Mitochondrial Transfer through Pericardium Contributes to Functional Recovery in Ischemic Cardiomyopathy

Although mesenchymal stem cell transplantation has been ecacious in the treatment of ischemic cardiomyopathy, the underlying mechanisms remain unclear. Herein, we investigated whether mitochondrial transfer could explain the success of cell therapy in ischemic cardiomyopathy. Mitochondrial transfer was examined in co-cultures of human adipose-derived mesenchymal stem cells and rat cardiomyocytes under hypoxic conditions. Functional recovery was monitored in a rat model of myocardial infarction following human adipose-derived mesenchymal stem cell transplantation. In vitro, we observed mitochondrial transfer, which required formation of cell-to-cell contacts and synergistically enhanced energy metabolism. Rat cexhibited mitochondrial transfer three days following human adipose-derived mesenchymal stem cell transplantation to the ischemic heart surface post myocardial infarction. We detected donor mitochondrial DNA in the recipient myocardium concomitant with a signicant improvement in cardiac function. In conclusion, mitochondrial transfer is vital for successful cell transplantation therapies and promotes improved treatment outcomes in ischemic cardiomyopathy.


Introduction
Ischemic heart disease is one of the leading cause of death worldwide. Various cell therapies have been applied to remediate acute myocardial infarction and ischemic cardiomyopathy. The modes of action of these therapies include angiogenesis, cardiomyocyte regeneration, protection from ischemia reperfusion injury, and activation of hibernating myocardium, mediated by cytokines, exosomes, and microRNAs secreted from transplanted cells. However, the mechanisms underlying these responses remain to be fully elucidated [1][2][3][4].
Recently, a phenomenon, termed mitochondrial transfer, has been described in animal models of acute lung injury following mesenchymal stem cell (MSC) transplantation. Mitochondrial transfer is thought to exert protective effects on recipient alveolar epithelial cells, thereby representing a potential cell therapy mode of action against heart failure [5][6][7][8][9]. Furthermore, bidirectional mitochondrial transfer has been observed in cardiomyocytes co-cultured with MSCs [5,10,11]. Cardiomyocytes receiving mitochondria show elevated expression of mitochondrial proteins, increased oxidative phosphorylation, and improved ATP production [12,13]. Meanwhile, in vivo, mitochondrial transfer is expected to confer resistance to cytotoxicity, exert anti-apoptotic effects, and increase ATP production in recipient cardiomyocytes, thus facilitating an overall improvement in myocardial function [14][15][16].
To the best of our knowledge, no previous study has evaluated whether mitochondrial transfer occurs in vivo following cell transplantation to the heart. Thus, the role of mitochondrial transfer in the success of cell transplantation and its effect on heart function remains unknown. In this study, we demonstrated mitochondrial transfer from transplanted cells to cardiomyocytes in vivo and proposed this as the mechanism underlying successful cell therapies for ischemic cardiomyopathy (Fig. 1).

Results
Donor-derived mitochondrial DNA is detected in cardiomyocytes in vivo Three days following transplantation of hADSCs harboring pre-stained mitochondria, we detected stained mitochondria in the recipient rat heart and epicardium ( Fig. 1A-D). Electron microscopy revealed tunnellike connections between hADSCs and cardiomyocytes ( Fig. 1E-G). Laser microdissection was used to obtain heart samples for DNA sequencing (Fig. 1H). The mitochondrial genome fragment was PCRampli ed, and the resultant amplicon length was consistent with those in previous reports (Fig. 1I) [21].

Mitochondrial transfer occurs within 24 h following transplantation
For intravital imaging, we transfected hADSCs with CellLight to induce mitochondrial expression of red uorescent protein ( Fig. 2A-B and Supplementary Video S1A online). Upon transplantation, the graft formed a layer on the heart where hADSCs were equally distributed ( Fig. 2C and Supplementary Video S1B online). Three to 4 h post transplantation, mitochondria from the hADSCs were detected in the myocardial layer of the recipient ( Fig. 2D and Supplementary Video S1C online). Residual cardiomyocytes in the brotic area contained granules that were considered to represent the transferred mitochondria ( Fig. 2E-F). When co-culture of hADSCs and cardiomyocytes was performed for 24 h under hypoxic conditions, the hADSCs were found to move around the cardiomyocytes in a xed eld of view ( Fig. 3A-B and Supplementary Video S2A online). In the cardiomyocytes, the intensity of hADSC-derived mitochondrial luminescence signi cantly increased during the 24-h co-culture (Fig. 3C). The xed sample also exhibited granular hADSC-derived mitochondria in the rat cardiomyocytes (rCM) ( Fig. 3D and Supplementary Video S2B online).
hADSC co-culture synergistically enhances the OCR of cardiomyocytes The OCR of rCMs was synergistically enhanced by hADSC co-culture, being signi cantly higher than the sum of the OCRs of rCMs and hADSCs cultured independently. This effect was cancelled by the addition of a non-speci c gap junction inhibitor αGA ( Fig. 3E-F). We successfully obtained adult rCMs using a Langendorff circuit and captured the motion of individual rCMs ( Fig. 3K-M). An additional 24-h culture with supernatant hADSCs had no effect on the contraction of rCMs (Fig. 3N).
hADSC transplantation improves cardiac function in a dose-dependent manner We monitored luciferin signals using an in vivo imaging system at least four weeks post transplantation. On the fourth week, the heart surface was observed directly (Fig. 4B-C). Histological analysis of these samples revealed surviving hADSCs on the surface or in the connective tissue between the heart and chest wall (Fig. 4D). The supernatant of cultured and grafted hADSCs contained various cytokines, including VEGF, HGF, interleukin (IL)-6, and IL-10. Among them, secretion of VEGF and IL-6 was stimulated by hypoxia (Fig. 4E). hADSC transplantation exhibited a therapeutic effect on cardiac function in the infarcted rats. The effect was temporary in low dose-and medium-dose groups, limited to the early phase following transplantation. In the max-dose group, improvement was greater than that in the rest at weeks 2 -6. However, at week 8, cardiac function improved in both the low-dose and medium-dose groups equally. No effect against cardiac remodeling was observed, as indicated by left ventricular (LV) enddiastolic diameter (LVDd) and LV posterior wall thickness; however, the LV end-systolic diameter (LVDs) tended to become smaller with increasing dose ( Fig. 4F and Supplementary Fig. S1 online). Although the addition of αGA to hADSCs blocked the improvement of cardiac function three days post transplantation, it did not affect the long-term improvement (Fig. 4G). The optimal transplantation time was found to be 1 -2 weeks following the onset of myocardial infarction ( Supplementary Fig. S2A-J online).
hADSC transplantation increases intramyocardial mitochondrial DNA and ATP No difference was observed in brotic area across the groups during the rst two weeks following implantation. However, a difference emerged after eight weeks. Additionally, semiquantitative assessment revealed less interstitial brosis in the hADSC-treated group compared to the control group. rCM diameter was signi cantly smaller in the high-dose and maximum dose groups, and the highest capillary densities and lowest collagen accumulation were observed in the medium-dose group ( Fig. 5A-B). Mitochondrial number and cristae structure were relatively normal compared to that in the untreated group, as indicated by electron microscopy. Other organelles, such as the endoplasmic reticulum, myo bers, Z-disk, and intercalated disks, were morphologically preserved (Fig. 5C). Following treatment, the amount of rat mitochondrial DNA in the myocardium increased with time. However, mitochondrial DNA from the infarct area exhibited a decrease by day 56 post treatment ( Fig. 5D). Donor-derived mitochondrial genes were not su ciently ampli ed for detection. One day after transplantation, cytosolic ATP levels in the hADSC-treated group were approximately ten-fold higher than those in the control group ( Fig. 5E).

Discussion
Following transplantation of hADSCs onto the ischemic heart surface, donor-derived mitochondrial DNA was detected in the DNA extracted from cardiomyocytes, providing the rst evidence of mitochondrial transfer from ADSCs to the ischemic myocardium in vivo. This mitochondrial transfer occurred shortly following transplantation, as evidenced by histological evaluation and intravital imaging, suggesting an in uence on early improvement of cardiac function. Furthermore, ATP concentration in the intramyocardial infarct area increased 24 h post transplantation and stabilized after a week. Previous reports have suggested retention of these mitochondria for approximately 1 week [28], that is consistent with our ndings. However, direct histological identi cation of mitochondria over time was challenging. Although the detailed mechanisms underlying mitochondrial transfer have not been established, our data con rm cell-cell communication and indicate transfer of mitochondria and other organelles between cells [29]. Speci cally, when cardiomyocytes were exposed to hADSC supernatant, no immediate contraction-enhancement effect was observed on cardiomyocytes, suggesting that a direct cell-cell communication is necessary for mitochondrial transfer to occur between hADSCs and cardiomyocytes [30]. In vitro experiments using a co-culture system demonstrated that ADSCs interact with and transfer mitochondria to cardiomyocytes, thereby improving their metabolic capacity, as measured using OCR. Moreover, addition of a connexin blocker to this in vitro co-culture system prevented the OCR enhancement effect of mitochondrial transfer and reduced the ability of mitochondrial transfer to improve cardiac function in vivo, suggesting the involvement of connexin in cell-cell communication underlying mitochondrial transfer. However, the survival period of transferred mitochondria, the signaling pathway leading to mitochondrial transfer, and the detailed mechanisms of recipient cell activation remain to be elucidated.
Ischemic diseases are often ameliorated by angiogenesis; however, this can lead to the dysfunction of myocardial and cardiac cells. Reactivation of these cells can result in long-term improvement of cardiac function. Although the transplanted cells survived on the surface of the heart for approximately four weeks, the transferred mitochondria could not be identi ed in heart specimens thereafter. Rather, the transferred mitochondria either disappeared within a few days following transplantation or were present in very low numbers. This suggests that hADSC treatment improves the pathophysiology of ischemic cardiomyopathy, including mitochondrial abnormalities, by providing a short-term supply of relatively healthy mitochondria and normalizing the long-term morphology and quantity of mitochondria in the recipient myocardium [14].
Regarding the pathophysiology of mitochondrial cardiomyopathy, mitochondrial protein abnormalities can cause chronic structural abnormalities in heart tissue. Thus, if transferred mitochondria activate pathological mitochondria, it may be bene cial for the subject [31][32][33][34][35]. In the present study, structural improvements, such as orientation of myocardial bers, suppression of CM hypertrophy, and suppression of brosis (remodeling suppression) were observed. These histological changes could be induced by cytokines from the transplanted cells. Moreover, considering the downward trend of ejection fraction eight weeks post transplantation, in the αGA group, mitochondrial transfer could be considered to contribute not only to short-term, but also long-term improvements in CM function.
Various transfer modes have been proposed for mitochondrial transfer, including involvement of microvesicles, tunneling nanotubes, cell fusion, and potential cell surface gap junction proteins [36][37][38]. In vitro, danger-associated molecular pattern proteins, mtDNA, Ca 2+ , CD38 signaling, NAD + , and NADH, released from damaged cells, act to trigger signals and actively promote mitochondrial transfer from donor cells via modes shown in Graphical Abstract [39]. Additionally, evaluation of the direct myocardial transplantation of isolated mitochondria suggest a mechanism whereby cardiomyocytes directly internalize mitochondria through the cell membrane. These processes are considered to be non-speci c and re ect passive transfer [40][41][42][43]. Herein, mitochondrial transfer was promoted by hypoxic stimulation in vitro, and mitochondrial transfer from the myocardium to hADSCs was prevented by a connexin blocker in vivo. Therefore, ischemic myocardial signaling may cause cell-cell communication via gap junctions, after which mitochondrial transfer occurs.
In this study, however, the transfer of mitochondrial DNA alone, migration of transplanted cells into the myocardium along with subsequent fusion with recipient cells, and mitochondrial transfer via microvesicles could not be excluded. The presence or absence of a signal from the recipient CM to the donor cell (to assess whether mitochondrial transfer is active or passive) was not determined, and any potential signal was not speci ed. Moreover, the amount and period of retention of transferred mitochondria, the mechanisms underlying subsequent increase in intramyocardial ATP, and the short-and long-term improvement of cardiac function remain to be fully elucidated. Nevertheless, our current ndings support the hypothesis of cell transplantation delivering living mitochondria directly to myocardial cells, along with cytokines that promote angiogenesis, inducing protective effects in the myocardium [8]. Taken together, our results suggest that ADSC-to-CM mitochondrial transfer occurs both in vitro and in vivo, thereby contributing to the recovery of early cardiac function, following ADSC transplantation in ischemic cardiomyopathy. Thus, enhancement of the mitochondrial transfer mechanisms may be expected to lead to enhanced e cacy of cell therapy.

Assessment of in vivo mitochondrial transfer
Transplantation of adipose tissue-derived stem cells (ADSCs) harboring stained mitochondria A myocardial infarction model in F344/NJcl-rnu/rnu rats (CLEA Japan, Inc., Tokyo, Japan) was generated by permanent ligation of the left anterior descending artery (LAD). Two weeks following LAD ligation, human ADSCs (hADSCs; 1 × 10 6 cells; Lonza, Tokyo, Japan) were transplanted into the infarct area using brinogen and thrombin solution (CSL Behring, Tokyo, Japan). The composition of this graft is shown in Supplementary Table S1 online [17]. The hADSCs were pre-stained with MitoTracker Red (Thermo Fisher Scienti c, Waltham, MA, USA) [18]. The rats were euthanized at appropriate time points following transplantation, and heart samples were collected (Fig. 6). Following formalin xation, para n-embedded heart sections were stained with an antibody against phalloidin (Cat. #A12379; Thermo Fisher Scienti c) and counterstained with Hoechst 33342 (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) (Supplementary material).

Observation of graft-heart boundaries with electron microscopy
Mitochondrial ultrastructure and spatial relationship between hADSCs and rat cardiomyocytes (rCMs) were assessed in tissue samples collected from areas including the boundary between heart and hADSC graft. Contrasted sections were imaged under an H-7500 transmission electron microscope (Hitachi, Tokyo, Japan) (Supplementary material).
Intravital imaging of mitochondrial transfer from hADSC graft to beating myocardium Myocardial infarction-induced adult male CAG/eGFP transgenic Sprague-Dawley (SD) rats (body weight, 200 -250 g; Japan SLC, Inc., Shizuoka, Japan) were used as recipients of hADSC transplants. The hADSCs were transfected using CellLight TM Mitochondria-RFP BacMam 2.0 (Thermo Fisher Scienti c) ( Fig. 2A). Graft composition and transplantation was performed as explained above (Supplementary  Table S1 online). The beating heart and the graft were stabilized using a custom-made suction device (Fig. 2B) [19]. Clear in vivo images were obtained by methods reported previously [20]. Observation was initiated immediately following transplantation and continued for maximum possible duration (Supplemental material).

Detection of donor mitochondrial DNA in recipient rCMs
Para n-embedded tissues from the heart and hADSC graft were harvested three days post transplantation. Tissue sections were stained with hematoxylin and eosin, and residual rCMs around the brotic tissue and just beneath the hADSC graft were collected using a laser-equipped microscope (Leica LMD7000). DNA was isolated, and a human mitochondrial DNA fragment was ampli ed using a customized primer pair speci c for human mitochondria (Euro ns Genomics, Louisville, KY, USA) (Supplementary Table S2

Measurement of oxygen consumption rate
Oxygen consumption rate (OCR) in living cells was measured using a Seahorse XF24 Extracellular Flux Analyzer (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's protocol. rCMs (2 × 10 4 cells/well) were cultured on 96-well Agilent Seahorse XF cell culture microplates under normoxic conditions for 48 h, after which the same number of hADSCs were added to each well and cultured with or without αGA, under normoxic conditions, for 24 h before analysis. The detailed culture conditions for each group are listed in Supplementary Table S5 online [24].
Quanti cation of cytokines from the hADSC graft hADSC grafts were cultured under normoxic or hypoxic (1 % O 2 ) conditions, and culture supernatant was collected after 24, 48, and 72 h. The secretion of hepatocyte growth factor (HGF), vascular endothelial growth factor (VEGF), IL-6, and IL-10, which has been reported to be associated with functional improvement, was measured by performing enzyme-linked immunosorbent assays (R&D Systems, Minneapolis, MN, USA).

Isolation and motion analysis of rCMs with/without hADSC supernatant
The 7-8-week-old male Crl:CD (SD) rats were anesthetized by inhalation of iso urane, and heart was harvested. The heart was connected to a Langendorff circulation circuit through aortic ligation to the cannula (Fig. 3G). The perfusion buffer (Tyrode's solution, Supplementary Table S6 online) containing 300 units/mL collagenase type II (Gibco; Thermo Fisher Scienti c) was perfused into the coronary arteries. The heart was homogenized on a petri dish, and cardiac cells were pipetted and transferred to a 96-well plate (Fig. 3H-I). The cardiomyocytes were stimulated with either 1 µM isoprenaline, 400 µM propranolol, or hADSC culture supernatant (Fig. 3J). Motion of rCMs was observed and analyzed using the SI8000 Cell Motion Imaging System (Sony, Tokyo, Japan) (Supplementary material online) [25][26][27].
Evaluation of in uence of mitochondrial transfer on cardiac function andmetabolism Graft survival following transplantation Luciferase-transduced hADSCs were transplanted into the nude rat myocardial infarction model in the same manner as described above (Fig. 4A). Luminescence intensity of the grafts was measured at the body surface. For histological analysis, hearts with residual graft samples were collected at days 1, 3, 7, and 14, and para n-embedded samples were stained using an antibody speci c to human mitochondria (Cat. #MAB1273; Merck Millipore, Burlington, MA, USA) (Supplementary material).
Histological analysis of the heart Heart samples were obtained from each group 8 weeks following surgery. To assess rCM diameter and brosis, heart sections were stained with periodic acid-Schiff and Picro Sirius red, respectively. The sections were also stained with anti-von Willebrand factor antibody, and capillary density was calculated in the peri-infarct area. For morphological evaluation of mitochondria in the heart, electron microscopy was performed as described above.
Quanti cation of intramyocardial ATP and mitochondrial DNA ATP content in the heart samples was assessed using a colorimetric ATP assay kit (ab83355; Abcam, Cambridge, UK). To quantify human and rat mitochondrial DNA in the recipient heart, we performed quantitative PCR. Total DNA was isolated from the infarct region and remote areas, and PCR was performed with SYBR Green (Thermo Fisher Scienti c) using speci c primer pairs (Supplementary Table  S1 online).

Statistical analysis
All values are expressed as means ± SEM. Statistical analyses were performed using JMP Pro 14 software (SAS Institute Inc., Cary, NC, USA). Comparison across multiple groups with normal distribution was performed using one-way analysis of variance (ANOVA), followed by Tukey's multiple comparisons test. Comparison between two groups was performed using an unpaired or paired two-tailed Student's ttest or Mann-Whitney U test, as appropriate. Comparison across multiple groups with non-normal distribution was performed using Kruskal-Wallis tests followed by Dunn's multiple comparisons tests. p 2. M. and Y. S. participated in research design and reviewed all data and the manuscript.
All authors have performed the following role: Drafting the work or revising it critically for important intellectual content; nal approval of the version to be published; agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Sources of Funding
This study was supported by ROHTO Pharmaceutical Co., Ltd.      Study protocol. Myocardial infarction was induced in 7-8-week-old male rats two weeks prior to transplantation. During transplantation, hADSCs with or without pre-stained mitochondria were implanted using brin glue on the surface of the heart. Following transplantation, heart samples were obtained at scheduled time points for various analyses and cardiac function was evaluated using echocardiography.
The hADSC grafts were analyzed in vitro.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.