in GD patients were also found to be significantly abnormal. 17 Our previous studies showed that the expression of some miRNAs was significantly changed in peripheral Tregs of GD patients, causing abnormalities in multiple signalling pathways. 1  Herein, we found that the number and function of Tregs were significantly decreased compared with healthy individuals. Meanwhile, miR-363-5p expression in Tregs was significantly upregulated, which was closely associated with immune inflammation and thyroid dysfunction in GD patients. MiR-363-5p could inhibit proliferation/ differentiation and function of Tregs by regulating STAT4-HSPB1-Notch1 axis through its target gene STAT4. This might be one of the important mechanisms for the abnormalities in Tregs in GD patients.
Our findings provided valuable clues to reveal the mechanisms of immune abnormalities in GD patients.

| Ethics statement
This study has been approved by the Ethics Committee of Shandong Provincial Hospital (NO 2015-054) and conformed to the Declaration of Helsinki. All subjects were informed of the nature of the study and were provided informed consent. All procedures were performed in compliance with relevant laws and institutional guidelines.

| Recruitment of subjects and sample collection
Twenty GD patients and healthy individuals were recruited from Shandong Provincial Hospital. GD patients were diagnosed according to the medical history, physical examination and laboratory tests including TSH, free triiodothyronine (FT3), free thyroxine (FT4) and TRAb based on 2018 European Thyroid Association Guideline for the Management of Graves' Hyperthyroidism. 18 The exclusion criteria were as follows: smoking, pregnancy, alcohol addiction, hypertension, diabetes mellitus, lipid dysregulation, BMI>27; usage of hormonal medication; and medical history of malignancy. Peripheral blood (5 ml) was collected from all subjects in the morning after an overnight fast (≥8 h). The serum samples were stored in −80°C for subsequent experiments.

| Flow cytometric analysis and Cytokine assay
Peripheral blood mononuclear cells were purified by Ficoll density centrifugation (Sigma-Aldrich) and stained with anti-CD4-FITC (BD Bioscience), anti-CD25-APC (eBioscience) for 30 min at 4°C in the dark. Following fixation and permeabilization, cells were stained with anti-Foxp3-PE (eBioscience). Flow cytometric analysis was performed using FACSCalibur (Becton Dickinson), and the data were analysed by FlowJo X software (Flow Jo). The levels of cytokine IL-2, sCD25, IL-10 and sCD14 were detected by the Multiplex Human Premixed Multi-Analyte Kit (R&D, USA) and high sensitivity ELISA kits (eBioscience), respectively. All procedures were performed according to the user manual.

| RNA extraction and Real-time quantitative PCR
Total RNA was extracted using TRIZOL reagent. MiRNA expression was detected by real-time PCR, and locked nucleic acid primers were purchased from TaKaRa. Human RNU6B was used as an internal reference. The mRNA expressions of STAT4, HSPB1, Notch1, IL-10 and Foxp3 were analysed using SYBR Green qPCR, and their primers were synthesized by Sangon Biotech. Human ACTB was used as an internal reference. Reverse transcription kit and TB Green Premix Ex Taq kit were purchased from TaKaRa. All primers used in this study are presented in Table S1.

| Transcriptome sequencing and data analysis
The mRNA was purified from 5 μg total RNA using poly-T oligoattached magnetic beads. After ultrasonic fragmentation, sequencing libraries were prepared using TruSeq RNA Library Prep Kit (Ilumina). PE150 sequencing was then performed on an Illumina HiSeq 2500 (Illumina). After removal of low-quality reads and sequences of adaptors and primers, all clean data were aligned with UCSC human reference genome (hg19) using HISAT package. The expression levels of all detected genes were standardized (FPKM, Fragments per Kilobase Million) using StringTie and ballgown. The genes with significantly greater than 1.5-fold change in expression levels were considered as differential genes. Pathway and functional enrichment analysis of target genes as well as differentially expressed genes was carried out using Enrichr online tools. 19,20

| Western Blotting
Cells were lysed by RIPA lysis buffer containing 1 mmol/L PMSF (Beyotime). After centrifugation at 12,000 g for 15 min, the supernatant was collected and protein concentration was determined using BCA (Thermo). Total protein was electrophoresed on 10% SDS-PAGE and electrically transferred to PVDF membranes (Bio-Rad). After blocking, the membranes were incubated with antiβ-actin (Santa Cruz Biotechnology), anti-phospho-Stat4 (Tyr693) and anti-Notch1 mAb (Cell Signaling Technology) overnight at 4°C.
After washing, the membranes were incubated with the peroxidaseconjugated secondary antibody (Proteintech Group) for 2 h. Protein bands were detected using ECL kit. Densitometry analysis was performed by Image pro plus 6.0 software (Media Cybernetics).

| Dual-luciferase reporter assay
The human STAT4 3'UTR region containing the predicted binding sequence of miR-363-5p seed sequence was amplified with normal primers (WT), and the mutant sequence of the aforementioned 3'UTR region was amplified with mutation-site primers (Mut). Then, these sequences were cloned into pmir-GLO vector (Promega) to construct wild-type and mutant luciferase reporter vectors, respectively. Next, these reporter vectors were co-transfected into 293T cells with miR-363-5p or corresponding negative control sequence (Ribobio), respectively. After transfection for 36 h, the luciferase activity was measured using Dual-Luciferase Reporter Assay System (Promega).

| Statistical analysis
Data were presented as mean ± SEM or median (IQR). The t test was used to detect the significances between two groups. One-way anova test with appropriate correction for multiple comparisons was used to detect significances between multiple groups. The p < 0.05 or corrected p < 0.05 was considered as statistically significant. All experiments were repeated independently at least three times.

| Significant abnormalities in peripheral inflammatory cytokines and Treg cells of GD patients
In the present study, 20 initial patients with GD (GD group) and 20 healthy individuals (control group) were recruited. The demographic characteristics of all subjects are presented in Table 1.
The levels of serum FT3, FT4 and TRAb in GD group were significantly higher than in control group (Figure S1A-C). The percentage of Treg (CD4+CD25+Foxp3+) cells in CD4+ T cells of peripheral blood in GD patients was significantly lower than in healthy individuals ( Figure 1A). The immunosuppressive function of Treg cells in the peripheral blood of GD patients was also obviously decreased ( Figure 1B). Moreover, compared with healthy individuals, the serum level of anti-inflammatory cytokine IL-10 was markedly decreased ( Figure 1C), while the serum levels of pro-inflammatory cytokines including sCD14, sCD25 and IL-2 were significantly increased in GD patients ( Figure 1D-F). These results suggested the significant abnormalities in immune and inflammation of patients with GD.

| MiR-363-5p was significantly upregulated in Treg cells and correlated with abnormalities in thyroid function and inflammation of GD patients
It had been reported that the expression of some miRNAs including miR-363-5p was significantly changed in peripheral Treg cells of GD patients. 1 As one of the most significantly altered miRNAs, the expression changes of miR-363-5p in Tregs of GD patients were confirmed using real-time PCR. The results showed that miR-363-5p was significantly upregulated in peripheral Tregs of GD patients compared with healthy individuals, which was consistent with previous studies reported ( Figure 1G  The t test was used to detect significant changes.
miR-363 was selected. The enrichment analysis of the genes in the intersection suggested that the potential target genes of miR-363 were mainly involved in cell migration, apoptosis, secretion and thyroid hormone synthesis and were associated with various immune diseases such as systemic lupus erythematosus and type 2 diabetes.
Especially, many immune-inflammatory functions were also significantly enriched, such as innate immunity, Th1 and Th2 cell differentiation, and T-cell activation ( Figure 2D). The pathway enrichment analysis indicated that these potential target genes of miR-363-5p were involved in many signalling pathways such as Notch, Nod1/2 and Toll-like receptors, thyroid hormones, various interleukins (IL-2, IL-6, etc.) and cytokine/chemokine signalling pathways ( Figure 2E).
All these results implied that miR-363-5p was strongly correlated with immune response and might regulate CD4+ T-cell function through several signalling pathways including STAT, Notch and Wnt.

| MiR-363-5p targeted STAT4 to inhibit its expression
The prediction results of various software indicated that STAT4 was the target gene of miR-363-5p. Moreover, mRNA sequencing also suggested that miR-363-5p agomir could significantly suppress STAT4 expression. The 3'UTR of STAT4 in several species had a conserved binding site for the miR-363-5p seed sequence with a binding energy of −13.7 kcal/mol ( Figure 3A). MiR-363-5p agomir significantly decreased STAT4 expression at both protein and mRNA levels in CEM cells. MiR-363-5p antagomir significantly upregulated STAT4 mRNA expression but increased STAT4 protein expression without significant differences ( Figure 3B,C, Figure S3A). To clarify whether Stat4 is a target of miR-363-5p, a dual-luciferase reporter plasmid containing wild-type (STAT4-WT) or mutant (STAT4-Mut) sequence of STAT4 3'UTR was constructed, respectively ( Figure 3D).
To enhance the reliability of results, let-7e was used as a negative control because STAT4 was not its target gene. The well-established target gene of miR-150, c-myb, was used as a positive control. These plasmids were co-transfected with miR-363-5p agomir in 293T cells, respectively. The results showed that the luciferase activity was significantly reduced in 293T cells co-transfected with miR-363-5p agomir and STAT4-WT plasmid, while no significant changes were observed in 293T cells co-transfected with miR-363-5p agomir and STAT4-Mut plasmid. Meanwhile, the luciferase activity was significantly decreased in 293T cells co-transfected with miR-150 and the plasmid containing c-myb 3'UTR, while there was no significant change in luciferase activity of 293T cells co-transfected with let-7e and STAT4-WT plasmid ( Figure 3E). These results confirmed that STAT4 was the target of miR-363-5p.

| MiR-363-5p regulated Notch1 signalling pathway through STAT4 and HSPB1
The mRNA sequencing results indicated that, in addition to its target gene STAT4, miR-363-5p also regulated many other genes including is an important transcription factor regulating the expression of numerous genes, many of which modulate Notch1, such as HSPB1.
Real-time PCR showed that miR-363-5p significantly downregulated HSPB1 expression in CEM cells, which was consistent with mRNA sequencing data ( Figure 4D). The HSPB1 expression vector could reverse the inhibitory effects of miR-363-5p on HSPB1 and Notch1 expression ( Figure 4E,F). Also, HSPB1 was not a target gene of miR-363-5p ( Figure S3B). Furthermore, the STAT4 expression vector could also alleviate the inhibitory effects of miR-363-5p on the expression of HSPB1 and Notch1 ( Figure 4G,H). Thus, these results demonstrated that miR-363-5p could downregulate HSPB1 by suppressing its target gene STAT4, thereby repressing Notch1 expression and regulating the Notch1 signalling pathway.

| MiR-363-5p suppressed Treg differentiation and function via STAT4-HSPB1-Notch1 axis and was associated with immune abnormalities in GD patients
To  was consistent with our findings in GD patients. This also suggests that the same gene may play different or even opposite roles in different diseases, which has been widely reported. [33][34][35][36] All these indicated that the roles of STAT4 in immune regulation including Tregs development, proliferation, differentiation and function were complicated and needed to be further investigated.
Notch1 is the key molecule in the regulation of T cells, and its role in autoimmune diseases remains obscure. Especially, there are still many contradictions in its role in the regulation of proliferation, differentiation and function of Treg cells. 37 Several studies showed that blocking Notch1 signalling significantly inhibited the function and proliferation of Treg cells, 38 which was consistent with our findings. However, some other studies showed that Notch1 pathway activation promoted Tregs apoptosis and inhibited its function 39 Blocking Notch1 could promote the function and proliferation of Tregs through increasing STAT5 phosphorylation. 40 The reasons and mechanisms for these conflicts remain unclear and might be lie in the differences in disease, animal models, inflammatory environments and cytokines. Obviously, more research is needed to fully reveal the roles of Notch1 in Treg cell function.
Our study revealed that STAT4 regulated Notch1 expression through heat-shock protein family B (small) member 1 (HSPB1).
HSPB1 is a highly conserved protein and suppresses apoptosis caused by stress signals. 41 HSPB1 is reported to be highly expressed in many cancers and participate in tumour cell proliferation. 41 Some studies also showed that HSPB1 was involved in the inflammatory response and had immunomodulatory activity. 42,43 We revealed for the first time that HSPB1 could participate in the dysfunction of peripheral Tregs in GD patients via mediating the regulation of Notch1 signalling by STAT4. However, the roles of HSPB1 in immune modulation and more autoimmune diseases are still unknown and deserve further study. It should be noted that miR-363-5p inhibitor had no significant effect on the expression of its target genes or downstream molecules in CEM cells. This might be due to the fact that miR-363-5p expression was already at a low level in CEM cells and the effect of its inhibitor was not obvious. However, due to the high level of miR-363-5p expression in Tregs from GD patients, the effect of its inhibitor on its expression and target genes becomes apparent. This phenomenon has also been reported in many studies. For example, miR-155 inhibitor could significantly reduce cell proliferation in HuT102 cells expressing high levels of miR-155, while this effect was far less pronounced in HuT78 cells expressing low level of miR-155. 44 In summary, in the present study, we found for the first time that miR-363-5p was significantly upregulated in peripheral Tregs of GD patients and played important roles in Treg dysfunction of GD patients; namely, miR-363-5p could suppress proliferation/differentiation and function of Tregs through STAT4-HSPB1-Notch1 axis, which might be an important mechanism for the decreased number and dysfunctions of Tregs in GD patients ( Figure 6H). Our findings improved our knowledge of miRNA function and Tregs regulation and deepened our understanding of the mechanism of immune abnormalities in GD patients.

ACK N OWLED G EM ENTS
This work was supported by the National Natural Science Foundation of China (82070820).

CO N FLI C T O F I NTE R E S T S
Authors declare no competing interests.

AUTH O R CO NTR I B UTI O N S
Xianlun Yin involved in conceptualization (equal); data curation (equal); and writing-original draft (lead). Junfeng Ge contributed to formal analysis (equal) and investigation (equal). Xiurong Ge contributed to investigation (equal) and resources (equal). Jing Gao contributed to methodology (supporting) and validation (supporting). Xinhuan Su contributed to data curation (equal) and methodology (supporting).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.