Mesenchymal stem cells‐derived extracellular vesicles containing miR‐378a‐3p inhibit the occurrence of inflammatory bowel disease by targeting GATA2

Abstract This study sought to determine whether mesenchymal stem cells‐derived extracellular vesicles (MSCs‐EVs) carrying microRNA‐378a‐3p (miR‐378a‐3p) could affect the pathogenesis of inflammatory bowel disease (IBD) by regulating the GATA‐binding protein 2 (GATA2)/aquaporin‐4 (AQP4)/peroxisome proliferator‐activated receptor α (PPAR‐α) axis. Initially, colon mucosa biopsy tissues were harvested from healthy controls and patients with IBD for qRT‐PCR and immunohistochemistry analysis. EVs harvested from MSCs and lipopolysaccharide (LPS) were used to stimulate the M064 cells to establish an in vitro inflammation cell model. Besides, 2,4,6‐trinitrobenzene sulfonic acid intracolon administration was performed to establish in vivo IBD mouse models. After loss‐ and gain‐of‐function assays, the regulatory role of MSCs‐derived EVs loaded with manipulated miR‐378a‐3p in IBD in relation to GATA2/AQP4/PPAR‐α were explored. Upregulation of GATA2 was identified in the colon tissue of IBD patients. GATA2, which was a target gene of miR‐378a‐3p, transcriptionally upregulated AQP4. After silencing of GATA2, LPS‐induced apoptosis of M064 cells was reduced by the downregulation of AQP4. Decreased AQP4 contributed to PPAR‐α pathway inactivation and weakened the LPS‐induced apoptosis of M064 cells. MSCs‐EVs delivering miR‐378a‐3p suppressed the GATA2/AQP4/PPAR‐α pathway, which reduced LPS‐induced apoptosis of M064 cells and the occurrence of IBD in mice. Altogether, the current study illustrated that MSCs‐EVs transfer miR‐378a‐3p to reduce the GATA2 expression, which downregulates AQP4 to block the PPAR‐α signalling pathway, thus suppressing the occurrence of IBD.


| INTRODUC TI ON
Inflammatory bowel disease (IBD), classified as Crohn's disease and ulcerative colitis, is considered as chronic remittent/progressive inflammatory response affecting the entire gastrointestinal tract or the colonic mucosa. 1 Pathologically, IBD is characterized by intestinal inflammation as well as extraintestinal manifestations such as keratopathy and episcleritis. 2 As IBD is a multifactorial disorder, it is associated with various modification ranging from dietary, genetic, immunological and microbial factors. 3 Statistically, the incidence of IBD is relatively low in Asian population relative to the Western population; however, a gradual rise has been apparent. 4 Currently, surgical intervention is regarded as the gold standard for the treatment of IBD, 5 and bacteriotherapy has been proposed as a promising protocol for the treatment of IBD. 6 Mesenchymal stem cells (MSCs) have been identified as multipotent cells that are capable of proliferating or replacing dead cells in the body.
MSCs also secrete immunomodulatory molecules. 7 Recently, the application of MSCs has been proposed as an effective treatment protocol of IBD. 8 Extracellular vesicles (EVs), including small EVs or exosomes, are small vesicles secreted by various cell types that can transfer intricate bioactive molecules for involvement in various pathophysiological pathways of intercellular communication. 9 In light of existing research, microRNAs (miRs) can essentially function as efficient diagnostic molecules for the detection and classification in IBD subtypes. 10 Intriguingly, miR-378a-3p could increase the IL-33 expression upon inflammation, which can exacerbate the pathogenesis of ulcerative colitis. 11 Based on our bioinformatic analysis, GATA-binding protein 2 (GATA2) was targeted by miR-378a-3p. An existing study identified GATA2 as a multi-catalytic transcription gene, which could serve as a regulator of inflammatory processes. 12 As previously reported, GATA2 can radically influence the inflammation in colitis. 13 Aquaporin 4 (AQP4) has been identified as a type of transmembrane protein of the aquaporin family and a vital water channel in the mammalian brain. 14 Moreover, an existing study determined that the lack of AQP4 could facilitate the alleviation of experimental colitis in a mouse model. 15 Peroxisome proliferator-activated receptors (PPARs) have been identified as ligandactivated transcription genes with notable functionality as mediators of various biological processes in IBD. 16 Strikingly, inhibited PPARα action due to palmitoylethanolamide could facilitate the alleviation of ulcerative colitis. 17 In light of the aforementioned literature, the current study sought to investigate whether MSCs-derived EVs (MSCs-EVs) containing miR-378a-3p could influence the development of IBD, with involvement of the GATA2/AQP4/PPARα axis.

| Ethical approval
The present study was performed with approval of the ethics com-

| Bioinformatics analysis
Through the Gene Expression Omnibus (GEO) database, we searched the IBD-related expression dataset GSE35609 was searched from the GEO database, which comprises of 4 normal samples and 5 disease samples. With the normal samples serving as control, the R language 'limma' package was used for a comprehensive differential analysis. With |logFC| > 1 and p value <0.05 as the screening criteria, the differentially expressed genes (DEGs) were identified. The binding domain of GATA2 in the AQP4 promoter region was predicted using the Jaspar database, while the AQP4-related genes were predicted by GeneMANIA database, and the related genes were enriched by KEGG pathway and analysed by the KOBAS3.0 database.
A combination of the starBase database and DIANA database were used to predict the upstream miRNA of GATA2. The EVmiRNA database was used to identify the miRNAs in the EVs of MSCs, and the miRNAs with more than 400 expressions were selected for subsequent analysis.  Table S1. These tissue samples were analysed using a combination of reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and immunohistochemistry.

| Isolation and identification of extracellular vesicles
The MSCs were cultured in complete medium containing FBS (centrifuged at 100,000× g for 2 h at 4°C to remove EVs). After incubation of the MSCs at the second passage for 48-72 h, the supernatant of MSCs was isolated and the EVs were extracted from MSCs (1.2 × 10 7 cells) by differential centrifugation (centrifuged at 300× g for 10 min, 3000× g for 10 min, and 20,000× g for 30 min). The number of EVs was quantified using the EXOCEP EVs Quantification Kit (System Biosciences Inc.).
Next, MSCs-EVs were subject to comprehensive evaluation.
Briefly, 20 μl EVs were dripped onto the copper wire, and the liquid was dried from the side using filter paper after allowing to rest for 3 min. Subsequently, 30 μl of the phosphotungstic acid solution (pH 6.8) was dripped onto the wire and re-dyed at room temperature for 5 min. Next, the MSC-EVs observations were documented under transmission electron microscopy (TEM). The particle size analysis 18 was performed using a nanoparticle tracking analyzer (NTA; NS300, Malvern Instruments Ltd.). Western blot assay was performed using anti-CD63 (sc-5275) and anti-CD9 (sc-9148), and calnexin (sc-23954).

| Extracellular vesicles uptake assay
The purified MSCs-EVs were incubated with PKH26 (Sigma-Aldrich Chemical Company) for 5 min at ambient temperature. The labelled MSCs-EVs were resuspended in basal medium prior to incubation with M064 cells at 37°C for 12 h. Next, 4′6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Chemical Company) was used to stain the nuclei for 10 min. The stained cells were observed under the IX53 fluorescence microscope (Olympus).

| Cell culture and establishment of inflammation model in vitro
Mouse colonic epithelial cell line M064 (SNPM-M064; Sunncell) and H&EK 293T cells (ATCC) were cultured in Dulbecco's modified Eagle medium (DMEM) (Gibco) medium containing a combination of 10% FBS, 100 μg/ml streptomycin and 100 U/ml penicillin in a 37°C incubator containing 5% CO 2 . Upon attaining 80%-90% cell confluence, the cells were passaged. The M064 cells were reacted with 1 μg/ml LPS (Sigma-Aldrich Chemical Company, prepared with 0.1% DMSO) for 12 h to successfully establish an in vitro inflammation model. 19 Meanwhile, the control cells were treated with 0.1% DMSO.
The miR-378a-3p mimic was transfected into MSCs according to the provided instructions, and the riboFECT™CP reagent (RiboBio) was added to transfect miR-378a-3p mimic or mimic NC (20 nmol) into 2 × 10 5 MSCs. After 48 h of culture, MSCs-EVs containing miR-378a-3p and MSCs-EVs-mimic NC were harvested through isolation and immediately co-cultured with the M064 cells for subsequent detection.

| Cell counting kit-8 (CCK-8) assay
M064 cells were seeded in a 96-well plate at the density of 10 4 cells per well for incubation in an incubator for 24 h. After LPS treatment for 12 h, 10 μl of the CCK-8 (Sigma-Aldrich Chemical Company) solution was added to each well. After incubation in a humidified incubator at 37℃ for 1 h, the absorbance value of each sample was estimated at the excitation wavelength of 450 nm with an Epoch microplate spectrophotometer (Bio-Tek). Six duplicated wells were set.  5'-CCCCATAGCAAACTAAGGGCT-3'. 21 The immunoprecipitated DNA was detected by qRT-PCR using iQ SYBR Green supermax (Bio-Rad Laboratories).

| qRT-PCR
The Trizol reagent (15596026, Invitrogen) was used to extract the total RNA content from colon tissues and cells. The RNA content was reversely transcribed into cDNA using the PrimeScript RT reagent kit (RR047A, Takara, Otsu). The Poly A tailed detection kit (b532451, Sangon; including Universal PCR primer R and U6 Universal PCR primer R) was adopted to isolate the cDNA content from the miRNA containing Polya tail. Fast SYBR Green PCR kit (Applied Biosystems) and ABI prism 7300 qRT-PCR system (Applied Biosystems) were used for qRT-PCR detection. U6 served as the internal reference for miR-378a-3p, and glyceraldehyde-phosphate dehydrogenase (GAPDH) served as internal reference for GATA2 and AQP4. The relative gene expression was analysed based on the 2 −ΔΔCT method.
The primer design is provided in Table S2.

| Western blot assay
The tissues and cells were isolated for lysis with the radioimmunoprecipitation assay (RIPA) lysate (Beyotime) containing 1% phenylmethylsulfonyl fluoride (PMSF) for 30 min on ice. The protein concentration was determined based on the BCA (Pierce) method.
The protein content was denatured using 5× loading buffer and boiled at 10°C for 10 min; the loading amount of the protein was 50 μg. The separation gel and concentration gel were prepared for electrophoresis. After electrophoresis, the bands containing the target protein were transferred onto a polyvinylidene fluoride membrane. Next, the membrane was immersed in 5% skimmed milk powder, sealed at ambient temperature for 1 h and treated with The Image J (NIH) software was used for grey value analysis.

| Establishment and of inflammatory bowel disease mouse model
A total of 50 male C57BL/6 mice (6 weeks old, weighing 18-22 g) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd.
The mice were housed at saturated humidity (45%-50%) and temperature (25-27°C) for 1 week with daily 12-h light/darkness cycles for acclimation to the experimental environment. The mice were fasted for 12 h before administration, and they had ad libitum access to food and water at other times. Next, 2,4,6-trinitrobenzene sulfonic acid (TNBS) was administered to the mice to establish the IBD model. 22 Briefly, the mice were anaesthetized by inhalation of diethyl ether (Sigma-Aldrich Chemical Company). Then, a 3-cm catheter was inserted into the anus. Next, 3.5 mg TNBS was dissolved in 100 μl of 30% ethanol solution and administered in the colon via the catheter. The control mice were given 100 μl of 30% ethanol solution under similar protocol. The IBD mice were used as control or treated with TNBS, with 10 mice in each group. The IBD model mice were treated with MSCs-EVs agomir NC + oe-NC, MSCs-EVs-miR-378a-3p agomir +oe-NC or MSCs-EVs-miR-378a-3p agomir + oe-GATA2, with 10 mice in each group. TNBS was administrated once a week. During the first and second regimens of TNBS administration, 200 μg of MSCs-EVs agomir NC or MSCs-EVs-miR-378a-3p agomir and 50 nmol of oe-NC or oe-GATA2 were injected intravenously into the mice. After 4 regimens of TNBS administration, the mice were euthanized by CO 2 asphyxiation, and the colitis tissues were isolated about 2 cm away from ileocecal region for follow-up detection.

| Determination of disease activity index (DAI)
Several parameters such as the body weight, faecal characteristics and faecal occult blood were documented. DAI was calculated according to the scoring system provided in Table S3.

| Haematoxylin and eosin staining
The colon segment (about 2-3 cm long) was excised, and the colon length from cecum to anus was measured. Next, the segment was fixed using 4% paraformaldehyde, paraffin-embedded and sectioned

| Immunohistochemistry
The colonic mucosa was fixed using 4% paraformaldehyde for 1 week. Next, the samples were embedded in paraffin and divided into 4 μm thick sections. The sections were treated with 0.1 M citric acid buffer (pH 6.0) and heated in a 100°C microwave oven for 10 min to facilitate antigen repair. The sections were then subject to overnight incubation with the corresponding primary antibodies anti-GATA2 (at a dilution ratio of 1:100, SC-267, Santa Cruz Biotechnology), anti-AQP4 (at a dilution ratio of 1:100, sc-32739, Santa Cruz Biotechnology) and anti-PPARα (at a dilution ratio of 1:100, sc-398394, Santa Cruz Biotechnology) at 4°C. Next, the secondary antibody goat anti-rabbit (ab6721, Abcam Inc.) in combination with HRP was subject to incubation at room temperature for 1 h. Immune response was generated using 0.05% diaminobenzidine (DAB) containing 0.01% hydrogen peroxide (H 2 O 2; Bioss). The slices were soaked in haematoxylin for 5 min, then in 1% hydrochloric acid alcohol for 4 s to revert to blue colour. The standard of proteinpositive cells was identified as the normal positive cells which were brownish yellow. The sum of integrated optical density was analysed using the Image-Pro Plus 6.0 software (Media Cybrnetics).

| Statistical analysis
All experimental data, expressed as mean ± standard deviation, were analysed using the SPSS 21.0 statistical software (IBM) and graphPad Prism 7.0. The cell experiments were conducted three times independently. Data between two groups were compared by the unpaired t test and those among multiple groups were compared by one-way analysis of variance (ANOVA), followed by Tukey's post hoc tests. In all statistical references, a value of p < 0.05, p < 0.01, p < 0.001 or p < 0.0001 was demonstrative of a statistically significant difference.

| GATA2 expression was elevated in colonic tissues of IBD patients and mice
Initially, a differential analysis was performed on the expression microarray GSE35609 obtained from the GEO database, which identified 535 DEGs ( Figure 1A). The 535 DEGs were further intersected with the transcription factors retrieved from Cistrome ( Figure 1B), and five candidate transcription factors (GATA2, ATF3, BCl3, RXRG and ARNT) were determined. These five transcription factors showed significantly differential expression in IBD. Moreover, our findings revealed that GATA2 was the most DEG among the five candidate transcription factors (logFC = 1.8). Next, qRT-PCR was conducted to detect the expression of GATA2 in the colonic mucosa of healthy controls and IBD patients. The results revealed that the expression pattern of GATA2 in the colonic mucosa of IBD patients was notably upregulated relative to the healthy controls ( Figure 1C).
Subsequently, we established an IBD mouse model for further investigation. We initially detected alterations in the weight and colon length of mice ( Figure 1D,E) and found that relative with the control mice, the weight of TNBS mice gradually decreased over time, and the colon length was also shortened significantly. Based on DAI score determination, relative to that in control mice, DAI of TNBS mice was markedly elevated ( Figure 1F). Haematoxylin and eosin staining ( Figure 1G) results elicited apparent inflammatory cell infiltration, crypt loss, mucosal destruction and oedema in the colon mucosa of TNBS mice, with an increased total injury fraction of TNBS mice in TNBS mice relative to the control mice. Subsequently, the results of qRT-PCR and Western blot assay demonstrated a markedly increased expression pattern of GATA2 in TNBS mice relative to the control mice ( Figure 1H,I, Figure S1A). Subsequently, immunohistochemistry was performed to determine the localization and expression pattern of GATA2 in the mouse colonic mucosa samples ( Figure 1J), and the results displayed a predominant expression pattern of GATA2 in the nuclei of epithelial cells from the mouse colonic mucosa tissues, and its expression was markedly increased in TNBS mice in comparison with the control mice. The aforementioned results suggested that GATA2 was upregulated in the colonic mucosa of IBD patients and mice.

| Silencing GATA2 promoted LPS-induced proliferation and reduced the apoptosis of M064 cells by inhibiting AQP4 expression
We initially detected the expression pattern of AQP4 in the colonic mucosa of healthy controls and patients with IBD ( Figure 2A) and determined that relative to the healthy controls, the AQP4 expression in the colonic mucosa of patients with IBD was significantly upregulated. The results of immunohistochemistry ( Figure 2B) elicited a prominent expression pattern of AQP4 in the cell membrane of epithelial cells from the mouse colonic mucosa tissues, and its expression pattern was markedly elevated in the TNBS mice relative to the control mice. Therefore, we speculate that GATA2 may elevate the expression pattern of AQP4 through transcription, thus increasing the occurrence of IBD.
Subsequently, we investigated the binding condition of GATA2 to the AQP4 promoter region, and identified multiple GATA2 binding domains in the AQP4 promoter region (Table S4)

| Silencing AQP4 inhibited PPAR-α pathway to promote LPS-induced proliferation and reduce apoptosis of M064 cells
We searched the AQP4-related genes ( Figure 3A), performed KEGG pathway enrichment analysis, ( Figure 3B) and found that the AQP4related genes were prominently enriched in the PPAR signalling pathway. PPARα is a member of the nuclear receptor superfamily PPARs 21,22 notably expressed in the intestinal cells, and can serve as modulators of inflammation and immune response. [23][24][25] We initially detected the expression pattern of PPARα in the colonic mucosa of healthy controls and patients with IBD, which revealed a considerably elevated PPARα expression in colonic mucosa of patients with IBD compared to the healthy controls ( Figure 3C). Immunohistochemistry ( Figure 3D) elicited identification of PPARα in the nuclei of epithelial cells from the mouse colonic mucosa tissues, with a radically rising level in the TNBS mice compared to the control mice. Further detection of the PPARα expression pattern in M064 cells after LPS treatment revealed that relative to the control cells, the PPARα level was notably higher in LPS-treated M064 cells ( Figure 3E, Figure S1E).

Subsequently, we investigated the interaction between AQP4
and PPARα through Co-IP and identified that AQP4 could definitively bind to PPARα ( Figure 3F). qRT-PCR revealed that compared to si-NC, si-AQP4 decreased AQP4 and PPARα expression pattern in LPS-treated M064 cells ( Figure 3G). Furthermore, the M064 cells were further treated with fenofibrate. Fenofibrate has extensive application in reducing the cholesterol level of patients at risk of cardiovascular disease. It can be used alone or in combination with statins for the treatment of hypercholesterolemia and hypertriglyceridemia. 26 Western blot results showed that si-AQP4 contributed to the significantly decreased protein expression pattern of AQP4 and PPARα in the presence of DMSO, while additional treatment of fenofibrate notably increased the PPARα expression pattern ( Figure 3H). Moreover, M064 cell proliferation increased F I G U R E 1 GATA2 is highly expressed in colonic tissues of IBD patients and mice. (A) The heatmap of differential gene expression pattern of IBD microarray was plotted. The sample number is shown in x-axis, and gene name is shown in y-axis; gene expression level clustering is shown in the dendrogram on the left, sample clustering in the dendrogram on the top and colour scale in the histogram on the right. (B) Retrieval of transcription factors in differential genes. Two circles in the figure represent transcription factors and differential genes, respectively, and the middle part represents the intersection of two groups of data. (C) qRT-PCR was performed to detect GATA2 expression pattern in colonic mucosa of healthy controls (n = 17) and IBD patients (n = 35). (D) Colon length of mice with different treatment was recorded (n = 10). (E) Weight loss of mice with different treatment during administration was recorded (n = 10). (F) Statistical chart was plotted for DAI score of mice with different treatment (n = 10). (G) Haematoxylin and eosin staining was performed to determine colonic tissue damage of mice with different treatment (n = 10, scale bar: 50 μm) (H) qRT-PCR was used to detect the mRNA expression pattern of GATA2 in colonic mucosa tissues of mice with different treatment (n = 10). (I) Western blot assay was used to detect the protein expression pattern of GATA2 in colonic mucosa tissues of mice with different treatment (n = 10). (J) The localization and expression pattern of GATA2 in mouse colonic mucosa tissues were determined with immunohistochemistry (n = 10, scale bar: 50 μm). Data between two groups were compared by unpaired t test and those among multiple groups at different timepoints by repeated measures of ANOVA, followed by Tukey's post hoc tests. ***p < 0.001. ****p < 0.0001

| miR-378a-3p targeted and inhibited GATA2
We predicted the upstream miRs of GATA2 using scientific databases such as starBase and intersected the prediction results with the miRs expressed in the EVs of MSCs retrieved from EVmiRNA ( Figure 4A). Finally, we identified three candidate miRs, namely miR-25-3p, miR-92a-3p and miR-378a-3p. Next, qRT-PCR was performed to detect the expression pattern of miR-378a-3p in the colonic mucosa of healthy controls and patients with IBD, which revealed considerably reduced miR-378a-3p expression pattern in the colonic mucosa of patients with IBD relative to the healthy controls ( Figure 4B). Intriguingly, the involvement of decreased miR-378a-3p has been identified in the pathogenesis of ulcerative colitis. 11 Dual-luciferase reporter gene assay displayed that compared with the mimic NC, miR-378a-3p mimic led to a notable reduction of luciferase activity in the GATA2-wild-type (WT) co-transfection group ( Figure 4C), thereby highlighting that miR-378a-3p can target GATA2. After LPS stimulation, a lowered miR-378a-3p expression pattern was evident in the M064 cells ( Figure 4D). qRT-PCR ( Figure 4E,F) demonstrated that compared with mimic NC, miR-378a-3p mimic markedly upregulated the expression pattern of miR-378a-3p in LPS-treated M064 cells, while that of GATA2 was significantly downregulated. Western blot assay revealed that the GATA2 protein level was decreased after miR-378a-3p elevation in LPS-treated M064 cells ( Figure 4G, Figure S1F). Conjointly, miR-378a-3p can target and inhibit GATA2 expression.

| Mesenchymal stem cells-derived extracellular vesicles carrying miR-378a-3p inhibited GATA2/ AQP4/PPAR-α pathway and reduced LPS-induced apoptosis of M064 cells
In this experiment, we extracted EVs from MSCs by differential centrifugation for observation under transmission electron microscopy (TEM; Figure 5A), which revealed that the EVs were round-or   Collectively, our results demonstrated that MSCs-EVs carrying miR-378a-3p can repress the GATA2/AQP4/PPARα pathway to impede LPS-induced apoptosis of M064 cells.

| DISCUSS ION
Inflammatory bowel disease is regarded as a non-communicable disease lacking effective pharmacological therapy. 27 In the current study, we sought to determine the regulatory mechanism of miR-378a-3p-containing MSCs-EVs in IBD and identified that it could inhibit the progression of IBD by blockade of the AQP4/GATA2/ PPARα axis.
Initially, our findings elicited the elevation of GATA2 in the colon tissues of IBD patients and mice, and its silencing could reduce the LPSinduced apoptosis of M064 cells. An existing study investigated the role of GATA2 in IBD and determined that the GATA2 deficiency might influence the development of IBD. 28 Additionally, previous research identified that the transcription factor GATA2 serve as an inducer for the inflammatory reaction in colitis. 13 Furthermore, the current study demonstrated that the inhibition of GATA2 could contribute to downregulated AQP4 and blockade of PPARα, thereby reducing the LPSinduced apoptosis of M064 cells. Notably, the interaction between GATA2 and AQP4 and that between AQP4 and PPARα have been rarely reported. The results of the current study demonstrated a positive regulation in IBD in them. Besides, accumulating evidence has documented the participation of AQP4 and PPARα in the development of IBD. 23,29 For instance, the involvement of AQP4 was identified in the pathogenesis of intestinal dysfunction after traumatic brain injury due to regulation of intestinal oedema. 30 Suppression of PPARα action by PEA could aid in the inhibition of ulcerative colitis. 17 Strikingly, PPARα activated by fenofibrate treatment could evidently exacerbate inflammation and tissue injury in acute colitis. 31 Essentially, the current study also determined that MSCs-EVscontaining miR-378a-3p could reduce the LPS-induced apoptosis of M064 cells and the occurrence of IBD in mice. An existing study has elicited the potential application of MSCs and EVs in the treatment of IBD. Previously, MSCs have demonstrated therapeutic potential in IBD. 32 EVs elicit functionality in the aspect of immune mediation as well as intestinal barrier integrity and regulation of IBD. 33 Besides, it was previously reported that exosomes derived from human umbilical cord MSCs could radically attenuate IBD in a mouse model. 34 Moreover, EVs derived from human bone marrow-derived MSCs could evidently alleviate colitis in IBD models and thus have been proposed as potential candidates in the treatment of IBD. 35 Furthermore, several studies have identified the ability of EVs to deliver miRs to influence the progression of several diseases including those associated with IBD. For instance, EVs containing miR-146a could target TRAF6 and IRAK1 to ameliorate experimental colitis. 36 Moreover, miR-200b-containing microvesicles could evidently alleviate experimental colitis-associated intestinal fibrosis via suppression of epithelial-mesenchymal transition. 37 The role of miR-378a-3p in intestinal diseases has been previously investigated. Existing findings revealed that miR-378a-3p elevated IL-33 expression in an inflammatory environment, thereby inducing the pathogenesis of ulcerative colitis. 11 Besides, CDK6 inhibition targeted by miR-378a-3p can serve as a protective barrier against intestinal injury induced by ionizing radiation. 38 Mechanistically, our findings elicited that miR-378a-3p could target and inhibit the expression of GATA2. Therefore, after delivering M064 cells by MSCs-EVs, miR-378a-3p downregulated GATA2 to block AQP4 and PPARα, which radically inhibited IBD development.
To conclude, the key findings in this study elicited that MSCs-EVs could carry miR-378a-3p to inhibit the transcription factor GATA2, which can downregulate the expression of AQP4 and block PPARα signalling pathway, thus inhibiting the occurrence of IBD (Figure 7). This study may provide a novel target and theoretical basis for the clinical treatment of IBD. However, further investigations are warranted to validate our results for clinical application.

CO N FLI C T O F I NTE R E S T S
The author declares no conflict of interest exists.  The data and materials of the study can be obtained from the corresponding author upon request.