M1 Macrophage exosomes MiR‐21a‐5p aggravates inflammatory bowel disease through decreasing E‐cadherin and subsequent ILC2 activation

Abstract Abnormal immune regulation is a key feature of the complex pathogenic mechanism of ulcerative colitis (UC). In particular, macrophages and group 2 innate lymphoid cells (ILC2s) are important components of natural immunity that have been shown to play important roles in the pathogenesis of UC, as well as decreased E‐cadherin expression on the colonic mucosa. However, it remains unclear how these components interact with each other. In this study, we investigated the molecular mechanisms of UC mediated by macrophage‐derived exosomes. We showed for the first time that miR‐21a‐5p expression is increased in the peritoneal exosomes of mice with dextran sulphate sodium induced enteritis and that miR‐21a‐5p expression correlates negatively with E‐cadherin expression in enterocytes. Moreover, we confirmed that miR‐21a‐5p was mainly derived from M1 macrophages and demonstrated that KLRG1, a surface inhibitory receptor on ILC2s, participated in excessive ILC2 activation in UC by promoting GATA‐3. In conclusion, our results suggest molecular targets and provide a theoretical basis for elucidating the pathogenesis of UC and improving its treatment.


| INTRODUC TI ON
Ulcerative colitis (UC) is a type of inflammatory bowel disease that mainly affects the rectum and colon and is characterized by the destruction of colonic mucosal barrier function (barrier disruption) as well as repeated diarrhoea containing mucus, pus and blood. The incidence of UC is rising worldwide, and the disease has become a significant nationwide health problem in China. 1,2 UC has a complex and multifactorial pathogenesis that includes genetic predisposition, epithelial barrier dysfunction, dysregulated immune responses, pathogenic infection and environmental factors. In addition, the intestinal mucosal immune system plays an important role in the occurrence, progression and outcome of UC 3 and abnormal immune regulation is a key feature of its pathogenesis.
Macrophages are an important component of natural immunity that form part of the first line defence in intestinal mucosa and have been shown to play an important role in the initiation of UC. [3][4][5][6] In humans, macrophages are classified into two subsets with seemingly different functions: M1 macrophages, also known as 'classically activated' macrophages, are mainly involved in biological processes that promote miR-21a-5p expression is increased in the peritoneal exosomes of mice with dextran sulphate sodium induced enteritis and that miR-21a-5p expression correlates negatively with E-cadherin expression in enterocytes. Moreover, we confirmed that miR-21a-5p was mainly derived from M1 macrophages and demonstrated that KLRG1, a surface inhibitory receptor on ILC2s, participated in excessive ILC2 activation in UC by promoting GATA-3. In conclusion, our results suggest molecular targets and provide a theoretical basis for elucidating the pathogenesis of UC and improving its treatment.

K E Y W O R D S
E-cadherin, group 2 innate lymphoid cell, macrophage-derived exosome, MiR-21a-5p, ulcerative colitis inflammation and inhibit tumours, whereas M2 macrophages, also known as 'selectively activated' macrophages, play important roles in tumour progression, parasite control and tissue repair. 7 The pathogenic M1 subset is significantly more prevalent in colons with inflammation induced by dextran sodium sulphate (DSS), and this difference is consistently observed in most locations. 9 Exosomes are a heterogeneous group of cell-derived membranous structures that originate from the endosomal system or are shed from the plasma membrane. Importantly, exosomes are present in biological fluids and are involved in multiple physiological or pathological processes. 14 For instance, serum exosomes isolated from DSS-induced mice have been shown to induce p38 and ERK phosphorylation as well as the production of tumour necrosis factor α (TNFα), suggesting that exosomes may play an important role in the pathogenesis of UC. 15 Moreover, the oral administration of autologous exosomes from the healing phase could represent a safe and effective personalized approach for treating UC. 16 However, it is not yet clear whether dysregulated M1 macrophage-mediated immune responses play a pathogenic role in UC via exosomes. The innate immune system provides the first line of defence against invading pathogens as well as important cues for the development of adaptive immunity. Innate lymphoid cells (ILCs) are a new class of lymphocytes that have recently been discovered to play an important role in innate immunity at the mucosal barrier and promote the establishment and maintenance of adaptive immunity. 17 According to their phenotypes and secreted cytokines, ILCs can be divided into three categories: ILC1, ILC2 and ILC3. ILC2s are important source of Th2-type cytokines and play important roles in the activation and maintenance of Th2-type immunity. 18 DSS-induced colitis displays a skewed Th2 immunity characterized by low levels of IFNγ and TNFα but high levels of IL-4, IL-6, IL-10, IL-17A and IL-17F. 19,20 In addition, the level of Th2 cytokines has been associated with the severity of UC and poor prognosis; therefore, IL-13 has been suggested as a new target for host-directed therapies against UC. [21][22][23] Forkel et al also found that UC inflamed colons contain significantly more ILC2s, 24 indicating that these cells may participate in the pathogenesis of UC by promoting Th2 immunity.
We used miRWalk to search potential target gene of miR-21a-5p.
The results showed that Cdh1, which is the gene that express E-cadherin, was one of the possible target gene (P =0.846) and they may interact through 3'UTR. Combined with the result of mouse intraperitoneal exsosomes high throughout sequencing, which indicated that miR-21a-5p had one of the most significant expression change between DSS-induced enteritis mice and healthy mice, we speculate that miR-21a-5p may attenuate ulcerative colitis by decrease the expression of E-cadherin. Besides, the correlation between E-cadherin and KLRG1 and the correlation between KLRG1 and GATA-3 have been proved by several studies. KLRG1 is a well-known marker of NK cells that is also expressed on ILC2s, and KLGR1 binds with its ligand, E-cadherin, indicates an immunosenescent phenotype. Therefore, in this study, we aimed to clarify the role and mechanism of macrophages and ILC2s in the pathogenesis of UC to provide potential targets and a theoretical basis for the development of new drugs to prevent and treat UC.

| Mice and DSS-induced colitis
Male C57BL/6 mice (7 to 8 weeks old, 20 ± 0.5 g) were purchased from the Medical Experimental Animal Center of Xiangya Second Hospital for use in our experiments with approval from the Animal Experimentation Committee of Xiangya School of Medicine, Central South University. Thirty male mice were maintained on a 12-hour/12-hour light/dark cycle in a temperature-controlled room (24°C). Firstly, mice were fed with normal food and water for one week to adapt. Then, they were randomly divided into two equal groups: a) DSS-induced colitis group and a control group. The DSSinduced colitis group received 3% DSS (MP Biomedicals, Ohio, USA) dissolved in drinking water for 6 days, followed by normal drinking water for 5 consecutive days. Control group was the same as the experimental group but with normal drinking water for 11 consecutive days. Weight change was recorded every day. Model success and colitis severity were evaluated using the weight loss and disease activity index (DAI).

| Exosome extraction
After the mice had been anaesthetized and killed, 5 mL of RPMI-kneaded for 2-3 minute. The plunger of the syringe was gently withdrawn and pushed several times and then withdraw to obtain peritoneal wash fluid, which was subjected to gradient centrifugation as follows: 1000 g for 10 minutes; supernatant removed; 2000 g for 10 mins; supernatant removed; 10 000 g for 30 minutes; supernatant removed; supernatant filtered with a 0.22 μm filter; 100 000 g for 120 minutes; and purified exosomes harvested from the precipitate. Exosomes were identified by particle size analysis using transmission electron microscopy and morphology ( Figure S1).
To produce the DSS-induced inflammation model in vitro, FHC cells were starved for 12 hours, treated with 1% DSS for 4 hours and then incubated with reduced serum culture medium at 37°C for 4 hours.

| Microphage differentiation
THP-1 cells were differentiated into M0, M1 and M2 macrophages as described previously (25). Briefly, THP-1 cells were treated with 200 nM phorbol 12-myristate 13-acetate (PMA) for 24h until they became adherent, indicating that an M0 phenotype had been induced ( Figure S2). To produce M2 cells, the M0 macrophages were treated with 20 ng/mL of IL-4 and IL-13 for a further 18 hours whereas M1 macrophages were induced by incubation with 100 ng/ mL of LPS and 20 ng/mL of IFNγ. The control group was treated with same amount of solvent, which is phosphate buffer saline, at the same time-point.

| Co-culture of exosomes and FHC cells
Exosomes were isolated as described above. The exosomes were added to FHC cells for 12 hours of incubation at 37℃. The PKH 26-labelled exosomes displayed red colour. Cytoskeleton was stained with phalloidin (green), and the cell nuclei were stained with DAPI (blue). Microscopic observation was performed at 0 hours, 4 hours and 12 hours.

| Reverse transcription-polymerase chain reaction (RT-PCR)
RNA was extracted using TRIzol (Thermo Fisher Scientific, New York, USA) according to the manufacturer's protocol. Real-time PCR was conducted using an iScriptTM two-step RT-PCR kit with SYBR Green (Invitrogen) and gene-specific primers (Table S1). PCR was performed in triplicate on Roche LightCycler® 480 (Roche, Basel, Swiss) with the following cycling parameters: 95 ℃ (1 minute) followed by 40 cycles of 95°C (15 seconds) and 60°C (60 seconds). The qRT-PCR data were quantified using the 2-△△ Ct method.

| Statistical analysis
SPSS for Windows (Version 21.0, SPSS Inc, Illinois, USA) was used for statistical analysis. Statistical differences between the experimental group and control group were assessed with the one-way ANOVA test. Correlations were evaluated using the Spearman rank correlation coefficient. Statistical significance was set at P <0.05.

| Macrophages produce exosome-derived miR-21a-5p
First, in order to confirm whether intraperitoneal macrophages attenuate ulcerative colonitis through exosomes, we co-cultured mouse colonic epithelial cells with intraperitoneal macrophages from healthy mice, mice with DSS-induced enteritis or mice with DSS-induced enteritis treated with the exosome inhibitor, GW4869.
Co-culture with intraperitoneal macrophages from mice with DSSinduced enteritis significantly increased the levels of inflammatory factors (IL-6, IL-11, IL-22, TNFα) in colonic epithelial cells compared to macrophages from healthy mice, whereas GW4868 eliminated this difference ( Figure 1A). To identify potential factors involved in this mechanism, we performed high-throughput sequencing on intraperitoneal exosomes extracted from mice with DSS-induced enteritis and healthy mice. Significantly higher miR-21a-5p levels were observed in the intraperitoneal exosomes extracted from mice with DSS-induced enteritis mice than in the control group, whereas miR-21a-5p levels in peripheral blood remained similar ( Figure 1B-D).
Next, we determined whether miR-21a-5p was secreted by macrophages by examining the numbers of intraperitoneal macrophages and miR-21a-5p expression in intraperitoneal exosomes from mice with DSS-induced enteritis pre-treated with Clophosome, which can efficiently exhaust macrophages. Clophosome not only depleted the levels of intraperitoneal macrophages, but also significantly decreased miR-21a-5p expression in intraperitoneal exosomes ( Figure 1E). Therefore, we compared miR-21a-5p expression in M1and M2-derived exosomes by inducing THP-1 cells to become M0 and then M1 or M2 macrophages. The efficacy of M0 induction was confirmed by flow cytometry, which indicated that 90% of cells were successfully induced. Moreover, comparing the expression of miR-21a-5p in the supernatant of M1 and M2 exosomes revealed that M1 exosomes displayed significantly higher miR-21a-5p expression than M0 exosomes ( Figure 1F), whereas M2 exosomes showed no significant difference ( Figure 1G).

| Exosome-derived miR-21a-5p aggravates DSSinduced enteritis by decreasing E-cadherin expression
After we confirmed that miR-21a-5p in M1 exosomes could attenuate DSS-induced enteritis, further investigation of the target gene of miR-21a-5p was made. Using miRwalk, a well-established target gene prediction website, to search potential target gene of miR-21a-5p. The results showed that Cdh1, which is the gene that express nuclei stained with DAPI, blue signals) at 0 hours and then start to merge at 4 hours. Finally at 12 hours, almost all red signals located inside of the FHC cells, indicating that FHC cells internalize the exosomes contained miR-21a-5p secreted by M1 cells. Therefore, we investigated whether a correlation existed between miR-21a-5p and E-cadherin by producing stable miR-21a-5p-overexpressing and -silenced FHC cell strains and examining E-cadherin mRNA expression ( Figure S5A). We found that miR-21a-5p up-regulation decreased E-cadherin mRNA expression, whereas its down-regulation increased E-cadherin mRNA expression ( Figure 2E). In addition, correlation analysis revealed a negative correlation between the intraperitoneal macrophage-derived exosomes of mice with DSS-induced enteritis and E-cadherin expression (R 2 = 0.8464; Figure 2F).

| Absence of E-cadherin promotes ILC2 activation via KLRG1/GATA-3
Because of the relationship between E-cadherin and ILC2, we con-  Figure 3A,B). Therefore, we performed immunoprecipitation assays on HEK293T cells transfected with Flag-KLGR1 and HA-E-cadherin 36 hours earlier and positively identified an interaction between E-cadherin and KLRG1 ( Figure 3C).

Finally, we validated the inhibitory effect of E-cadherin on ILC2s
by co-culturing E-cadherin-silenced colonic epithelial cells ( Figure   S5B) with newly sorted ILC2s and examining Th2 cytokine levels.
Silencing E-cadherin was found to increase Th2 cytokine levels ( Figure 3D), whereas newly sorted ILC2s treated with E-cadherin to induce in vitro activation secreted significantly less IL-5 and IL-13 and expressed lower levels of GATA-3, IL-5 and IL-13 mRNA ( Figure 3E).
Together, these results suggest that E-cadherin may inhibit the activation of ILC2s via KLRG1/GATA-3.

| ILC2s and Th2 immunity are activated in UC
To elucidate the role that ILC2s play in UC, we evaluated the change in proportion of ILC2s in the spleen lymphocytes of mice with DSSinduced enteritis using flow cytometry. A significant increase was observed in the proportion of ILC2s in the spleen and intestinal mucosa lamina propria lymphocytes of mice with DSS-induced enteritis compared to the control group ( Figure 4A-B). In addition, we investigated the change in the proportion of Th2 cytokines in the intestinal mucosa of mice with DSS-induced enteritis using ELISA, finding that the mice with DSS-induced enteritis displayed significantly higher IL-4, IL-5 and IL-13 levels than the control group ( Figure 4C).
Therefore, we sought to evaluate the change in the proportion of  Figure S6A); however, patients with active UC showed significantly higher Th2 cytokine levels than healthy volunteers ( Figure S6B).

| D ISCUSS I ON
Cell adhesion molecules mediate adhesion and binding between cells or between cells and the extracellular matrix. E-cadherin is a type of cell adhesion molecule that plays an important role in maintaining integrity by mediating mutual aggregation between cells 25,26 ; however, its dysregulation has been shown to participate in the pathogenesis and remission of UC. [25][26][27] In particular, the down- Therefore, miR-21a-5p may target E-cadherin and result in the destruction of the intestinal mucosal epithelium, thus suggesting a novel mechanism via which abnormally polarized macrophages lead to the destruction of the intestinal mucosal barrier via E-cadherin ( Figure 5).
KLRG1 is a well-known marker of NK cells that is also expressed on ILC2s and whose binding with its ligand, E-cadherin, indicates an immunosenescent phenotype. As a key feature of mature ILC2s, KLRG1 plays a critical role in maintaining ILC2 cell function, whereas reduced KLRG1 expression or binding with E-cadherin has been F I G U R E 4 ILC2 proportion increased in UC. ILC2 proportion increased in (A) spleen macrophages, (B) colonic mucosa and lamina macrophages in DSS-induced enteritis mice (n = 15) compared to control group (n = 15); (C) Th2 cytokines were significantly higher in colon tissue than in the control group. * P < 0.05, ** P < 0.01 shown to induce excessive ILC2 activation and Th2-type immunity. [28][29][30][31] In addition, ILC2s may contribute towards increased type 2 cytokine production in the absence of the suppressive E-cadherin ligation and decreased KLRG1-E-cadherin binding via the recently described mechanism of barrier sensing. 32  Therefore, we speculate that miR-21a-5p inhibits E-cadherin expression, thereby reducing its binding to KLRG1 and indirectly promoting ILC2 activation to produce the excessive pathogenic Th2 type immune response typically observed in UC ( Figure 5).
GATA-3 is a key transcription factor that regulates the differentiation, maturation and activation of ILC2s and thus controls their number and function. 33 In addition, studies have shown that sex hormones can affect KLRG1 and regulate the number and activity of ILC2s in lung tissue via GATA-3. 34 Our results confirmed that E-cadherin supplementation in vitro significantly reduced the levels of Th2 type cytokines (IL-5, IL-13) secreted by ILC2s as well as their GATA-3 expression. Previous studies have found that this pro-activation can be blocked by silencing GATA-3 in ILC2s, 31,32 suggesting that GATA-3 plays an important role in ILC2 activation.
Therefore, we speculate that reduced E-cadherin expression in intestinal mucosa epithelial cells reduces its binding to KLRG1 on the surface of ILC2s, allowing KLRG1 to increase GATA-3 expression in ILC2s and promote their activation ( Figure 5). In conclusion, our results suggest molecular targets and provide a theoretical basis for elucidating the pathogenesis of UC and improving its treatment.

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.