Exosomes Derived from Anoxia Preconditioned Mesenchymal Stem Cells alleviate Myocardial Ischemia Reperfusion Injury by Inhibiting Pyroptosis/ Caspase-1 Induced Apoptosis in The Deficiency of Gasdermin D


 BackgroundExosomes derived from adipose-derived mesenchymal stem cells can potentially protect cardiomyocytes from myocardial ischemia reperfusion injury. It's notable that exosomes derived from adipose-derived mesenchymal stem cells underwent anoxia preconditioning showed a better cardioprotective effect than that without anoxia. Here, in vitro and in vivo studies were used to investigate the cardioprotective effects against myocardial ischemia reperfusion injury of exosomes derived from adipose-derived mesenchymal stem cells with (Int-EXO) or without anoxia (NC-EXO), respectively. Methods: Adipose-derived mesenchymal stem cells and exosomes were identified by western blot, flow cytometry, transmission electron microscopy, and nanosight. An exosome tracer assay identified exosomes absorbed by cells. An in vitro model using mice cardiomyocytes for studying anoxia-reoxygenation and an in vivo mice model of MIRI were used to investigate the cardioprotective effects of NC-Exo and Int-Exo, respectively.ResultsWe discovered that treatment with NC-EXO and Int-EXO significantly reduced the infarct size and attenuated cardiomyocyte apoptosis, In addition, Int-EXO group had a less infarct size and apoptosis degree. The mechanism revealed by RNA sequencing showed that 40 miRNAs were up-regulated in Int-EXO compared to NC-EXO. 10 of these miRNAs could bind thioredoxin-interacting protein as a downstream target gene; among these, the top-discrepant miRNA224-5p was selected for further study. Dual luciferase reporter assay and rescue study verified TXNIP as a target gene for miR-224-5p. Furthermore, the cellular death signaling pathway which Int-EXO involved in mediating was in a special form of apoptosis, not pyroptosis, induced by activated thioredoxin-interacting protein-pyroptosis-caspase1 pathway in gasdermin D-deficient cells. ConclusionThe research demonstrated adipose-derived mesenchymal stem cells exosomes attenuated MIRI by inhibiting pyroptosis-induced apoptosis in cardiomyocytes which are lack of gasdermin D. The cardioprotective effect of Int-EXO was more significant than that of NC-EXO, possibly due to treated with anoxia preconditioning, adipose-derived mesenchymal stem cells product more miRNAs targeting thioredoxin-interacting protein in exosomes to alleviate pyroptosis-induced apoptosis. These findings provide new insights into the pathogenesis and methods for intervention of myocardial ischemia reperfusion injury.


Introduction
Myocardial ischemia reperfusion injury(MIRI) is caused by blood ow reperfusion after continuous myocardial ischemia due to excessive oxygen free radical formation, calcium overload, and other mechanisms that in uence the curative effect of reperfusion therapy [1]. Although great advancements in pharmacological and interventional therapies have been made to revascularize infarct-related arteries, MIRI, which is caused by secondary oxidative-stress damage in the myocardium, remains a major challenge in the clinical practice [2,3]. There is an unmet need to alleviate MIRI after infarct-related artery revascularization.
Emerging evidence points to a potential molecular mechanism for MIRI [4,5]. Secondary oxidative-stress damage in the myocardium triggered by excessive production of mitochondrial reactive oxygen species (mROS) is accompanied by revascularization of the infarct-related artery [4,6]. Moreover, sustained and excessive mROS production is a particular mediator of pyroptosis, which stimulates in ammasomes to join together and cleave caspase-1; this, in turn, triggers an in ammatory response that leads to the progression of MIRI [7]. Not only that, cleaved caspase-1 inhibits the activity of transcription factor GATA4 [8], which is closely related to heart development and is a link between pyroptosis and apoptosis to initiate comprehensive programmed cell death in the myocardium. Therapeutics targeting pyroptosis protect the ischemic-reperfusion injured heart from sterile in ammation, apoptosis, and dysfunction [9,10]. In addition, mROS has also been demonstrated to upregulate thioredoxin-interacting protein (TXNIP), which in turn represses the activity of hypoxia inducible factor-1, hence weakening myocardial tolerance to anoxia and inducing the formation of in ammasomes [11]. These ndings highlight the contributions of mROS in the induction of myocardial pyroptosis and apoptosis; however, the underlying mediators require further elucidation.
Exosomes, which are lipid bilayer extracellular vesicles with diameters ranging from 30 nm to 150 nm, are released by almost all types of cells including mesenchymal stem cells. They can modulate intercellular communications by transferring proteins, mRNAs, and miRNAs between cells, and they have the potential for cell-speci c targeting [12]. Exosomes derived from cells in different stages, states, species, and environments have distinct inclusions [13]. The process of exosomal action usually inherits the biological characteristics of the source cell to intervene in the trajectories of various physiological and pathological processes, including cell proliferation, differentiation, secretion, migration, and death. Changes in microenvironment and the state of cells can enhance or weaken the biological function of secreted exosomes. Exosomes have been reported to be important regulators of various cardiovascular diseases [14]. For example, exosomes derived from mesenchymal stem cells have the potential of protecting cardiomyocytes from MIRI and non-speci c in ammation [15]. Cardiac broblast-derived exosomes facilitate pathological cardiac hypertrophy via activating the renin angiotensin system in cardiomyocytes [16], and exosomes derived from cardiomyocytes contribute to cardiac brogenesis via myocyte-broblast cross-talk [17]. Obviously, the cell source of exosomes greatly determines their action modality and functional outcome.
Given the fact that anoxia preconditioning can enhance cardiomyocyte tolerance against an oxygende cient environment, in ammation, and oxidative stress [18,19], we hypothesized that exosomes derived from mesenchymal stem cells undergoing ischemic preconditioning have a more signi cant potential to ameliorate MIRI, oxidative stress damage, and sterile in ammation. We demonstrate that exosomes derived from adipose-derived mesenchymal stem cells (ADSCs) pretreated with anoxia preconditioning could be absorbed by cardiomyocytes and thereby ameliorate pyroptosis/caspase-1 induced apoptosis. Interestingly, these exosomes show a more signi cant cardioprotective effect than ADSCs without anoxia preconditioning. Furthermore, we showed that miR-224-5p is enriched in ADSCs-derived exosomes pretreated with anoxia preconditioning and that TXNIP is a downstream target of miR-224-5p, which can bind to and inhibit the antioxidant protein thioredoxin and is a key event linked to NLRP3 in ammasome via reactive oxygen species [20,21]. Inhibition of TXNIP can suppress the activation of in ammasomes and pro-caspase-1, thus ameliorating pyroptosis, perpetuating GATA4 levels in cardiomyocytes, and alleviating cardiomyocyte apoptosis, which just existed in GSDMD-de cient cells.

Animal studys and mice model of MIRI
All animal experiment procedures after appraisals, feasibility studies were approved by the Shanghai Ninth People's Hospital institutional ethics committee (SH9H-2017-A39-1). All experimental contents acted in accordance with the guidelines of the Directive 2010/63/EU of European Parliament. Six weeks old male speci c pathogen free(SPF) Wild-type (WT) C57BL/6 mice were purchased from Experimental Animal Center of Shanghai Ninth People's Hospital. The mice were fed normally under SPF conditions in experimental animal center until eight weeks. Mice were rstly induced anesthesia in an anesthesia box with 3% concentration of iso urane(Sinopharm Chemical reagent co.,Ltd, Shanghai, China) for 1 minites and then mice were maintained anesthesia with 1.5-2% concentration of iso urane through an anaesthetic mask. After mice were under complete anesthesia, invasive experiments were allowed to carry out. In order to minimize potential errors, animal operations(model of MIRI, Evensblue/TTC stain) were performed by Professor Erhe gao(MD,PhD,Director of Small Animal Surgery and Physiology Core, Lewis Katz School of Medicine, Temple University, USA). Mice were euthanized with excessive carbon dioxide inhalation.
A total of 80 healthy (WT C57BL/6) male mice (eight weeks old, ~ 24 g) were kept under SPF experimental animal feeding surroundings (temperature, 20-24 o C; humidity, 50-60), and fed with standard mouse chow and water ad with ad libitum feeding. Mice model of MIRI was operated under anesthesia with 1.5-2% concentration of iso urane through an anaesthetic mask by ligating the left coronary artery (LCA) for 30 min with a silpknot, then released the slipknot to induce reperfusion. Successful model of MIRI was con rmed by electrocardiograph(ECG) dynamic changes (ST segment elevation). After releasing the slipknot ligating LCA with a certain period of time, that is blood reperfution, hearts were harvested for further studys. A sham operated group was used as a control and underwent the same procedure, with the exception that the knot was left untied. For a detailed description of the procedure, please review Gao et al., 2010 [22].

Transthoracic echocardiography
Echocardiography was applied to assess on the baseline(day0), 1st, 3rd and the 7th day after reperfusion by using a Vevo 770 high-resolution imaging system. After mice were anaesthetized with 1.5-2% concentration of iso urane, echocardiography assessment was carried out to evaluate cardiac function. The ejection fraction (EF) and fractional shortening (FS) on the baseline(day0), 1st, 3rd and the 7th day were measured.

Evaluation of area at risk and infarct size
After reperfusion for 4 hours, the chest wall was reopened to expose the heart after mice were anaesthetized with 1.5-2% concentration of iso urane through mask. The LCA was re-ligated at the same degree of previous ligation, and 1% Evans Blue was injected through the ascending aorta with the distal artery nipped until the non-infarction zone turned blue. The heart was harvested, washed in saline, and sliced horizontally into ve pieces from left ventricular apex to the degree of ligation. All tissue pieces were incubated in 1% TTC for 20 min at 37 °C. Image Pro Plus 6.0 was used to calculate infarct area and area at risk(AAR). AAR/ left ventricle area (LV) and infarct size (IS) /AAR were analized.

Isolation and culture of mouse adipose-derived stem cells
Mesenchymal stem cells were harvested from C57BL/6 mice adipose tissue, as previously described 23 . In brief, adipose tissue isolated form proximal limb was digested by 0.075% collagenase IV(9001-12-1; Sigma, USA) for 60 min at 37 °C in a mechanical horizontal rotator at a speed of 180 rpm/min. The collagenase IV was dissolve in complete medium contained DMEM/F12(SH30023.01; Hyclone, USA) and 10% fetal bovine serum (FBS, ST303302, PAN, Germany). Then the mixture was centrifugated at a speed of 200 × g for 10 min. After centrifugation the supernatant was discarded and the cell pellet was resuspended in complete medium(DMEM/F12 and 10% exosome-free serum). The medium was placed in incubator at 37 °C in a 5% CO 2 atmosphere [23] for three days without taking out or shaking.
Cardiomyocytes were isolated from 1 to 2 day-old C57BL/6 mice as described previously [24]. In Brief, mice were disinfected quickly in a 75% ethanol solution, and hearts were quickly extracted and digested in 0.125% concentration trypsin diluted with PBS (15090-046,Gibco.USA) at 4 °C overnight. Subsequently, supernatant was collected and centrifuged at 200 × g for 5 min and the cell pellet was cultured in DMEM containing 10% FBS at 37 °C in a 5% CO2 atmosphere. After 1.5 hour of incubation to induce broblast adhering on dishes, then supernatant containing cardiomyocytes were plated in 0.1% gelatin-coated dishes in a density of 1 × 10 6 cells/mL. The purity of isolated cardiomyocytes was estimated by α-actinin staining (Fig. 1B).

Anoxia/ reoxygenation model of cardiomyocytes
To establish Anoxia/ reoxygenation model of cardiomyocytes, cardiomyocytes were incubated in oxygenfree DMEM medium(low glucoce) in an atmosphere content of 95%N2 and 5%CO2 for 2 h,then recovering oxygen and replacing oxygen-normal DMEM(low glucoce) medium for 1 h.

ADSCs treated with anoxia preconditioning
ADSCs were seeded at 1-2 × 10 6 cells/100-mm cell-culture dishes in exosome-free complete medium for 24 h, then medium was exchanged for DMEM/F12 which was incubated with 100%N2 overnight to remove oxygen existed in medium as possible. For anoxia preconditioning, the cells were subjected to repeated cycles of anoxia (60 min with oxygen-free medium) with intermittent reoxygenation (30 min with normal DMEM/F12) for ve cycles in an anoxic chamber(Forma-1025 Anaerobic System, Thermo, USA). Then ADSCs were incubated with serum-free medium in normal cultivation environment(21%O2, 5%CO2) and supernate was collected after 24 h.

Isolation and characterization of exosomes
Exosomes were isolated by differential centrifugation in conditioned media. In brief, the cell supernatant was centrifuged at 2,000 × g for 30 min at 4 °C to remove cell debris and cell fragments; cell supernatant was then collected and centrifuged at 100,000 × g for 70 min to precipitate exosomes to the bottom of centrifuge tube. After that, supernatant was discarded in order to remove the contaminated protein, and exosomes were suspended in PBS and centrifuged at 100,000 × g for 70 min. Finally, the sub-packaged precipitates were preserved at − 80 °C; repeated freeze/thaw cycles were avoided. Size distribution and concentration of exosomes were veri ed by a NanoSight NS300 instrument (Malvern Instruments, Malvern, UK), and the morphology of exosomes was de ned using transmission electron microscopy (TEM) (Fig. 1G). Expression of the exosomal markers TSG101, CD9 (positive controls), and calnexin (negative control) was assessed by western blot [25].

Exosomes Injection
Mice were randomly divided into 4 groups including sham group (n = 20), MIRI group (n = 20), NC-Exo group (n = 20), and Int-Exo group (n = 20). The groups were handled under the following conditions: 1) sham group mice undergoing sham surgery; 2) MIRI group mice undergoing coronary artery ligation for ligation for 30 min and drawing out the suture to loose the slipknot for reperfusion, then NC-Exo (10 µg per gram of body weight) were injected through the tail vein; 4) Int-Exo group mice undergoing coronary artery ligation for 30 min and drawing out the suture to loose the slipknot for reperfusion, then Int-Exo (10 µg per gram of body weight) were injected through the tail vein.

Analysis of exosomes uptake into cardiomyocytes
Exosomes and mice cardiomyocytes were labeled with PKH67 (MINI67, Sigma-Aldrich,USA) and phalloidin (A12379s, ThermoFisher,USA), respectively, according to the manufacturer's protocol. A total of 5 µL of PKH26-stained exosomes was added to the sample, followed by a 2 hours incubation for internalization by the cell. After washing twice with PBS, slides were xed in 4% paraformaldehyde and cell nuclears were stained with DAPI (C1005, Beyotime, China). Fluorescent images were collected by using a inverted microscope (BX63, Olympus, Japan) at x 630 magni cation (Fig. 1F).

Cell transfection
Cells were seeded in a 6-or 24-well plate, and transfected with miR-224-5p mimics, inhibitors, or plasmid DNA using a Lipofectamine 3000 transfection reagent (Cat. 11668019,Invitrogen,USA) following the manufacturer's protocol. Experiments were performed 24 h after transfection.

Immuno uorescence assay
Cells cultured on confocal plates were xed in 4% paraformaldehyde for 15 min, then washed with PBS for 3 times, and cells were permeabilized with 0.5% Triton X-100(C1016, Beyotime, China) for 15 min. Cells were blocked with a primary antibody dilution buffer for 1 hour at room temperature. Next, cells were incubated with mice anti-α-actinin primary antibody (1:1000, ab9465, Abcam,USA) overnight at 4 °C. Primary antibody was removed and the cells were washed twice with PBS. Then, the cells were incubated with a secondary antibody (1:1000, 4409, CST,USA) for 2 hours at room temperature. The cells nuclears were then stained with DAPI for 5 min at room temperature. Fluorescent images were collected by using a inverted microscope (BX63, Olympus, Japan) at x 630 magni cation.

Dual luciferase reporter assay
HEK 293T cells were cultured in a 24-well plate. When cell con uence reached approximately 70%, cells were transfected with PGL3 luciferase plasmids containing WT, NC and mutated TXNIP 3'UTR and miR 224-5p mimic (Ribobio, Shanghai, China). Then cells were re-cultured in a lucifugal 96well plates 12 hours later. After 36 hours, the ratio of re yluciferase and ranilla luciferase was detected using the Dual-GLO™ Luciferase Assay System (E2920,Promega,USA).

Real-time PCR
Total RNAs were extracted with RNAiso Plus extraction reagent following the manufacturer's instructions (No. 9108, Takara, Dalian, China). Stem-loop primers(purchased from Ribobio Biotech, guangzhou, China) were used to produce cDNA of miRNA by using PrimeScriptTM RT reagent Kit (RR047A, Takara, Dalian, China). The miRNA-cDNA was ampli ed by using TB Green® Premix Ex Taq™ II (RR820L, Takara, Dalian, China) through ABI-7500 Real-Time PCR Detection System (Applied Biosystems, USA). U6 was used as an internal control.

Western blotting
Protein was extracted from cardic tissues, cardiomyocytes. 10ug total protain was separated on 12% SDS-PAGE at 80 V for 1.5 hour, and then transferred to a PVDF membrane (IPVH00010,Millipore,USA) at 300 mA for 1 hour. After blocking. Membranes were incubated with primary antibodies at 4 °C overnight and secondary antibodies at room temperature for 1 hour. The primary antibodies used in the study included rabbit anti-TSG101 ( . Immunoreactive proteins were visualized using ECL substrate kit (ab65623, Abcam, USA) and two-color infrared uorescence imaging system(Odyssey CLX, LICOR, USA). Tubulin-α was applied as internal control.

Statistical analysis
Data analysis was performed through SPSS 19.0 software. Shapiro-Wilk test was applied to assess whether data conformed to normal distribution. Statistical analysis was performed using Pearson Chi square test(n ≥ 5) or Fisher exact test(n 5)with subsequent multiple comparisons using Chi square with Bonferroni correction for categorical variables. One-way ANOVA with subsequent post hoc multiple comparisons using the Student-Newman-Keuls test for continuous variables. Kruskal-Wallis test was applied to nonparametric testing of multiple independent samples and a Dunn-Bonferroni test for post hoc comparisons.

Identi cation of ADSCs and ADSCs-derived exosomes.
To examine whether ADSCs were successfully isolated from mice adipose tissue, we used CD29, CD45, and CD90 as cell surface markers, and CD34, CD105, and CD106 as negative controls. As shown in Fig. 1A, more than 95% of cells isolated from adipose tissue were positive for surface markers (CD29, CD45, and CD90) and hardly express the negative markers (CD34, CD105, and CD106). Subsequently, exosomes were obtained through sequential centrifugation and characterized by nanosight and transmission electron microscope and TEM to con rm the diameters of vesicles (ranging from 50 to 150 nm, with a median average size 115 nm) (Fig. 1C,E). The isolated vesicles also contained the exosome-associated markers (CD9, CD63, and tumor susceptibility gene (TSG101)), but did not contain calnexin (Fig. 1D).

Exosomes derived from ADSCs with anoxia preconditioning alleviated pyroptosis in cardiomyocytes
To con rm whether ADSCs-derived exosomes with anoxia preconditioning (Int-Exo) can more markedly in uence pyroptosis in MIRI,pyroptosis pathway was detected in cardiomyocytes. As shown in Fig. 2(A-B), Int-Exo and exosomes derived from untreated ADSCs (NC-Exo) decreased the expression of TXNIP, and cleaved-caspase-1 (TXNIP, cleaved-caspase1; NC-Exo vs IR group, p = 0.0007, p = 0.0002, respectively; Int-Exo vs NC-Exo group, p = 0.0194, p = 0.0007, respectively) 3.3. Exosomes derived from ADSCs with anoxia preconditioning ameliorated apoptosis in cardiomyocytes It has been proven that cleaved caspase-1 can deactivate GATA4 which is a key transcription factor closely related to heart development and inhibition of apoptosis [8]. To explore the biological functions of the pyroptosis-GATA4-apoptosis axis, we conducted an Annexin V-PI assay and Western blot to evaluate the in uence of Int-Exo and NC-Exo on cardiomyocyte apoptosis. As shown in Fig. 2(C-F)

Cardiomyocytes exhibit a lower expression level of GSDMD
To verify the nal result of activation of TXNIP-pyroptosis/caspase-1 pathway being in the form of apoptosis rather than GSDMD-induced-pyroptosis in cardiomyocytes, which may partly caused by the de ciency of GSDMD, we investigated the expression differences of GSDMD in cardiomyocytes and myeloid cells (Fig. 2G-H). Results of western blot exhibit that pretreated with anoxic/reoxygenation, no differences of cleaved GSDMD expression between CM and CM + IR group(p = 0.991). Meanwhile,stimulated with anoxic/reoxygenation, expression of GSDMD in RAW264.7 + IR group was signi cant higher than that in CM(p 0.0001), CM + IR group(p 0.0001) and RAW264.7 group(p 0.0001). In addition, no differences of total GSDMD expression were detected between CM and CM + IR group(p = 0.9898). However, total GSDMD expression in RAW264.7 group is much more higher than that in CM group(p 0.0001) and same trend was observed in CM + IR and RAW264.7 + IR group(p 0.0001).
3.5. ADSCs-derived exosomes with anoxia preconditioning ameliorated in-situ apoptosis in MIRI To further validate the anti-apoptotic effects of NC-Exo and Int-Exo in vivo, mice were intravenously injected with ADSCs-derived exosomes before being subjected to MIRI. In-situ apoptosis of myocardial tissue was estimated by Tunel assay. Figure 3A showed that the count of apoptotic cardiomyocytes in the Exo group was signi cantly less than that in the IR group. In particular, the Int-Exo group exhibited the lowerest percentage of apoptosis (NC-Exo vs IR group; p = 0.0003; Int-Exo vs NC-Exo group; p = 0.0141).
3.6. ADSCs-derived exosomes with anoxia preconditioning reduced myocardial infarct size Infarct size(IS) and AAR of the myocardium after MIRI in WT C57BL/6 mice pre-treated with ADSCsderived exosomes was evaluated. Figure 3B showed that the AAR/LV between the Exo groups, IR group, and sham group mice were similar, which indicated that the models were successful and comparable.
Furthermore, NC-Exo and Int-Exo groups exhibited signi cantly reduced infarct size compared to the IR group. Among them, the Int-Exo group had the most minimal infarct size (NC-Exo vs IR group; p = 0.0007; Int-Exo vs NC-Exo group; p = 0.0028).
3.9. ADSCs-derived exosomes with anoxia preconditioning improved cardiac function after MIRI Echocardiography revealed that the ejection fraction (EF) and the left ventricular fractional shortening (LVFS) in I/R mice treated with exosomes were signi cantly higher than those of I/R mice after 3 and 7 days of reperfusion Fig. 3(G-H), especially in the Int-Exo group (EF of 3rd day after reperfusion, NC-Exo vs IR group, p = 0.020, Int-Exo vs NC-Exo group, p = 0.036; FS of 3rd day after reperfusion, NC-Exo vs IR group, p = 0.033, Int-Exo vs NC-Exo group, p = 0.015; EF of 7th day after reperfusion, NC-Exo vs IR group, p = 0.022, Int-Exo vs NC-Exo group, p = 0.017; FS of 7th day after reperfusion, NC-Exo vs IR group, p = 0.058, Int-Exo vs NC-Exo group, p = 0.014).
3.10. Pyroptosis-related miRNAs were identi ed as a possible target of TXNIP To explore the underlying mechanism leading to the higher inhibiting effect of Int-Exo on pyroptosis (when compared to NC-Exo), microRNA sequencing was performed to explore the differential expression of miRNAs in the two types of exosomes. A total of 41 miRNAs were upregulated in Int-Exo compared to NC-Exo, including 10 that were predicted by target-scan, miR, and ENCORI to associate with TXNIP. The differential expression of these miRNAs was con rmed by Realtime-PCR, which is shown in Fig. 4A-B.

TXNIP is one of the target genes of miR-224-5p
Among the 10 upregulated miRNAs predicted to associate with TXNIP, miR-224-5p was selected for further study because its upregulation was the most signi cant. The results of Dual Luciferase Reporter assay to detect whether TXNIP was the target gene of miR-224-5p (Fig. 4C-D), the uorescence intensity in the miR-224-5p + TXNIP WT-3' UTR group was signi cantly reduced when compared to that of other groups (p = 0.009). These results indicated that miR-224-5p regulated translation of TXNIP by targeting to 3' UTR region of TXNIP.

miR-224-5p ameliorated apoptosis of cardiomyocytes with anoxia-reoxygenation by targeting TXNIP
A rescue study was performed to further validate whether miR-224-5p can ameliorate apoptosis in MIRI, partly by targeting the TXNIP-pyroptosis-GATA4 pathway. The results showed that miR-224-5p inhibitor decreased the expression of miR-224-5p, while increasing TXNIP expression (compared with the IR group,p = 0.022) in cardiomyocytes (Fig. 4E-F). The results of AnnexinV-PI (Fig. 4G-H) indicated that the apoptotic ratio of cardiomyocytes that underwent anoxia-reoxygenation was higher than that of the control group (p 0.0001) (Fig. 4G(a)) and 4G(b)). As the expression of miR-224-5p was suppressed, the apoptotic ratio of cardiomyocytes was further increased (compared to the IR group, p = 0.0115) (Fig. 4G(e)). Moreover, inhibition of TXNIP in cardiomyocytes that underwent anoxia-reoxygenation signi cantly reduced the apoptosis ratio (compared with IR group, p = 0.0025) (Fig. 4G(d)). However, if miR-224-5p is inhibited, when the expression of TXNIP is suppressed, the pro-apoptotic effect due to miR-224-5p de ciency could be reversed (compared to the IR + MiR224 Inhibitor group, p < 0.0001) (Fig. 4G(f)). These results demonstrated that miR-224-5p reduces the apoptosis of cardiomyocytes that underwent anoxia-reoxygenation, partly by targeting TXNIP.

Discussion
Exosomes are a e cient nanoscale-vehicles which have good characteristics such as excellent biocompatibility, low immunogenicity, and can cross through the blood-brain barrier. Given that exosomes, may "target to" speci c tissue and cells, targeted delivery of small-molecule and nucleic acid drug through exosomes should result in more effective treatment and fewer nonspeci c delivery. Thus, exosomes also have the potential to deliver cardioprotective non-coding RNA, cytokines and smallmolecule protein for the treatment of ischemic heart disease [26]. Although great advancements have been made, clinical transformation of exosomes still exsited some problems to overcome: (1) the technology of exosome isolation is not impeccable enough to meet clinical needs, (2) a synthetic vesicle and related technologies to edit exosomal content is provoking widespread concern over its biosecurity, (3) the long-time use safety of exosomes in clinical trials remains unclear [27]. Thus, our aim is to improve the biological e cacy of exosomes by using a secure anoxia-preconditioning method, thereby improving the cardioprotective effect of exosomes derived from mesenchymal stem cells, and to further elucidate the underlying mechanisms.
Remote ischemic preconditioning (RIPC) is de ned as repeated stimulation of the nonadjacent tissue, organ with transitory, noninvasive ischemia, which may reduce organ MIRI [28]. Paracrine action was partly thought to underlie the mechanism of RIPC, in which exosomes plays a core role [29], which enlightened us to treat ADSCs with intermittent anoxia. Our ndings revealed that exosomes derived from ADSCs that underwent anoxia preconditioning can serve as promising carriers for suppressing the activation of the cardiac pyroptosis pathway, the latter of which is a core element in inducing myocardial apoptosis. Importantly, in this study, RNA sequencing identi ed that there were approximately 40 miRNAs enriched in ADSCs-derived exosomes with anoxia preconditioning. According to bioinformatic analysis, 10 of these miRNAs critically mediate the pyroptosis pathway by targeting TXNIP and stabilizing the biological function of GATA4 to suppress the apoptosis of cardiomyocytes during MIRI. Thus, it is reasonable to assume that anoxia preconditioning on ADSCs could promote cells to load more miRNAs targeting TXNIP. Furthermore, the mechanism revealed in our study may partly explain the cardioprotective effect that Int-Exo exerts to reduce apoptosis of cardiomyocytes.
GATA4, as a transcription factor, with a characteristic protein structure of the zinc nger, has been proved as a vital cardioprotective factor and an indispensable mediator of adaptability of the myocardium to external pressure. When the expression of GATA4 is suppressed, toleration and survivability of cardiomyocytes against external pressure was affected, such as cardiotoxic drug and ischemic cardiomyopathy [30]. Furthermore, upregulation of GATA-4 would invert the trend of cardiomyocytes apoptosis [31]. Consequently, GATA-4 has emerged as an a potential target spot of therapeutic intervention to ischemic cardiomyopathy [32]. As transcription factor, GATA4 can activate the transcription of Bcl-2 through directly binding to Bcl-2 promoter region and up-regulate transcriptional level of Bcl-2, which inhibit cell apoptosis through 3 aspects: (1) Bcl-2 Inhibits leakage of glutathione to against oxidation state of thiol in mitochondria to stable mitochondrial membrane potential (MMP). (2) Bcl-2 regulates function of mitochondrial permeability transition pore to inhibit the release of cytochrome c(cyt-c) and apoptosis inducing factor(AIF). (3) Bcl-2 positions apoptosis protease activating factor-1(APAF-1) to mitochondrial membrane in order to inhibit activation of caspsase-3 by blocking the conbination of APAF-1 and cyt-c [33][34][35]. The results of precious studys indicated that regulating GATA-4 could ameliorate cardiomyocyte apoptosis in a early stage in a early stage when MMP appears collapse. Our data demonstrated that a relatively high-expression of GATA4 and Bcl-2 protein reduced the levels of apoptosis in cardiomyocytes and myocardium undergoing MIRI, after pretreatment with Int-Exo, which was partly induced by suppressing pyroptosis pathway and activation of caspase-1.
Recently, caspase-1 was famous for its proin ammatory function in pyroptosis pathway, videlicet, the primary molecules of execution of the pyroptosis process, during which caspase-1 cleaves and promote IL-1β, IL-18, and gasderminD (GSDMD) [36] to be activated state. Sepsis related activation of pyroptosis pathway was con rmed to have relevance to several cardiac diseases such as acute myocardial infarction, heart failure, and myocardial contractile dysfunction [37,38]. Moreover, actived caspase-1 is a upstream regulatory molecular of GATA4 to directly induce its inactivation. It is reported that P4 position containing hydrophobic amino acid is the ideal binding site of cleaved caspase-1. It's worth noting that GATA4 protein has such characteristics in two functional sites binded and cleaved by caspase-1, that is, YMAD 168 within the major transcription activation domain and WRRD 230 within the rst zinc nger were the target sites of caspase-1. Whichever the site was cleaved by caspase-1, transcriptional activity of GATA4 is bound to be affected [39]. Our ndings showed that speci c miRNAs transferred from Int-Exo to cardiomyocytes bound to the 3'UTR sequence of TXNIP, suppressed the pyroptosis-caspase1 pathway, thus protecting GATA4 from cleavage and alleviating apoptosis.
In most studies, various types of programmed cell death are considered independent, discrete, and unrelated. Apoptosis and pyroptosis are different in morphology, function, and mechanism [40]. On a cellular level, while membrane shrinking and the formation of vesicular bodies (apoptotic bodies) is the outcome of apoptosis. As for pyroptosis, it is a in ammasome-induced programmed cell death in form of loss of membrane permeability, membranolysis and release of in ammatory cytokines, triggered by GSDMD [41,42]. GSDMD was de ned as a functional executive molecule to cause cell death at the end of pyroptosis. Actived caspase-1 cleaves GSDMD between its N-terminal and C-terminal domains. Then the former was capable of bioactivities that actived GSDMD cause pores on cytomembrane to induce cell death [43]. Currently, it is general deemed that caspase-1 is an in ammation-related protein. A probably reason is that the majority of researches involved in in ammasomes applying in ammatory cell or myeloid cells as cell model that abundantly express GSDMD [44]. However, earlier studies indicated that actived caspase-1 could cause cell apoptosis in the de ciency of GSDMD through Bid induced cell apoptosis and other pathway [45,46]. In fact, caspase-1 was the vital role of both GSDMD-de cient and GSDMD-abundant programmed cell death pathways, in consideration of activation of pyroptosis/caspase1 pathway could induce apoptosis rather than pyroptosis in cell types that do not or only weakly express GSDMD [47,48]. We displayed a signal path diagram to graphically explain the pyroptosis/caspase1 induced apoptosis mediated by Int-Exo(Graphical Abstract). In our studies, as a type of non-myeloid cell, cardiomyocytes showed a relatively low expression of total GSDMD in protein degree.
Although with the activation of pyroptosis/caspase-1 pathway, the cleavage of GSDMD was also inapparent (Fig. 4), which were consistent with Tsuchiya K's results [45]. That is the de ciency of GSDMD blocks the last stage of pyroptosis-related membranolysis, thus the main manifestation of MIRI in cardiomyoctes still was apooptosis even with the activation of pyroptosis pathway. The pyroptosis/caspase-1-induced apoptosis occurs in cardiomyocytes through cleaving GATA4. Accordingly, the level of cleaved-caspase-1 in caspase-1-induced apoptosis under MIRI conditions may be inversely proportional to the expression of GATA4. Moreover, once miRNAs in Int-Exo bind and inhibit the expression of TXNIP, namely reducing the activation of the in ammasome and cleavage of caspase-1, the trend of inactivated GATA4 is reversed. The results reveal that pyroptosis-GATA4-apoptosis axis may exist and partly play a key role in ROS-induced MIRI.
Limitations still existed in our study: (1) Our results suggested exosomes derived from anoxia preconditioned ADSCs can perform a more obvious effect against MIRI, it is important to note that miRNA differences were not just limited in miRNAs targeting TXNIP, which maybe the main difference in int-Exo, such as protein, cytokines, other non-coding RNA may do anti-MIRI a favor. Thus we cannot exclude miRNAs' cardioprotection function through targeting other genes. (2) Due to the functional limitation of currently available antibody which could not speci cally binding to NLRP3 and ASC protein assembled into in ammasome. Thus western blot results in this research of NLRP3 and ASC just re ected whole protein degree in cells, which may explain non-signi cant expression difference in NLRP3 and ASC in IR and Exo groups. Further investigations are expected to elucidate the role of exosomes in protecting cardiomyocytes from MIRI.
In conclusion, our study indicated that ADSCs-derived exosomes that undergo intermittent anoxia preconditioning exerted a more signi cant cardioprotective effect against MIRI, because these exosomes were loaded with higher levels of miRNAs targeting TXNIP. In addition, as non-myeloid cells, activation of the pyroptosis pathway in cardiomyocytes would induce cell apoptosis in a GSDMD-de cient form rather than pyroptosis caused by the GSDMD protein. Thus, Int-Exo worked as a more effective anti-apoptotic factor by inhibiting the activation of the in ammasome to increase expression of GATA4. These ndings provide new insights into the pathogenesis of MIRI and methods for intervention at early stages of MIRI.
Ethics approval and consent to participate All animal experiment procedures after appraisals, feasibility studies were approved by the Shanghai  Figure 1 Identi cation of ADSCs and ADSCs-derived exosomes A)CD29, CD45, and CD90 were used as cell surface markers; CD34, CD105, and CD106 were used as negative controls. B) The purity of neonatal mice cardiomyocytes was estimated to be higher than 95% based on the results of the α-actinin staining. C) TEM characterization of exosomes. D) Western blot examination of exosome biomarkers in ADSCsderived exosomes and ADSCs calnexin were used as negative control. E) Size distribution by intensity was detected through nanosight. F) Exosome tracer assay was used to identify exosomes that were absorbed by cells.