Cardiomyocytes capture stem cell-derived, anti-apoptotic microRNA-214 via clathrin-mediated endocytosis in acute myocardial infarction

Extracellular vesicles (EVs) have emerged as key mediators of intercellular communication that have the potential to improve cardiac function when used in cell-based therapy. However, the means by which cardiomyocytes respond to EVs remains unclear. Here, we sought to clarify the role of exosomes in improving cardiac function by investigating the effect of cardiomyocyte endocytosis of exosomes from mesenchymal stem cells on acute myocardial infarction (MI). Exposing cardiomyocytes to the culture supernatant of adipose-derived regenerative cells (ADRCs) prevented cardiomyocyte cell damage under hypoxia in vitro. In vivo, the injection of ADRCs into the heart simultaneous with coronary artery ligation decreased overall cardiac infarct area and prevented cardiac rupture after acute MI. Quantitative RT-PCR–based analysis of the expression of 35 known anti-apoptotic and secreted microRNAs (miRNAs) in ADRCs revealed that ADRCs express several of these miRNAs, among which miR-214 was the most abundant. Of note, miR-214 silencing in ADRCs significantly impaired the anti-apoptotic effects of the ADRC treatment on cardiomyocytes in vitro and in vivo. To examine cardiomyocyte endocytosis of exosomes, we cultured the cardiomyocytes with ADRC-derived exosomes labeled with the fluorescent dye PKH67 and found that hypoxic culture conditions increased the levels of the labeled exosomes in cardiomyocytes. Chlorpromazine, an inhibitor of clathrin-mediated endocytosis, significantly suppressed the ADRC-induced decrease of hypoxia-damaged cardiomyocytes and also decreased hypoxia-induced cardiomyocyte capture of both labeled EVs and extracellular miR-214 secreted from ADRCs. Our results indicate that clathrin-mediated endocytosis in cardiomyocytes plays a critical role in their uptake of circulating, exosome-associated miRNAs that inhibit apoptosis.

Substantial improvements in primary prevention efforts to reduce risk factors of cardiovascular disease, including smoking, low-density lipoprotein cholesterol, and blood pressure, have decreased the incidence of myocardial infarction (MI) 3 (1); however, cardiovascular disease, including ischemic heart disease, remains the leading cause of mortality worldwide. Cellbased therapy represents a potential frontier to transform the treatment and prognosis of heart failure and acute myocardial infarction through repair of damaged cardiac tissue (2). Numerous early-phase randomized clinical trials have investigated the clinical benefits of cell transplantation in patients with acute MI. Bone marrow cells represent the most commonly utilized cellular source, although the beneficial effects of their transplantation in patients with MI are controversial (3). Recently, mesenchymal stem cells and cardiac progenitor cells have served as alternate cellular sources of cell-based therapy (4). Cardiac progenitor cells are capable of cardiac regeneration (5); moreover, their paracrine effects improve tissue resistance to ischemic stress (6). Mesenchymal stem cells produce and secrete a broad range of cytokines, chemokines, and growth factors that are potentially involved in cytoprotection and angiogenesis (7). In cell-based therapy, both cell types secrete paracrine factors in extracellular vesicles (EVs), such as exosomes and microvesicles, to improve cardiac function (8). The concept of EVs as messengers in intracellular communication implies that EVs secreted from endothelial cells and cardiac fibroblasts may interact with cardiomyocytes to induce physiological changes (4). Although EVs derived from endothelial cells, cardiac fibroblasts, and transplanted cells have emerged as key mediators of intercellular communication to improve cardiac function (9), the mechanisms by which cardiomyocytes effect transmembrane signaling in response to EVs remain unclear. Here, we report that clathrin-mediated endocytosis in cardiomyocytes plays a critical role in transporting circulating microRNAs (miRNAs) into the cell.

Adipose-derived regenerative cells (ADRCs) protect against ischemia-induced cardiac dysfunction
To investigate whether mesenchymal stem cells effect cardiac protection through secretion of paracrine factors, isolated mouse adult cardiomyocytes under hypoxia conditions were treated with the supernatant of bone marrow-mononuclear cell (BM-MNC) or ADRC cultures (Fig. 1A). After 6-h exposure to normoxia or hypoxia, hypoxia-damaged cardiomyocytes were ascertained by trypan blue vital staining. Hypoxia decreased cardiomyocyte cell viability, whereas this response was more effectively ameliorated by treatment with the supernatant from ADRC culture than that from BM-MNCs.
To determine whether treatment with ADRCs might attenuate the progression of myocardial ischemia-induced cardiac dysfunction in mice, we ligated the left coronary artery in mice. ADRCs or BM-MNCs were injected in the heart simultaneously with coronary artery ligation, and infarct changes were determined in triphenyltetrazolium chloride (TTC)-stained sections 2 days after the ligation (Fig. 1B). ADRC treatment decreased infarct area in the heart in a dose-dependent manner (Fig. 1C). Moreover, echocardiography to assess cardiac functionality showed that ADRC treatment protected mice from cardiac dysfunction (Fig. 1D). The therapeutic effect afforded by ADRCs was similar to that with BM-MNCs. In addition, the treatment reduced MI-induced increases of brain natriuretic peptide (BNP) and atrial natriuretic peptide (ANP) in the heart, which are elevated in heart failure ( Fig. 1, E and F), but did not recover MI-induced decreases of myosin heavy chain isoforms. Next, we investigated whether the ADRC treatment resulted in beneficial effects after MI. Notably, starting at 3-7 days after the coronary artery ligation, control mice showed increased mortality compared with the ADRC treatment group (Fig. 1G). Necropsy indicated that 15 of 27 mice with MI died with the presence of a blood clot within the chest cavity, which results from cardiac rupture ( Fig. 1H and Fig. S1A). Conversely, ADRC treatment reduced cardiac rupture following coronary artery ligation.
The most dramatic complications of acute MI involve cardiac rupture of acutely infarcted tissue, with the tear usually preceded by a large infarct with subsequent expansion and occurring near the junction of the infarct and normal muscle (10). To examine whether ADRC treatment reduced cardiac apoptosis induced by coronary artery ligation, we identified apoptotic cardiomyocytes by terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining of heart sections (Fig. S1B). ADRC treatment significantly reduced apoptosis in cardiomyocytes at 1 and 2 days following coronary artery ligation (Fig. 1I). Additionally, although mesenchymal stem cells secrete angiogenic factors and enhance angiogenesis in ischemic hind limbs (11), ADRC treatment did not significantly enhance angiogenesis in the heart (Fig. S1C). These findings suggested that ADRC treatment reduces cardiomyocyte apoptosis following coronary artery ligation, ultimately preventing MI-induced cardiac rupture.

Cardiac protection by ADRC-secreted miR-214
EVs facilitate intercellular communication in physiological and pathological processes and contain various factors, including growth factors, cytosolic signaling proteins, and extracellular matrix proteins. Exosome-meditated transfer of miRNA has emerged as a novel mechanism of genetic exchange between cells. We investigated the expression of 35 anti-apoptotic and secreted miRNAs in cardiomyocytes, BM-MNCs, and ADRCs (Table S1). Quantitative RT-PCR (qRT-PCR) revealed that these cells all express various anti-apoptotic and secreted miRNAs with miR-214 being the most abundant in ADRCs ( Fig. 2A). qRT-PCR analysis of various tissues and cells indicated that miRNA-214 is exclusively expressed in ADRCs (Fig.  2B). miRNA-214 and miRNA-134 expression was significantly increased in cardiomyocytes treated with ADRC culture supernatant compared with that in controls (Fig. 2C); this was further increased under hypoxia conditions. In vivo, ADRC treatment increased miR-214 expression in the heart 1 and 2 days after coronary artery ligation (Fig. 2D). To examine whether miR-214 derived from ADRCs was increased in cardiomyocytes, an miR-214 inhibitor or miRNA inhibitor negative control was nucleofected into ADRCs (Fig. 2E). miR-214 inhibitor significantly decreased miR-214 expression in ADRCs and suppressed the increase of miR-214 expression in cardiomyocytes induced by treatment with ADRC supernatant (Fig. 2F), suggesting that miR-214 in the supernatant of ADRC cultures was captured in cardiomyocytes. miR-214 silencing in ADRCs significantly impaired its anti-apoptotic effects in vitro (Fig. 2G).
To examine the effect of miR-214 in cardiomyocytes, rat neonatal cardiomyocytes under hypoxia were treated with an miR-214 mimic (Fig. 2H). This treatment significantly reduced hypoxia-induced apoptosis, suggesting that miR-214 has a protective function in cardiomyocytes. Moreover, ADRCs treated with an miR-214 inhibitor decreased the cardiac protective effects of ADRCs, whereas ADRCs treated with an miR-214 mimic enhanced the effects (Fig. 2, I-K). miR-214 has been shown to improve cardiomyocyte survival in MI through regulating miRNA expression, such as Bcl2l11 and Slc8a1 (12), potentially because miRNA association with its mRNA target generally results in target mRNA degradation (13). qRT-PCR analysis showed that the expression of Bcl2l11 and Slc8a1 increased 2 days after MI, whereas ADRC treatment significantly suppressed expression of these genes (Fig. 2L). Conversely, miR-214 blockade in ADRCs significantly impaired the ADRC-mediated effects on gene expression. These findings suggested that miR-214 secreted from ADRCs promotes cardiomyocyte survival after MI.

Clathrin-mediated endocytosis in cardiomyocytes
miRNAs secreted from cells can affect gene expression in target cells via exosomes (14). To examine whether exosomes secreted from ADRCs are transported into cardiomyocytes, exosomes were isolated from the ADRC cultures ( Fig. 3A) and labeled with the fluorescent dye PKH67. Cardiomyocytes were then cultured with the labeled exosomes (Fig. 3B). Three-dimensional reconstruction of confocal microscopy images showed the labeled exosomes in the cytoplasm of the

MicroRNA endocytosis in cardiomyocytes
pressed ADRC supernatant-induced decrease of hypoxiadamaged cardiomyocytes in a dose-dependent manner (Fig.  3H). These results demonstrated that clathrin-mediated endocytosis in cardiomyocytes mediates capture of vesicles including miR-214, ultimately rescuing cardiomyocytes from hypoxia damage.

Clathrin-mediated endocytosis after MI
AP2, which triggers clathrin assembly, constitutes the central node for regulating the initiation of clathrin-coated pits. Adap-tor association kinase 1 (AAK1) phosphorylates AP2 at Thr-156 to stabilize the conformational changes that activate AP2. Immunoblotting showed that Thr-156 phosphorylation of AP2 in the heart increased upon MI, suggesting that clathrin-mediated endocytosis is activated under these conditions (Fig. 4, A  and B). To investigate whether communication between cardiomyocytes and ADRCs affords cardiac protection through clathrin-mediated endocytosis in vivo, male mice were treated with ADRCs alone or in combination with chlorpromazine following coronary artery ligation (Fig. 4C). Notably, chlorprom-

MicroRNA endocytosis in cardiomyocytes
azine treatment failed to show ADRC-induced cardiac protection (Fig. 4, C-E). To determine whether exosomes are transported into the cardiomyocytes in vivo, PKH67-labeled exosomes were injected in the heart concomitant with coronary artery ligation (Fig. 4, F-I). A single injection promoted the spread of PKH-labeled exosomes across an ϳ700 ϫ 700 ϫ 700-m cubic area in the heart (Fig. 4, F and G). Among 786 PKH67-labeled exosomes in the heart, 67.7% of the PKH67labeled exosomes were in ␣-actinin-positive cardiomyocytes, and 32.3% were in ␣-actinin-negative noncardiomyocytes or extracellular area. MI increased the intensity of the labeled exosomes in the heart, suggesting that ischemic conditions enhanced endocytosis in vivo. Conversely, chlorpromazine treatment suppressed the ischemia injury-induced cardiac endocytosis (Fig. 4, H and I). These results demonstrated that clathrin-mediated endocytosis is activated during acute MI and plays a critical role in communication between cardiomyocytes and ADRCs in cell-based therapy.

Discussion
miRNAs comprise a class of small regulatory RNAs that anneal with complementary sequences in target mRNAs, usually suppressing protein translation (17). Numerous studies have shown that miRNAs correlate with cardiovascular diseases (18,19), thus constituting potential therapeutic targets (13,20). Exosome-mediated miRNA transfer represents a novel mechanism of genetic exchange between a donor and recipient cell (21). Our qRT-PCR analysis showed that miR-214 is abundantly expressed in ADRCs and that treatment with ADRC supernatant increased its expression in cardiomyocytes and reduced cardiac rupture after acute MI. Similarly, transfection of miR-214 protects against hydrogen peroxide-induced cardiomyocyte apoptosis in vitro (22). Chronic miR-214 overexpression has no morphological effect on the heart (23), whereas its genetic deletion in mice increases apoptosis induced by ischemia-reperfusion and causes deterioration of cardiac function in vivo (24). In vitro, miR-214 secreted from bone marrowderived mesenchymal stem cells protects cardiac stem cells from oxidative damage (25). We found that miR-214 secreted from ADRCs suppressed genes potentially involved in cardiac apoptosis, such as Bcl2l11 and Slc8a1. For example, in response to ischemia, Bcl-2-associated X protein (Bax) activity eventually results in mitochondria-dependent cardiac myocyte death via the caspase protease system (26). The Bcl2l11-encoded protein induces apoptosis by activating Bax or neutralizing antiapoptotic proteins (27). Finally, the Na ϩ /Ca 2ϩ exchanger encoded by Slc8A1 causes Ca 2ϩ overload-associated apoptosis in cardiomyocytes under cardiac stress (24). Left ventricular rupture occurs as a consequence of excessive inflammation and matrix metalloproteinase activity after acute MI (28,29). Although the relationship between left ventricular rupture and infarct size is unclear (30), apoptosis has been reported to be involved in left ventricular rupture after MI (31,32). These findings suggest that miR-214 -mediated anti-apoptotic effects contribute to preventing cardiac rupture.
EVs, which contain a subset of proteins, lipids, and small RNAs to facilitate cell-to-cell communication in the regulation of cancer and the immune system (14,33), are emerging as potential therapeutic delivery agents and recognized as playing important roles in cardiac repair following MI (34). EV-mediated function in MI, the types of donor cells that secrete EVs, and their biochemical composition have been investigated (4,8,12). However, the particular endocytosis mechanism utilized by cardiomyocytes to obtain EV contents was unknown (35). Our findings using endocytosis inhibitors revealed that cardiomyocytes receive circulating miR-214 in EVs through clathrin-mediated endocytosis, which is regulated by multiple factors across initiation, cargo selection, maturation, and fission steps (Fig. 4J) (16). AP2 triggers clathrin assembly and plays a central role in endocytosis initiation (36). Specifically, AAK1mediated AP2 phosphorylation causes conformational changes in AP2 that enhance its membrane-binding ability (37). The generation of reactive oxygen species in response to hypoxia causes phosphorylation of AP2 by AAK1 in opossum kidney cells (38). We found that AP2 was phosphorylated in the heart after MI, suggesting that hypoxia conditions may trigger clathrin-mediated endocytosis. Our data further show that dynamin, which regulates endocytic vesicle fission in the final step of clathrin-mediated endocytosis (39), contributes to the ADRC supernatant-induced decrease of hypoxia-damaged cardiomyocytes. The damaged cardiomyocytes with activated clathrin-mediated endocytosis may uptake EVs more aggressively than normal cardiomyocytes, suggesting that ADRC-derived EVs would serve as a useful tool to preferentially deliver miRNAs into the damaged cardiomyocytes. In conclusion, our data show that miRNA-214 in EVs secreted from ADRCs prevents cardiac rupture induced by acute MI and that clathrin-mediated endocytosis in cardiomyocytes plays a critical role in transporting the circulating miRNAs into the cell.

In vitro experiments
Adult mouse cardiomyocytes were isolated as described previously (40). The culture media were removed, and adherent cells were cultured with DMEM or conditioned medium from ADRCs or BM-MNCs. Inhibitors were added at 1 h after cell plating and mixed with DMEM or conditioned medium. For hypoxia studies, cells were exposed to hypoxia for 6 h. Hypoxic conditions were generated by using an AnaeroPack system MicroRNA endocytosis in cardiomyocytes (Mitsubishi GAS Chemical, Tokyo, Japan). Neonatal rat ventricular myocytes (NRVMs) were isolated from 1-2-day-old rat neonates (41). Primary NRVMs were incubated in DMEM with 10% fetal bovine serum.
Cell viability was assessed by a trypan blue exclusion assay. In 6 h of culture under normoxic or hypoxic conditions, trypan blue (0.4%; Gibco) was added into the culture dish (final concentration, 0.04%). The number of viable (unstained) and nonviable (blue-stained) cells was recorded in 10 random microscopic fields.

Animal experiments
MI was induced in 8-week-old male C57BL/6J mice (Nihon Crea, Tokyo, Japan) as described previously (42). After the left anterior descending artery was ligated, mice were administered an intramyocardial injection of cells (60 l of 0.5 ϫ 10 4 , 5 ϫ 10 4 , or 50 ϫ 10 4 ADRCs in PBS) or PBS alone (60 l) into two infarct border zones and two infarct zones (15 l in each site) with a 30-gauge needle. Mice were sacrificed at 2 or 28 days after surgery. The heart was incubated with TTC to determine the infarcted area. Transthoracic echocardiography was performed on mice using a Vevo 1100 (Fujifilm VisualSonica Inc., Bothell, WA) with a 40-MHz probe. The cause of death was determined by necropsy, and cardiac rupture was diagnosed by the presence of a blood clot within the chest cavity as described previously (43).

mRNA and microRNA expression analysis
To analyze mRNA expression, RNA was extracted from periinfarcted areas in the left ventricle using an RNeasy Mini kit (Qiagen, Chatsworth, CA) following the manufacturer's protocol and reverse-transcribed using the ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan). qRT-PCR analysis was assessed using SYBR Premix Ex TaqII (TaKaRa Bio, Shiga, Japan) and the following primers MicroRNA was obtained by using the miRNeasy Micro kit (Qiagen) and reverse-transcribed to cDNA using the miScript II RT Kit (Qiagen) following the manufacturer's protocol. qRT-PCR analysis was performed using the miScript SYBR Green PCR kit (Qiagen) with the miScript universal primer and miRNA-specific forward primers. All miScript Primer Assays (see Table S1) were purchased from Qiagen. Data are presented following normalization with Rnu6b.

Isolation and analysis of exosomes
ADRCs were cultured in DMEM supplemented with 10% exosome-depleted fetal bovine serum (Gibco, A25904DG) over 48 h. ADRC-derived exosomes were isolated from the conditioned medium of ADRCs using ExoQuick-TC exosome precipitation solution (EXOTC10A-1, System Biosciences, Palo Alto, CA) following the manufacturer's protocol. The protein levels of MFG-E8 and syntenin, a representative marker of exosomes, were detected by Western blotting.

Exosome labeling and uptake by cardiomyocytes
ADRC-derived exosomes were labeled using the PKH-67 green fluorescent cell linker kit (PKH67GL, Sigma-Aldrich). Following incubation for 5 min, exosomes were washed several times using an Amicon Ultra-0.5 centrifugal filter device (nominal molecular weight limit 100,000) (Millipore, Billerica, MA). Adult mouse cardiomyocytes were incubated with the labeled exosomes for 1-2 h with or without hypoxia. Exosome uptake was quantified as a percentage of fluorescence intensity relative to the reference of 2-h culture without hypoxia. Exosome localization was viewed using a BZ-X710 microscope (Keyence, Osaka, Japan) with a sectioning technique.
To study exosome uptake in vivo, mice were administered an intramyocardial injection of ADRC-derived exosomes into two infarct border zones (15 l in each site) following MI. Some mice were injected with the mixture of exosomes and 10 mol/ liter chlorpromazine to study the pathway of exosome uptake. Cellular uptake of ADRC-derived exosomes by cardiomyocytes was observed using a BZ-X710 microscope. Exosome uptake was quantified as the percentage of fluorescence intensity relative to the reference of sham mice injected with exosomes without 10 mol/liter chlorpromazine.

Study approval
All procedures of animal care and animal use in this study were approved by the Animal Ethics Review Board of Nagoya University School of Medicine.

MicroRNA endocytosis in cardiomyocytes Statistical analysis
Data are presented as the means Ϯ S.D. for given experiments. All statistical analyses for experiments were performed using GraphPad Prism 6 (GraphPad Software, La Jolla, CA). Comparisons between two groups were performed using an unpaired Student's t test, whereas those between more than two groups utilized ANOVA with Tukey's post hoc testing. Survival curves were analyzed with Kaplan-Meyer estimators and logrank (Mantel-Cox) testing. Significance was defined as p Ͻ 0.05 (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001).