IFN-γ Enhances the E cacy of Exosomes Derived from Mesenchymal Stem Cells in Myocardial Infarction Rats via miR-21 Stimulated by STAT1

Jian Zhang The First A liated Hospital of Nanjing Medical University Yao Lu The First A liated Hospital of Nanjing Medical University Yangming Mao The First A liated Hospital of Nanjing Medical University Yue Yu The First A liated Hospital of Nanjing Medical University Tianyu Wu The First A liated Hospital of Nanjing Medical University Wei Zhao The First A liated Hospital of Nanjing Medical University Yeqian Zhu The First A liated Hospital of Nanjing Medical University Pengcheng Zhao The First A liated Hospital of Nanjing Medical University Fengxiang Zhang (  njzfx6@njmu.edu.cn ) The First A liated Hospitol of Nanjing Medical University https://orcid.org/0000-0003-0125-3926

Introduction IFN-γ produced by activated T cells and natural killer (NK) cells plays physiologically key role in maintaining innate and adaptive immunity responses. 1 As a cytokine that can promote immunomodulation, IFN-γ has been widely studied for its excellent anticancer activity. 2 After interacting with the recipient cells through IFN-γ receptor (IFNGR), IFN-γ subsequently activates downstream signal transduction pathways and transcriptionally stimulates the expression of various genes involved in immune regulation and other biological activities. 3 Studies have illustrated that MSCs activated with IFNγ elicit more powerful immunomodulatory effects by the upregulation of immunoactive factors. 4,5 MSCs are stromal cells with self-renewal ability and multi-differentiation ability, 6 and have been used for cardiovascular diseases as regenerative therapy. 7 Recent studies have shown that the therapeutic effects of MSCs are mainly derived from the paracrine mediated by Exos. 8 The Exos are 50-200 nm vesicles secreted into the extracellular space, and shuttle a variety of microRNA (miRNA), long noncoding RNA (lncRNA) and proteins to perform the function of cell-cell communication. 9 Previous studies have discovered that MSC-derived Exos had the function of treating ischemic heart disease by inhibiting cell apoptosis, promoting angiogenesis and regulating macrophage polarization. 10,11 Subsequently, relevant studies suggest that Exos derived from modi ed MSCs exhibit more powerful protective e cacy. Wang et al. reveal that compared with normoxic condition, Exos derived from hypoxic condition exert better ability to inhibit cell death through miR-125b. 12 Adiponectin stimulates the release of Exos to enhance the treatment of heart failure in mice driven by MSCs. 13 Recently, our studies prove that hypoxia induction and macrophage migration inhibitory (MIF) modi cation can strengthen the therapeutic capacity of MSC-derived exosomes on myocardial infarction (MI) through inhibiting apoptosis and promoting angiogenesis. 14 However, whether Exos derived from MSCs stimulated by IFN-γ perform more e cient protective effects against MI than Contrl-Exo remains unknown.
Therefore, Ctrl-Exo and IFN-γ-Exo were extracted to treat both H9c2 and HUVECs cells under OGD condition and rat MI models to explore their therapeutic effects and potential mechanisms. On this basis, we found that IFN-γ treatment can heighten the anti-apoptotic ability and angiogenesis of Exos derived from MSCs to better preserve heart function. This progress was partly achieved by the increased expression of cardioprotective miRNA-21 through IFN-γ treatment. This Exo-mediated delivery of miRNAs from MSCs under various treatment conditions might be an effective alternative for MI treatment.

Ethics statements
This study protocol conforms to the Guide for the Care and Use of Laboratory Animals [National Exosomes Extraction and Characterization MSCs (1×10 6 ) were cultured to 70% con uence and then treated with 6ml of Exosome-free FBS (Gibco) for 2d. The cell culture medium (6ml) was collected in a 15ml centrifuge tube and centrifuged at 1,500 × g for 30 min to remove the cells and debris. Transferred the processed supernatant to another 15ml centrifuge tube containing 2ml RiboTM exosome extraction reagent (for cell culture medium, C10130-2, Ribobio, China). These mixtures were incubated overnight at 4°C, and then centrifuged at 2,000 ×g for 30 min. The supernatant was discarded, and the exosomal pellet was resuspended in 100μl PBS.
The exosomal surface markers were detected using western blotting with anti-TSG101, CD63, and CD81 antibodies (Abcam, UK). The external electron microscope (TEM) was utilized to observe the appearance of Exos. The Exos were xed with 1% glutaraldehyde, and then coated with 1% phosphotungstic acid on a copper mesh. JEM-2100 transmission electron microscope (JEOL, Tokyo, Japan) was used to observe the sample. Nanoparticle tracking analysis (NTA) was applied to analyze the size and distribution of Exos. We recorded and tracked the Brownian motion of Exos in PBS (Carlsbad, California). The particle size distribution data was obtained with Stoke-U.S. A ZetaView PMX 110 system (Particle Metrix, Germany) was used for NTA.

Exosomes uptake assay
In order to demonstrate the uptake of Exos by H9c2 cells and HUVEC cells, Exos were labeled with Dil (red uorescent dye, C1036, Beyotime, China) and co-cultured with recipient cells at 37°C for 6h or 24h and then washed with PBS and x with 4% paraformaldehyde for 20 min. The nuclei were stained with 6diamino-2-phenylindole (DAPI) (0.5g/ml, Beyotime) for 10 min, and observed with a confocal microscope.

Apoptosis assays
Flow cytometry (KeyGEN Biotech, China) detected cell apoptosis. H9c2 cells were seeded in 1×10 5 /6-well tissue culture plates overnight, and treated with different Exos or PBS before OGD. The cells were washed with PBS and stained with Annexin V uorescein isothiocyanate and propidium iodide apoptosis kit (KeyGen Biotech, China). Flowjo Software version 10.0 (Tree Star, USA) was used to analyze apoptotic cells. TdT-mediated dUTP Gap End Labeling (TUNEL) Apoptosis Detection Kit (Roche, USA) was also used to detect cell or tissue apoptosis according to the manufacturer's instructions. The formula for calculating the percentage of apoptotic nuclei was the total number of TUNEL stained nuclei divided by the total number of TUNEL positive nuclei.

Migration Assay
Cultured HUVECs in a 6-well plate, scraped the con uent layer with the tip of a P200 pipette. Then washed and incubated the cells after adding 100 µg/well of different Exos. Images were taken before and 6h, 12h after incubation, and Image J software (NIH) was used to determine the reduction ratio of the scratch area.

Tube Formation Assay
HUVECs were treated with PBS or different Exos. Then cells were washed with PBS and seeded (30,000 cells/well) in 96 well plates coated with growth factor reduced Matrigel (Corning, United States). After 6h, capillary-like tube formation was observed and photographed. Tube length and number of branches were analyzed with Image J software (NIH).

Quantitative real-time PCR (qRT-PCR)
The total cellular and exosomal RNA was extracted using Trizol reagent (Life Technologies, USA) according to the manufacturer's instructions. 14 A stem-loop-speci c primer method was used to measure miR-21-5p expression, as described previously. 20 The sequences of primers used in the study were shown in Table S1. The relative expression was calculated using the following equation: relative gene expression = 2− (ΔCtsample − ΔCtcontrol). All samples were measured in triplicate.

Transfection experiment
Transfection of miR-21-5p mimics (50 nmol/L) and negative control miRNA (50 -100 nmol/L) synthesized by Guangzhou Ribobio were carried out into H9c2 cells using riboFECT TM CP Reagent (Ribobio, China) according to the instructions of manufacturer. The full-length BTG2 sequence and empty vector as negative control were inserted into a pcDNA 3.1 plasmid (GenePharma, China) to transfect H9c2 cells using Lipofectamine 2000 (Invitrogen, USA) according to the manufacturer instructions. STAT1 siRNA and Negative control FAM manufactured by Suzhou GenePharma were delivered into MSCs by Lipofectamine 2000. qRT-PCR were performed to determine transfection e cacy. At 48h after transfection, different groups of cells were harvested for further study.

Western blot
Protein extraction and Western blot (WB) analysis were performed as previously stated. 14 Brie y, cells were washed with PBS and lysed with lysis buffer on ice for 20 minutes. The total cell protein concentration was detected by the BCA Protein Assay Kit. The total protein (20μg) was separated using SDS-PAGE (Invitrogen) and transferred to a PVDF membrane (Roche). The membrane was blocked with 5% bovine serum albumin (0.1%) in TBS-Tween and incubated against the desired antibody. The primary antibodies used for Western blot analysis are listed in Table S2. Bands were visualized using enhanced chemiluminescence reagents and analyzed using a gel documentation system (Bio-Rad Gel Doc1000 and Multi-Analyst version 1.1).

MI Model, Histological Analysis, and Immuno uorescence Staining
Eight-week-old male Sprague-Dawley (SD) rats obtained from the experimental animal center of Nanjing Medical University were randomly divided into 4 groups: sham operation group (Sham group, n=6), PBS injection group (PBS group, n=6), Ctrl-Exo injection group (Ctrl-Exo group, n=6), and IFN-γ-Exo injection group (IFN-γ-Exo group, n=6). Brie y, as previously described 14 , left anterior descending artery (LAD) was ligated, and Exos (50µL, 1µg/µL) or PBS was injected around the infarct area in rats. All surgeries and subsequent analyses were blinded for intervention. Echocardiography (Vevo 3100) was performed to determine the left ventricular ejection fraction (LVEF) and left ventricular short axis shortening rate (LVFS) after 2w and 4w.
The rats were sacri ced 4w after surgery. In ammatory cell in ltration and cell arrangement was evaluated by Hematoxylin-Eosin (HE) staining. Masson's trichrome staining was used to evaluate brosis and collagen area after MI, CD31 immuno uorescence staining was to observe the distribution of microvessels. Apoptosis was detected by TUNEL staining (Roche, USA). The primary antibody was anti-CD31 (ab7388; British abcam). DAPI was used for nuclear counterstaining. The images were further analyzed using a uorescence microscope (Zeiss, Germany) and Image J software (NIH).

Statistical Analysis
Continuous variables and categorical variables were described as mean ± SD and percentages, respectively. Independent-Sample T-test was used to compare continuous variables between the two groups. One way Analysis of variance (ANOVA) followed by Tukey's correction was used for comparison of three or more groups. All statistical tests were performed using GraphPad Prism software version 8.0, and p < 0.05 was considered statistically signi cant.

Characterization of control and IFN-γ-primed Exos
Exosomes were isolated from the control MSCs and 50 ng/ml IFN-γ-primed MSCs supernatant and identi ed the appearance using TEM. The pictures showed that the Ctrl-Exo and IFN-γ-Exo were typical lipid bilayer membrane encapsulated nanoparticles with a diameter between 30 and 150 nm (Fig. 1A). The marker proteins TSG101, CD81, and CD63 were positively expressed in two groups (Fig. 1B). Fig. 1C showed that the peak diameter of Exo were 123.4 nm and 116.2 nm in Ctrl-Exo and IFN-γ-Exo group, respectively. NTA analysis showed that IFN-γ treatment can increase the number of Exos secreted by MSCs. Similarly, IFN-γ increased the protein concentration in the Exo suspension observed by western blot.
The labeled Exos with Dil dye were co-cultured with HUVEC cells and H9c2 cells for 6h and 24h respectively. Confocal microscopy showed that Dil-labeled Exos can be observed around the nucleus within 6h, and most of the Exos were absorbed within 24h (Fig. 1D, E). In summary, both control MSCs and IFN-γ-primed MSCs can secrete Exos with common vesicle characteristics, and these Exos can be absorbed by H9c2 cells and HUVEC cells in a time-dependent manner in vitro.

Pro-angiogenesis and anti-apoptotic effects of IFN-γ-Exo in vitro
Both TUNEL staining and ow cytometry showed that the percentage of apoptotic cells were signi cantly reduced in IFN-γ-Exo group compared with OGD and Ctrl-Exo group under OGD condition ( Fig. 2A, B). Fig.  2C showed that the apoptosis-related proteins Bax and Cleaved-caspase-3 were signi cantly reduced in the IFN-γ-Exo group, while the anti-apoptotic protein Bcl2 increased compared with the OGD and Ctrl-Exo group. Angiogenesis and migration rate of HUVECs signi cantly increased in the IFN-γ-Exo group compared with control and Ctrl-Exo groups (Fig. 2D, E). These results suggested that IFN-γ-Exo confer superior protective effects on H9c2 and HUVECs compared to Ctrl-Exo in vitro.
IFN-γ-Exo exerted stronger cardioprotection against myocardial damage than Ctrl-Exo in vivo AMI rat model was used to determine the cardioprotective effects of MSC-derived Exos. Ctrl-Exo, IFN-γ-Exo (100 µl, 1 µg/µl), or 100 µl PBS were intramyocardially injected at the time of surgery (Fig. 3A). Dillabeled Exo was intramyocardially injected and then colocalized with cardiomyocytes 6h after injection, suggesting an e cient in vivo uptake of the Exos by heart tissue (Fig. 3B). LVEF and LVFS from 4 weeks' echocardiography were signi cantly improved in IFN-γ-Exo group compared with PBS and Ctrl-Exo groups (Fig. 3C), while 2 weeks' result showed no difference between Ctrl-Exo and Ctrl-Exo groups (Fig. S1). Quanti cation of the infarct area suggested that IFN-γ-Exo can maximize the reduction of the brotic scar area after MI (Fig. 3D). The above results indicated that IFN-γ-Exo has a better therapeutic effect against myocardial ischemia and hypoxia injury compared with Ctrl-Exo.
IFN-γ-Exo promote angiogenesis and cardiomyocyte survival in vivo HE staining showed that the degree of in ammatory cell in ltration in the IFN-γ-Exo group was lower than that of the Ctrl-Exo and PBS group (Fig. 4A). TUNEL analysis showed that the proportion of apoptotic cells signi cantly decreased in the IFN-γ-Exo group compared with PBS and Ctrl-Exo groups (Fig. 4B). CD31 staining found that the number of regenerative blood vessels in the IFN-γ-Exo group was signi cantly higher than that of the other two groups (Fig. 4C). Combined with in vivo and in vitro experiments, we found that IFN-γ-Exo can exert a better protective effect on myocardium after ischemia and hypoxia injury.
We compared the expression of putative ve miRNA in control and IFN-γ-primed MSCs and the results (Fig. S2) showed that miR-21 was the most signi cantly elevated miRNA in IFN-γ-primed MSCs compared with the control ones (Fig. 5A). Further, we found that miR-21 was signi cantly enriched in IFN-γ-Exo compared with Ctrl-Exo (Fig. 5B).
In order to investigate the role of miR-21 in regulating apoptosis, miR-21 mimics and negative control (NC) were transfected into H9c2 cells prior to exposure to OGD conditions. Apoptosis caused by ischemia and hypoxia were signi cantly abrogated by markedly increased miR-21. Moreover, western blot analysis con rmed that the protective effect of miR-21 against OGD-induced injury relied on the upregulation of Bcl2 and decreasing Bax and Cleaved-caspase-3 (Fig. 5C).

STAT1 activator promoted expression of miR-21 in IFN-γprimed MSCs
In order to explore whether transcription factors (TF) were involved in the transcriptional regulation of miR-21, we employed Illumina HiSeq 2,500 high-throughput sequencing for mRNA expression pro ling of control and IFN-γ-primed MSCs to identify the functional transcription factor in IFN-γ-primed MSCs. With a 2-fold change and P < 0.05 as the threshold cutoff, we identi ed 55 signi cantly differentially expressed TFs, of which 37 were upregulated in IFN-γ-primed MSCs (Fig. 5D). Next, we used the TransmiR v2.0 database 15 (http://www.cuilab.cn/transmir) to predict the transcription factors that might regulate miR-21 (Fig. 5E) and then crossed with the upregulated expressed TFs to nally get 3 promising TFs (STAT1, STAT2, and FOXC1) (Fig. 5F). Among them, consistent with sequencing results, STAT1 was the most signi cantly increased in IFN-γ-primed MSCs compared with control MSCs veri ed by qPCR (Fig. 5G).
Consistent with previous studies, 16, 17 we found that STAT1 were induced by IFN-γ in MSCs and cellular miR-21 were signi cantly upregulated. Further, the IFN-γ induction of miR-21 was abolished following STAT1 downregulation (Fig. 5H). In addition, we cloned the wild-type fragment miR-21 promoter into the upstream region of the luciferase reporter gene (pGL3-basic) to obtain pGL3-miR-21 WT and found that pGL3-miR-21 WT was activated by STAT1 overexpression. These results suggested potential transcriptional regulation of miR-21 by STAT1 (Fig. 5I).

BTG2 was identi ed as a target gene of miR-21 and promoted H9c2 cell apoptosis under OGD conditions
Using TargetScan and miRDB, BTG2 was identi ed as a potential target for miR-21 (Fig. 6A). Luciferase reporter assays con rmed the association between miR-21 and BTG2: the luciferase activity of BTG2-wt in miR-21 transfected cells was signi cantly inhibited, while that of the BTG2-mut remained unchanged (Fig. 6B). Furthermore, the mRNA and protein levels of BTG2 were signi cantly decreased in the miR-21 mimic groups, respectively (Fig. 6C, D). The above results indicated that there is a negative regulatory relationship between miR-21 and BTG2.
Subsequently, BTG2 was overexpressed to explore its effect on H9c2 cells apoptosis under OGD conditions. Fig 7A showed that western blot analysis con rmed that BTG2 was successfully overexpressed in H9c2 cells via transfection with BTG2 plasmid. Further experiments revealed that apoptosis induced by ischemia and hypoxia is further aggravated by BTG2 overexpression. The expression of Bax and Cleaved-caspase-3 protein were increased signi cantly in the BTG2 group, while the expression of Bcl2 protein was decreased. The same trend was veri ed in TUNEL staining and ow cytometry, that is, the incidence of apoptosis in the BTG2 group was signi cantly increased compared with ctrl (OGD) and vector group (Fig. 7B, C).
In order to deeply explore the functional relationship between miR-21 and BTG2, we used reply experiment to verify the effect of miR-21 mimics and BTG2 plasmid on H9C2 under OGD conditions. Consistent with the previous ndings 18 , ischemia and hypoxia not only increase the expression of Bax and Cleavedcaspase-3 but also increase the expression of BTG2, and this elevation were inhibited by miR-21 mimics. However, the anti-apoptotic effect of miR-21 on H9C2 cells under OGD conditions were obviously abrogated by BTG2 overexpression (Fig. 7D). These results indicated that the overexpression of BTG2 reverses the protective effect of miR-21 on OGD-treated H9c2 cells.

Discussion
This study found that compared with Ctrl-Exo, IFN-γ-Exo accelerated migration, tube-like structure formation, and prevented H9c2 from OGD-induced apoptosis; IFN-γ-Exo also reduced brosis size, cardiomyocytes apoptosis and improved recovery in cardiac function; Further research has revealed that IFN-γ-Exo attenuates OGD-induced injury in H9c2 cells through upregulating miR-21 expression stimulated by STAT1, which directly targeted on BTG2 (Fig. 8).
Growing evidence shows that Exos released from MSCs can protect ischemic cardiomyocytes from death, improve ventricular remodeling, and preserve heart function. 19 On this basis, how to improve the therapeutic effect of Exo is a point worthy of our consideration. Researchers have devoted a great deal of efforts to nding a variety of modi cation methods that can reinforce the therapeutic effect of MSCderived Exos. Hypoxia, as the most common treatment, elicited MSC-derived Exos exhibit more effective cardioprotection mediated by UCA1 and miR-125b. 12,20 Genetic modi cation can also enhance the treatment of heart disease by adjusting the expression of various miRNA. Exos secreted by MSCs overexpressing GATA-4 retain a large amount of anti-apoptotic miRNA for cardioprotection. 21 The Exos derived from MSCs modi ed by MIF can treat acute MI by limiting apoptosis and promoting angiogenesis. 14 Some drugs and cytokines have also been selected to improve the characteristics of stem cells. Atorvastatin, as a common clinical lipid-lowering drug, has also been proven to increase the therapeutic effect by up-regulating the expression of H19. 22 IFN-γ, as an immune-related cytokine, enable MSC-derived Exos to show stronger immunomodulation e cacy and therefore perform better in the treatment of colitis. Our research disclosed that Exo derived from IFN-γ-primed MSCs have more su cient anti-apoptosis and angiogenesis ability compared with the control group both in cardiomyocytes under OGD conditions and MI rat models. These results indicated that IFN-γ treatment may be an alternative to strengthen the therapeutic effect of exosome derived from MSCs, so we have conducted a more in-depth study on the changes in the contents of Exos.
Minimizing the loss of myocardial cells and restoring the microcirculation in the marginal zone of the infarction are common strategies for the treatment of MI. 23 miR-21 is a typical cardioprotective miRNA, which has various therapeutic effects such as anti-apoptosis, anti-brosis, and inhibiting in ammation. 24 Subsequently, in order to explore the mechanism of miR-21 being upregulated we used RNA sequencing to explore the transcription factors that were upregulated after IFN-γ treatment and nally selected STAT1 that may regulate the expression of miR-21. Some studies have reported that STAT3 can stimulate the expression of miR-21 by binding to the promoter region. 25,26 As an important member of the signal transduction and transcription activator protein family, STAT1 has a similar structure and function with STAT3. 27 Our research revealed that STAT1 induced by IFN-γ can upregulate miR-21 expression by binding to the promoter region.
The angiogenesis effect of miR-21 has been extensively studied, so we further explored the anti-apoptotic effect of miR-21 under ischemia and hypoxia. 28, 29 BTG2 (B cell translocation gene 2) is the rst gene found in the BTG/TOB gene family, and it exerts a tumor suppressor effect in various cancer types. 30 Previous studies have found that down-regulation of BTG2 by miR-21 can protect cardiomyocytes from doxorubicin treatment. 31 Functional studies have shown that overexpression of BTG2 can exacerbate the apoptosis of H9c2 cell apoptosis under ischemic and hypoxic conditions. Moreover, BTG2 reversed the protective effect of miR-21 on hypoxia-induced injury in H9c2 cells. At the same time, BTG2 was found to be up-regulated under OGD conditions, which suggests that BTG2 plays an important role in myocardial infarction, so it is worthy of in-depth study.

Conclusions
In summary, in this study, we have documented that Exos derived from control and IFN-γ-primed MSCs attenuate myocardial injury. IFN-γ-Exo exerted superior therapeutic e cacy, which were mainly mediated by increasing the expression of miR-21-5p. Brie y, our ndings provide insight into promoting the ability of anti-apoptosis and angiogenesis from MSC-derived Exos and suggest a promising approach to treat ischemic heart disease. in Ctrl-Exo and IFN-γ-Exo groups. (C) The particle size distribution and particle concentration were analyzed by nanoparticle tracking analysis (n=3). Confocal images showed that red uorescence of dye levels are presented as the average expression normalized to β-Tublin (n=3). (D) Migration was monitored for 6 and 12 h after scratching in HUVECs cultured with PBS, Ctrl-Exo and IFN-γ-Exo (n=3). (E) Tube formation of HUVECs incubated with PBS, Ctrl-Exo and IFN-γ-Exo (n=3). Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA followed by Bonferroni's correction. *P< 0.05, **P < 0.01, ***P< 0.001, NS not signi cant.  groups. Quantitative analysis of the apoptotic rate at the border zone among the different groups (3 random elds per animal; n=5). Data are presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA followed by Bonferroni's correction. *P< 0.05, **P< 0.01, ***P < 0.001, ****P< 0.0001.  Relative protein levels were presented as the average expression normalized to β-Tubulin (n=3). Data were Page 23/25 presented as mean ± SEM. Statistical analysis was performed with one-way ANOVA followed by Bonferroni's correction. *P< 0.05, **P< 0.01, ***P < 0.001, ****P< 0.0001. nuclei; blue, DAPI-stained nuclei. Scale bars=200μm. And ow cytometric analysis (n=3) (C). (D) H9c2 cells treated with control, OGD condition, OGD condition+miR-21 and OGD condition+miR-21+BTG2.