JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY VOL. 77, NO.

8, 2021

ª 2021 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

PUBLISHED BY ELSEVIER

Mitochondria-Rich Extracellular Vesicles

From Autologous Stem Cell–Derived
Cardiomyocytes Restore Energetics of
Ischemic Myocardium
Gentaro Ikeda, MD, PHD,a Michelle R. Santoso, BS,a Yuko Tada, MD, PHD,a Albert M. Li, AB,b Evgeniya Vaskova, PHD,a
Ji-Hye Jung, PHD,a Connor O’Brien, MD,a Elizabeth Egan, MD, PHD,c Jiangbin Ye, PHD,b Phillip C. Yang, MDa

ABSTRACT

BACKGROUND Mitochondrial dysfunction results in an imbalance between energy supply and demand in a failing
heart. An innovative therapy that targets the intracellular bioenergetics directly through mitochondria transfer may be
necessary.

OBJECTIVES The purpose of this study was to establish a preclinical proof-of-concept that extracellular vesicle (EV)-
mediated transfer of autologous mitochondria and their related energy source enhance cardiac function through resto-
ration of myocardial bioenergetics.

METHODS Human-induced pluripotent stem cell–derived cardiomyocytes (iCMs) were employed. iCM-conditioned
medium was ultracentrifuged to collect mitochondria-rich EVs (M-EVs). Therapeutic effects of M-EVs were investigated
using in vivo murine myocardial infarction (MI) model.

RESULTS Electron microscopy revealed healthy-shaped mitochondria inside M-EVs. Confocal microscopy showed that
M-EV–derived mitochondria were transferred into the recipient iCMs and fused with their endogenous mitochondrial
networks. Treatment with 1.0
108/ml M-EVs significantly restored the intracellular adenosine triphosphate production
and improved contractile profiles of hypoxia-injured iCMs as early as 3 h after treatment. In contrast, isolated mito-
chondria that contained 300
more mitochondrial proteins than 1.0
108/ml M-EVs showed no effect after 24 h. M-EVs
contained mitochondrial biogenesis-related messenger ribonucleic acids, including proliferator-activated receptor g
coactivator-1a, which on transfer activated mitochondrial biogenesis in the recipient iCMs at 24 h after treatment. Finally,
intramyocardial injection of 1.0
108 M-EVs demonstrated significantly improved post-MI cardiac function through
restoration of bioenergetics and mitochondrial biogenesis.

CONCLUSIONS M-EVs facilitated immediate transfer of their mitochondrial and nonmitochondrial cargos, contributing
to improved intracellular energetics in vitro. Intramyocardial injection of M-EVs enhanced post-MI cardiac function
in vivo. This therapy can be developed as a novel, precision therapeutic for mitochondria-related diseases including heart
failure. (J Am Coll Cardiol 2021;77:1073–88) © 2021 by the American College of Cardiology Foundation.

Listen to this manuscript’s

audio summary by
Editor-in-Chief From the aStanford Cardiovascular Institute and Division of Cardiovascular Medicine, Department of Medicine, Stanford Uni-
Dr. Valentin Fuster on versity School of Medicine, Stanford, California, USA; bDepartment of Radiation Oncology, Stanford University School of Medicine,
JACC.org. Stanford, California, USA; and the cDivision of Infectious Diseases, Department of Pediatrics, Stanford University School of
Medicine, Stanford, California, USA.
The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’
institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information,
visit the Author Center.

Manuscript received September 18, 2020; revised manuscript received December 7, 2020, accepted December 23, 2020.

ISSN 0735-1097/$36.00 https://doi.org/10.1016/j.jacc.2020.12.060

1074 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

P harmacologic therapies have involving CD38 (23). Extracellular mitochondria were

ABBREVIATIONS
AND ACRONYMS improved survival in patients with internalized by injured neurons and contributed to
heart failure (HF) over the past 3 de- better neurologic outcomes in a mice stroke model.
ATP = adenosine triphosphate
cades (1). Despite these pioneering medical These data suggest that EVs facilitate a secure and
CMR = cardiac magnetic
therapies, the number of HF patients is highly efficient transfer of their mitochondrial cargo
resonance
increasing steadily and represents the top into the recipient cells.
COX = cytochrome c oxidase
diagnosis of hospital admissions today. This Induced pluripotent stem cells (iPSCs) have
D15s = induced pluripotent
stem cell–derived
fact highlights the need for the development tremendous therapeutic potential for the treatment
cardiomyocytes cultured for of innovative treatment strategies (1,2). HF of cardiovascular disease (24). The promise of preci-
shorter than 15 days represents bioenergetic imbalance. The sion medicine can be realized by using patient-
D50s = induced pluripotent disruption of the balance between energy specific iPSC derivatives such as induced pluripotent
stem cell–derived
supply and demand is important for the path- stem cell–derived cardiomyocytes (iCMs), which
cardiomyocytes cultured for
longer than 50 days ogenesis of HF (3). Cardiac tissue from pa- secrete EVs containing functional mitochondria. The
ERR = estrogen-related tients with dilated, hypertrophic, and goal of this study was to establish a preclinical proof-
receptor ischemic cardiomyopathies exhibits mito- of-concept that EV-mediated transfer of autologous
ETC = electron transport chain chondrial structural abnormalities and mitochondria restores myocardial bioenergetics.
EV = extracellular vesicle diminished adenosine triphosphate (ATP) Moreover, the role of the nonmitochondrial EV cargo
FCM = flow cytometry production despite increased metabolic en- in activating PGC-1a –mediated mitochondrial
HF = heart failure
ergy demands in the failing heart (4–7). biogenesis in the recipient iCMs was elucidated
Although proliferator-activated receptor g (Central Illustration).
iCM = induced pluripotent stem
cell–derived cardiomyocyte coactivator (PGC)-1 a serves as a master regu-
SEE PAGE 1089
iPSC = induced pluripotent lator of mitochondrial biogenesis and func-
stem cell tional capacity, PGC-1 a expression levels are
METHODS
Ld-EV = large vesicle-depleted decreased in the myocardium of HF patients
extracellular vesicle (8,9). Insufficient energy generation results
This study has been approved by Stanford Univer-
LV = left ventricular in cardiomyocyte contractile abnormality,
sity School of Medicine Institutional Review Board
M-EV = mitochondria-rich myocardial dysfunction, and, ultimately,
extracellular vesicle committee and Institutional Animal Care and Use
decompensated HF. The current standard
MI = myocardial infarction
Committee (Institutional Review Board approval
pharmacologic regimen, including beta-
number: 31517, Administrative Panel on Laboratory
mRNA = messenger ribonucleic blockers (10,11), ivabradine (12,13), and
acid Animal Care approval number: 33723, and Institu-
renin-angiotensin-aldosterone antagonism
mtDNA = mitochondrial tional Animal Care and Use Committee assurance
(14,15), attempt to correct this imbalance by
deoxyribonucleic acid number: A3213-01). All animals received humane
reducing cardiac workload, namely, energy
PBS = phosphate-buffered care and treatment in accordance with the Guide for
saline
demand. These therapeutics are essentially
the Care and Use of Laboratory Animals (25).
noncurative because they do not target the
PGC = proliferator-activated Detailed methods are described in the Supplemental
receptor g coactivator primary energy source of the failing heart.
Appendix.
RFP = red fluorescent protein Therefore, it is essential to develop a novel
therapy that targets the intracellular ener- PREPARATION OF HUMAN BLOOD MONONUCLEAR

getic supply directly (16). CELL-DERIVED iPSCs AND iCMs. The monoclonal

Cell-based therapy is reported to improve the car- iPSCs lines were generated by transfection of human
diac function in animal models of myocardial infarc- blood mononuclear cells with nonintegrating Sendai
tion (MI). Further analysis demonstrated that only a virus. iPSCs were chemically differentiated into iCMs.
small proportion of the transplanted cells engraft in The maturity of iCMs was evaluated by polymerase
the recipient heart (17). These data support a “para- chain reaction (PCR) and western blot analysis.
crine hypothesis” of stem cell therapy and suggest Primers and antibodies used can be found in
that the isolation of the secreted factors may have Supplemental Tables 1 and 2.
important applications for HF therapy. We and others ISOLATION OF M-EVs AND LD-EVs. The iCM-
have reported that extracellular vesicles (EVs) effi- conditioned medium was collected and
ciently transfer their cargo into the recipient cells, mitochondria-rich extracellular vesicles (M-EVs) were
facilitating various intercellular communications obtained based on a differential ultracentrifugation
(18,19). Recently, it has been reported that extracel- method. Floating cells, cell debris, and apoptotic
lular mitochondria exist inside the EVs (20–22). As- bodies were removed by centrifugation at 1,000g for
trocytes actively released EVs that contain functional 10 min at 4 C. Pelleted vesicles were obtained by ul-
mitochondria via calcium-dependent mechanisms tracentrifugation of the supernatant at 10,000g for
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1075
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

C ENTR AL I LL U STRA T I O N Extracellular Vesicle-Mediated Autologous Mitochondrial Transplantation

Ikeda, G. et al. J Am Coll Cardiol. 2021;77(8):1073–88.

We obtained human induced pluripotent stem cells (iPSCs)-derived cardiomyocytes (iCMs) and succeeded in collecting mitochondria-rich extracellular vesicles (M-EVs)
from iCM-conditioned medium. The autologous M-EVs can be applied to patients with mitochondria-related diseases including heart failure. RNA ¼ ribonucleic acid.

30 min at 4  C. Large vesicle-depleted extracellular beads (Supplemental Figures 1C and 1D). Flow
vesicles (Ld-EVs) were obtained by filtering of M-EVs cytometry (FCM) was conducted at the Stanford
through 220-nm polyvinylidene fluoride mem- Shared FACS Facility on a BD LSRII instrument (Bec-
brane filter. ton, Dickinson, and Company, Franklin Lakes,
NANOPARTICLE TRACKING ANALYSIS. Size and New Jersey).
quantity of EVs were evaluated by nanoparticle
IN VITRO HYPOXIA MODEL. iCMs were placed in a
tracking analysis with a NanoSight LM20 (NanoSight,
Hypoxia Inductor Chamber (StemCell Technologies,
Malvern, United Kingdom). Samples were loaded into
Vancouver, British Columbia, Canada). Oxygen con-
the sample chamber with sterile syringes and imaged
tent was reduced to 0.8% to 1.0% using N2/CO 2 gas
using a 640-nm laser.
mix, and monitored continuously. The chamber was

FCM-BASED EV ANALYSIS. A total of 1
10 8

EVs placed in a 37  C incubator for 24 h.

were washed and suspended in 200 m l of iCM culture ATP MEASUREMENT. Intracellular ATP was deter-
medium and stained with 100 nmol/l MitoTracker mined by CellTiter-Glo luminescence kit (G7570,
Deep Red (MTDR) (Thermo Fisher Scientific, Wal- Promega Corporation, Madison, Wisconsin). Opaque-
tham, Massachusetts) and 100 nmol/l MitoTracker walled 96-well plates with iCMs (10,000 cells/well)
Green (MTG) (Thermo Fisher Scientific) for 10 min at were prepared. CellTiter-Glo luminescence test solu-
room temperature and protected from light. We tion was added and incubated for 30 min at room
confirmed that iCM culture medium was EV-free temperature. Luminescent signal was determined by
(Supplemental Figure 1A). Less than 1% of EVs from a luminescence microplate reader.
mitochondria-free red blood cells were positive for MOUSE MI MODEL. Adult female CD1 mice (age: 80 to
mitochondrial markers (Supplemental Figure 1B). EV 100 days) were purchased from Charles River Labo-
diameter was approximated using the size standard ratories (Wilmington, Massachusetts). Experimental
1076 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

F I G U R E 1 Characterization of M-EVs

A B
Unstained D15 M-EVs D15 Ld-EVs
iCMs conditioned-medium
10
5
10
5
10
5
+Filter
10
4
10
4
10
4
MTG+ M-EVs
1,000 g for 10 min 3 3
10 10 10
3
500

Collect Supernatant 10
2
10
2
10
2 400

Events
300
10,000 g for 30 min
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

SSA
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
200
D50 M-EVs D50 Ld-EVs
Pellet 100
5
10 10
5
+Filter 0
2 3 4 5
4
10 10
4
0 10 10 10 10
Mitochondria-rich EVs
MTDR
3 3
10 10
(M-EVs)
2 2
10 10

1 2 3 4 5 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10

MTG

C D E
ETC Complexes ATP Level
ATP5A (V) (Fold)
1.5
UQCRC2 (III) *

1.0
SDHB(II)
0.5
COXII (IV)
NDUFB8 (I) 0
Vs

s
EV
Vs

-E
EV

-
-E

Ld
-
M

Ld

5
D1
5

D1
5
D1

D1

(A) Differential ultracentrifugation was employed to isolate mitochondria-rich extracellular vesicles (M-EVs). (B) Representative flow cytometry dot plots and his-
tograms of M-EVs and large vesicle-depleted extracellular vesicles (Ld-EVs). (C) Transmission electron microscopy of induced pluripotent stem cell–derived car-
diomyocytes (iCMs) cultured for shorter than 15 days (D15) M-EVs. (D) The protein expression of the electron transport chain (ETC) in D15 M-EVs. (E) M-EVs, but not Ld-
EVs, produced extracellular adenosine triphosphate (ATP) (n ¼ 4 per group). *p < 0.0001 by an unpaired t-test. ATP5A ¼ ATP synthase a-subunit; COX
II ¼ cytochrome c oxidase subunit 2; D50s ¼ induced pluripotent stem cell–derived cardiomyocytes cultured for longer than 50 days; MTDR ¼ MitoTracker Deep Red;
MTG ¼ MitoTracker Green; NDUFB8 ¼ nicotinamide adenine dinucleotide dehydrogenase-1b subcomplex subunit 8; SDHB ¼ succinate dehydrogenase complex iron
sulfur subunit B; SSA ¼ side-scatter area; UQCRC2 ¼ ubiquinol-cytochrome c reductase core protein 2.

groups were randomized across multiple cages, lit- PBS), or 30 m l PBS were injected at 3 sites around
ters, and location of mouse cages in the husbandry infarct regions. The chest was closed, and animals
room. Mice were endotracheally intubated. Anes- were weaned from the ventilator and extubated.
thesia was maintained with inhalational 1.5% to 2.0% STATISTICAL ANALYSIS. Data are expressed as mean
isoflurane. A thoracotomy was performed, and an  SD. We analyzed the differences between the 2
8-0 silk ligature was placed around the left anterior groups using unpaired Student’s t-tests; the differ-
descending artery 1 mm below the atrioventricular ences among 3 groups or more were assessed using
border. M-EVs (1
108 /30 m l phosphate-buffered sa- analysis of variance and post hoc Tukey or Bonferroni
line [PBS] per group), Ld-EVs (1
108/30 m l PBS per multiple comparison tests with Prism software
group), isolated mitochondria (1.0 m g protein/30 ml version 4.0 (GraphPad Software, San Diego,
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1077
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

California). We considered p values <0.05 statisti- contained 4.7
more mitochondria than <240-nm
cally significant. EVs did (Supplemental Figure 3C). We next assessed
the in vitro effects of the conditioned medium on the
RESULTS mitochondrial function. iCMs were treated with the
conditioned medium for 24 h and the 3-(4,5-
IMMATURE iCMs SHOW ROBUST EXPRESSIONS OF dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bro-
TRANSCRIPTIONAL COACTIVATORS THAT REGULATE mide (MTT) reduction was measured as an indicator
MITOCHONDRIAL BIOGENESIS. iCMs mature in long- of mitochondrial function. The conditioned medium
term culture (26,27). Based on these findings, we from D15s or D50s did not change MTT reduction in
characterized iPSC–derived cardiomyocytes cultured the iCMs without hypoxic injury (Supplemental
for shorter than 15 days (D15s) versus those for longer Figure 3D). Then iCMs were subjected to hypoxia for
than 50 days (D50s) after the observation of initial 24 h and treated with the conditioned medium at the
spontaneous contractions. Consistent with previous time of reoxygenation. Treatment with the D15-
reports, the D15s group, when compared with D50s, conditioned medium significantly increased MTT
showed immature phenotypes in terms of sarcomeric reduction, whereas D50-conditioned medium showed
structure, electrical conduction, and Ca 2þ handling fewer effects. The 220-nm filtration of the condi-
(Supplemental Figures 2A and 2B). D50s showed tioned medium further compromised the effects,
higher expression levels of electron transport suggesting that the mitochondria-containing EVs
chain (ETC) proteins/transcripts and mitochondrial (>220 nm) enhanced mitochondrial functions in the
deoxyribonucleic acid (mtDNA) copy numbers hypoxia-injured iCMs.
(Supplemental Figures 2C to 2E). In contrast,
CHARACTERISTICS OF M-EVs. To augment the
messenger ribonucleic acid (mRNA) levels of mito-
therapeutic effects, we obtained M-EVs by differ-
chondrial dynamics-related genes were decreased in
ential ultracentrifugation of the conditioned me-
D50 (Supplemental Figure 2D). Two types of
dium according to previously published
mitochondria-specific labeling (28) that distinguish
microvesicle purification (31) and mitochondria
functional mitochondria (MTDR) and total mitochon-
isolation methods (Figure 1A). Nanoparticle tracking
dria (MTG) revealed that D50s showed more functional
analysis revealed that the size of M-EVs from both
mitochondria (Supplemental Figure 2F). Intracellular
D15s and D50s ranged from 100 to 600 nm and the
ATP levels were higher in D50s (Supplemental
mean diameters were 295 and 237 nm, respectively
Figure 2G). As D15s continue to differentiate into
(Supplemental Figure 4A). D15s produced higher
mature cardiomyocytes, an increase in mitochondria is
numbers of M-EVs that contained more mitochon-
reported to be orchestrated by PGC-1a –mediated
dria than D50s did (Figure 1B, Supplemental
mitochondrial biogenesis (29,30). Consistent with this,
Figure 4B). We found that 87% of mitochondria
the less mature D15s showed higher activity levels of
inside D15 M-EVs retained their membrane poten-
mitochondrial biogenesis through increased expres-
tials. Double staining with wheat germ agglutinin, a
sion levels of PGC-1a and estrogen-related receptor g
cell membrane dye, and MTG revealed that the
(ERR g ) (Supplemental Figures 2C and 2H).
majority (81%) of wheat germ agglutinin–positive
IMMATURE iCMs SECRETE MITOCHONDRIA-CONTAINING EVs contained mitochondria (Supplemental
EVs THAT IMPROVE MITOCHONDRIAL FUNCTION OF Figure 4C). Filtration of M-EVs through 220-nm fil-
HYPOXIA-INJURED CARDIOMYOCYTES. FCM revealed ter resulted in Ld-EVs that did not contain mito-
that D15-conditioned medium contained more extra- chondria (Figure1B, Supplemental Figure 4D).
cellular mitochondria than D50-conditioned medium Nanoparticle tracking analysis confirmed that the
did (Supplemental Figure 3A). Extracellular mito- filtration technique depleted the EVs >220 nm
chondria from D15s showed greater proportions of (Supplemental Figure 4E). Respectively, 40% and
MTDRþ-functional mitochondria. Filtration of the 43% of D15 M-EVs were positive for b1-integrin and
conditioned medium with a 220-nm filter reduced the CD63 (Supplemental Figure 4F). Transmission elec-
quantity of extracellular mitochondria (Supplemental tron microscopy of M-EVs revealed round-shaped
Figures 3A and 3B), suggesting that extracellular EVs with diameters ranging from 98 to 677 nm
mitochondria were incorporated in >220-nm EVs. (mean: 204 nm) (Figure 1C). Extracellular mito-
FCM side-scatter area revealed that >240-nm EVs chondria with preserved structural integrity were
1078 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

F I G U R E 2 M-EVs Facilitate Transfer of Their Mitochondrial Cargo and Restore Intracellular Bioenergetics in Hypoxia-Injured iCMs

A 3-Hours After Treatment B 24-Hours After Treatment

C D15 M-EVs D D50 M-EVs

3-Hours 24-Hours 3-Hours 24-Hours
4 4
** *
* *
Intracellular ATP (Ratio)

Intracellular ATP (Ratio)

* *
3 3

2 2

1 1

0 0
Hypoxia – + + + + + – + + + + + Hypoxia – + + + + + – + + + + +
M-EVs – – M-EVs – –

E D15 Ld-EV F D50 Ld-EV

3-Hours 24-Hours 3-Hours 24-Hours
4 4
*
Intracellular ATP (Ratio)

Intracellular ATP (Ratio)

3 3

2 2

1 1

0 0
Hypoxia – + + + + + – + + + + + Hypoxia – + + + + + – + + + + +
Ld-EVs – – Ld-EVs – –

Continued on the next page

JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1079
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

observed inside the EVs. ETC proteins were detec- Oligo-EVs were similar to M-EVs (Supplemental
ted in M-EVs, but not in Ld-EVs (Figure 1D). M-EVs Figure 6A). Mitochondria inside Oligo-EVs were
contained more mtDNA and ATP than Ld-EVs did dysfunctional (Supplemental Figures 6B and 6C).
(Figure 1E, Supplemental Figure 4G). However, 3.0
108/ml Oligo-EVs significantly
improved ATP levels at 24 h after treatment
M-EVs TRANSFER THEIR MITOCHONDRIAL CARGO INTO
(Supplemental Figure 6D). These findings suggest
HYPOXIA-INJURED iCMs AND RESTORE INTRACELLULAR
that D15 M-EVs contain nonmitochondrial cargos that
BIOENERGETICS. We next observed the transfer of M-
enhance bioenergetics in the hypoxia-injured iCMs.
EV mitochondria into the recipient cells. We labeled
M-EV mitochondria with red fluorescent protein NONMITOCHONDRIAL CARGOS IN D15 M-EVs

(RFP) and native mitochondria in the recipient cells ENHANCE MITOCHONDRIAL BIOGENESIS AND

with green fluorescent protein using BacMam tech- BIOENERGETICS IN THE RECIPIENT iCMs. The RNA/

nology. FCM revealed that M-EVs containing RFPþ- protein profiles of EVs are reported to mirror those of
mitochondria were secreted from D15s transduced their parent cells (34). D15s showed ongoing mito-
with BacMam mitochondria-RFP (Supplemental chondrial development and higher expressions of
Figure 5). iCMs transduced with BacMam mRNA that regulate mitochondrial biogenesis. We
mitochondria–green fluorescent protein were sub- therefore hypothesized that D15 M-EVs contain more
jected to hypoxia and treated with 3.0
10 /ml RFP - 8 þ mRNAs that facilitate mitochondrial biogenesis. D15 Ld-
M-EVs. Confocal microscopy revealed that M-EV EVs showed higher expression levels of PGC-1a and
mitochondria (red) were transferred into the recipient ERRg mRNA than D50 Ld-EVs did (Figure 3A). These
iCMs and some of them were fused with the host mRNAs were uniformly distributed in the D15 M-EVs,
mitochondrial networks (green) at 3 and 24 h after D15 Ld-EVs, and MitoTracker-positive EVs that were
treatment (Figures 2A and 2B). purified by fluorescence-activated cell sorter
We next examined the effects of M-EVs on intra- (Supplemental Figures 7A and 7B). Treatment with D15
cellular bioenergetics. When iCMs were subjected to M-EVs, but not D50 M-EVs, induced up-regulation of
hypoxia, ATP levels fell as expected (Figure 2C). mitochondrial biogenesis-related mRNAs in the
Treatment with D15 M-EVs at the time of reoxygena- hypoxia-injured iCMs at 24 h after treatment
tion increased ATP levels in a dose-dependent (Figure 3B). The protein expressions of PGC-1a and ETC
8
manner and 1.0
10 /ml or 3.0
10 /ml M-EVs 8 complexes were up-regulated in the iCMs treated with
showed significant effects at 3 and 24 h after treat- D15 M-EVs, whereas D50 M-EVs showed no effect
ment (Figure 2C). D50 M-EVs also showed significant (Supplemental Figure 7C). D15 Ld-EV treatment showed
effects, albeit to a lesser extent (Figure 2D). We and a trend toward an increase in PGC-1a and ETC protein
others (19,32) reported that EVs from cardiomyocytes expressions. We measured the mitochondrial respira-
deliver their RNA or protein cargos into the recipient tory function with a mitochondrial stress test using the
cells, providing various therapeutic effects. There- Seahorse XF96 extracellular-flux analyzer (Agilent,
fore, we next examined whether the non- Santa Clara, California). D15 M-EV treatment signifi-
mitochondrial cargo in M-EVs had significant effects cantly increased basal, ATP-linked, and maximal respi-
on the intracellular bioenergetics. Although Ld-EVs ration in hypoxia-injured iCMs (Figure 3C). D15 Ld-EVs
showed no significant effect at 3 h after treatment, also restored basal and maximal respiration. These
3.0
108 /ml D15 Ld-EVs significantly improved data suggested that nonmitochondrial cargo in D15 M-
intracellular ATP levels at 24 h after treatment EVs enhance the intracellular bioenergetics through
(Figure 2E). In contrast, D50 Ld-EVs showed no effects activation of mitochondrial biogenesis.
within 24 h (Figure 2F). To confirm the effects of PGC-1a KNOCKDOWN COMPROMISES THE RESTORATION
nonmitochondrial cargo, D15s were treated with oli- OF BIOENERGETICS AND MITOCHONDRIAL BIOGENESIS
gomycin A (33), a specific ATP synthase inhibitor, and IN THE HYPOXIA-INJURED iCMs. We evaluated the role
collected M-EVs (Oligo-EVs). The size distribution of of PGC-1a in the effects of M-EV therapy. D15s were

F I G U R E 2 Continued

(A,B) Representative microscopic image of iCMs treated with M-EVs containing RFPþ mitochondria. Transferred M-EV–mitochondria (red) were observed in the
recipient iCMs harboring GFPþ mitochondria (green). Some of the transferred mitochondria were fused with the native mitochondrial networks (white arrows).
Bars ¼ 50 and 20 mm. (C to F) ATP levels in iCMs treated with M-EVs or Ld-EVs (wedge: 0; 1.0
107; 3.0
107; 1.0
108; or 3.0
108; all per ml). (Left) 3 h after
treatment. (Right) 24 h after treatment (n ¼ 4 to 6 per group). *p < 0.05 and **p < 0.001 by a 1-way analysis of variance followed by Tukey test. GFP ¼ green
fluorescent protein; RFP ¼ red fluorescent protein; other abbreviations as in Figure 1.
1080 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

F I G U R E 3 Nonmitochondrial Cargo in D15 M-EVs Restored Bioenergetics Through Activation of PGC-1 a –Mediated Mitochondrial Biogenesis in the Recipient iCMs

A mRNA Levels in B
Ld-EVs (Fold) mRNA Levels in iCMs
10

NDUFA1

COX6A1
ATP5A1
PGC-1α

SDHA
TFAM
NRF1
ERRγ

CYC1

PKM
CS
* *
Control
1 D15 M-EV
D50 M-EV

Min Max
0.1
PGC-1α ERRγ
D15 Ld-EVs D50 Ld-EVs

C Antimycin A
Oligomycin A FCCP + rotenone
200 200
OCR (pmol min-1)

OCR (pmol min-1)

150 150
**
*** *
*
100 100 ***
**
50 50

0 0
0 20 40 60 80 100
AT tio al

t d

Le on)

Re tio al

ac ry

tio al
ira ke
ra as

ar ira im

ira ri
ap ato
Re P- n

H + ion

ak

sp n

ito ty

n
Time (Min)
sp n d
t
sp Lin
B

ro

Sp esp ax

i
ir

Re ho
(P

R M

Normoxia Hypoxia + D15 M-EVs

pi

C
s
Re

Hypoxia + PBS Hypoxia + D15 Ld-EVs

n-
No

Normoxia Hypoxia + D15 M-EVs

Hypoxia + PBS Hypoxia + D15 Ld-EVs

(A) Peroxisome proliferator-activated receptor g coactivator 1a (PGC-1a) and estrogen-related receptor g (ERRg) messenger ribonucleic acid (mRNA), normalized to
small nuclear RNA U6, were up-regulated in D15 Ld-EVs (n ¼ 4 per group). *p < 0.05 by an unpaired t-test. (B) Quantitative gene expression in iCMs, normalized to
actin b, ACTB, shown as fold change relative to normoxic iCMs (n ¼ 4 to 5 per group). (C) Seahorse extracellular-flux assays measuring oxygen-consumption rate (OCR)
(n ¼ 5 to 6 per group). *p < 0.01, **p < 0.001, and ***p < 0.0001 by a 1-way analysis of variance followed by Tukey test. ATP5A1 ¼ ATP synthase F1 a-subunit;
COX6A1 ¼ cytochrome c oxidase subunit 6A1; CS ¼ citrate synthase; CYC1 ¼ cytochrome c1; FCCP ¼ carbonyl cyanide-4-(trifluoromethoxy)-phenylhydrazone;
Max ¼ maximum; Min ¼ minimum; NRF1 ¼ nuclear respiratory factor 1; NDUFA1 ¼ nicotinamide adenine dinucleotide dehydrogenase-1a subcomplex subunit 1;
PBS ¼ phosphate-buffered saline; PKM ¼ pyruvate kinase M1/2; SDHA ¼ succinate dehydrogenase complex flavoprotein subunit A; TFAM ¼ mitochondrial transcription
factor A; other abbreviations as in Figure 1.
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1081
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

F I G U R E 4 PGC-1 a Knockdown Compromised the Restoration of Mitochondrial Bioenergetics and Biogenesis in the Hypoxia-Injured iCMs

A mRNA Levels C
in Ld-EVs PGC-1α / ETC Complexes
10
PGC-1α
ATP5A (V)
* *
1 UQCRC2 (III)
SDHB (II)
COXII (IV)
0.1 NDUFB8 (I)
β-actin

0.01 Hypoxia – – + + + + + + + +
PGC-1α ERRγ Ld-EV (shCTRL) – – – – + + + – – –
shCTRL shPGC-1α Ld-EV (shPGC-1α) – – – – – – – + + +

B Intracellular ATP 1.5

* * **** **
(Ratio) * * **** ***
2.5
** * 1.0
2

1.5 0.5

1
0
0.5
(I)

I)

)
1α

(II

IV

(V
(II
C-

B8

I(
B

5A
C
PG

XI
CR
UF

P
SD

CO

0
AT
UQ
ND

3-Hours 24-Hours
Normoxia Hypoxia + PBS Normoxia Hypoxia + PBS
Hypoxia + Ld-EVs (shCTRL) Hypoxia + Ld-EVs (shCTRL)
Hypoxia + Ld-EVs (shPGC-1α) Hypoxia + Ld-EVs (shPGC-1α)

(A) Relative PGC-1a and ERRg mRNA expression, normalized to U6, in Ld-EVs from D15s transduced with pLKO.1 vector expressing control short, hairpin ribonucleic acid
(shCTRL) or anti-peroxisome proliferator-activated receptor g coactivator 1a short, hairpin ribonucleic acid (shPGC-1a) (n ¼ 3 per group). *p < 0.0001 by an unpaired
t-test. (B) PGC-1a knockdown reduced restoration of ATP levels in hypoxia-injured iCMs at 24 h (n ¼ 5 to 8 per group). *p < 0.05 and **p < 0.01 by a 1-way analysis of
variance followed by Tukey test. (C) Representative immunoblots and quantitative analyses of PGC-1a, ETC complexes, and b-actin (n ¼ 4 to 6 per group). *p < 0.05,
**p < 0.01, ***p < 0.001, and ****p < 0.0001 by a 1-way analysis of variance followed by Tukey test. Abbreviations as in Figures 1 and 3.

transduced with lentivirus harboring anti-PGC-1 a on the restoration of ATP levels (Figure 4B) and PGC-
short hairpin RNA. We confirmed that anti-PGC-1 a 1a /ETC protein levels in hypoxia-injured iCMs at 24 h
short hairpin RNA, compared with lentivirus after treatment (Figure 4C). These data suggest the
harboring nontargeted short hairpin RNA, decreased significant role of PGC-1a –related mRNA and protein
the expression levels of PGC-1 a mRNA and protein in the M-EV therapy.
(Supplemental Figures 8A and 8B). Ld-EVs from D15s
transduced with anti-PGC-1 a short hairpin RNA ISOLATED MITOCHONDRIA DID NOT RESTORE
showed marked reduction in PGC-1 a and ERRg mRNA BIOENERGETICS IN HYPOXIA-INJURED iCMs WITHIN
levels (Figure 4A). Importantly, PGC-1a knockdown 24 H AFTER TREATMENTS. We next obtained iso-
significantly reduced the effects of Ld-EVs treatment lated mitochondria from homogenized D15s (Isolated-
1082 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

F I G U R E 5 D15 M-EVs Improved Mitochondrial Function, Contractile Property, and Cell Survival in Hypoxia-Injured iCMs

Functional
A Mitochondria (%)

****
Normoxia Hypoxia *
PBS PBS D15 M-EVs Isolated-Mito ****
100 * *
5 5 5 5
10 10 10 10 Normoxia
MTDR

10
4
10
4
10
4
10
4 80
3 3 3 3
Hypoxia + PBS
10 10 10 10
60 Hypoxia + D15 M-EVs
2 2 2 2
10 10 10 10
2 3
10 10 10 10
4 5 2 3
10 10 10 10
4 5 2 3
10 10 10 10
4 5 2 3
10 10 10 10
4 5
40 Hypoxia + Isolated-Mito

MTG
20

B Contractile Profiles

Max Contractile Max Relaxation

D Peak
Rate Rate
0.020 0.20 0
*** ** *** ***
0.015 0.15 Normoxia

–0.05 Hypoxia + PBS

0.010 0.10
Hypoxia + D15 M-EVs

0.005 0.05 Hypoxia + Isolated-Mito

–0.10
0

C Annexin / PI (%)

Normoxia Hypoxia ***

***
PBS PBS D15 M-EVs Isolated-Mito
80 ** * ****
**
3 3 3 3 ****
Propidium Iodide

10 10 10 10
60
** *
2 2 2 2
10 10 10 10
40
0 0 0 0
2 2 2 2 *
–10 –10 –10 –10 **
3 4 5 3 4 5 3 4 5 3 4 5
20
0 10 10 10 0 10 10 10 0 10 10 10 0 10 10 10

Annexin V 0
Viable Early phase Late phase
Apoptosis
Normoxia Hypoxia + PBS
Hypoxia + D15 M-EVs Hypoxia + Isolated-Mito

(A) Representative flow cytometry dot plots of iCMs labeled with MTG/MTDR (n ¼ 5 to 6 per group). (B) D peak and maximum contractile rate in iCMs
(n ¼ 5 to 6 per group). (C) Representative flow cytometry dot plots of iCMs stained with annexin V/propidium iodide (PI) (n ¼ 4 to 5 per group).
*p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a 1-way analysis of variance followed by Tukey test. Isolated-Mitos ¼ isolated mitochondria
from homogenized induced pluripotent stem cell–derived cardiomyocytes cultured for fewer than 15 days; other abbreviations as in Figures 1 and 3.

Mitos) and confirmed that the Isolated-Mitos that contain 100
more mitochondrial proteins than
(3.0 m g/ml protein) harbored the same amounts of 3.0
10 8/ml M-EVs had no effect on the recovery of
mitochondrial proteins (voltage-dependent anion ATP levels within 24 h (Supplemental Figure 9C).
channel, VDAC, and cytochrome c oxidase subunit 4, These data suggested that the EVs facilitated the
COX IV) in approximately 3.0
108 /ml D15 M-EVs transfer of their mitochondrial cargo into the recip-
(Supplemental Figure 9A). Hypoxia-injured iCMs ient iCMs.
were treated with RFP-labeled Isolated-Mitos MITOCHONDRIA INSIDE EVs ARE MORE RESISTANT
(3.0 m g/ml protein) and observed by confocal micro- TO CA 2 D OVERLOAD AND OXIDATIVE STRESS THAN
scopy. Isolated-Mitos (red) were rarely detected in- ISOLATED MITOCHONDRIA ARE. Extracellular envi-
side the recipient iCMs within 24 h after treatment ronment with high Ca2þ concentration (w2.0 mmol/l)
(Supplemental Figure 9B). Isolated-Mitos (300 m g/ml) and oxidative stress may compromise the viability of
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1083
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

F I G U R E 6 The Effects of Intramyocardial Injection of D15 M-EVs on Post-MI Cardiac Remodeling

A B LVEF (%)
D15 D15 Isolated 60
** *
Sham PBS M-EVs Ld-EVs -Mito **** ****
**** **** Sham
Diastole

40 ** *
* PBS
D15 M-EV
D15 Ld-EV
20
Systole

Isolated-Mito

0
Week-2 Week-4

C LVEDV (μL) D LVESV (μL) E LV Mass (mg)

300 * 300 150
* *** ***
**** **** ** ** *
** ** **
**** **** **
Sham
200 200 *** **** 100
PBS
D15 M-EV
D15 Ld-EV
100 100 50
Isolated-Mito

0 0 0
Week-2 Week-4 Week-2 Week-4 Week-2 Week-4

F
Viability (%)
Manganese-Enhanced MRI (MEMRI)
** **
Sham PBS
100 **** ****
*** ****
*
80 Sham
PBS
60 D15 M-EV
D15 Ld-EV
M-EVs Ld-EVs Isolated-Mito 40
Isolated-Mito
20

0
Week-2 Week-4

(A) Representative images of short-axis acquisitions at mid-left ventricle (LV). (B to E) Left ventricular ejection fraction (LVEF), left ventricular end-diastolic volume (LVEDV), left
ventricular end-systolic volume (LVESV), and LV mass at weeks 2 and 4 after myocardial infarction (MI), evaluated by cardiac magnetic resonance. (F) Representative images and
quantitative analyses of myocardial viability as visualized by manganese-enhanced cardiac magnetic resonance imaging (MEMRI). Red arrowheads indicate nonviable regions
(n ¼ 8 per group). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by a 2-way analysis of variance followed by Tukey test. Abbreviations as in Figures 1, 3, and 5.

exogenous mitochondria injected into the injured reduced MTDR signals in Isolated-Mitos
heart tissues. Therefore, we investigated whether (Supplemental Figure 10A). EV mitochondria
mitochondria in M-EVs versus in Isolated-Mitos retained MTDR signals up to concentration of
maintain function after incubation in buffers with 100 mmol/l H2 O 2 within 24 h, whereas MTDR signals
various concentrations of Ca 2þ or H2O 2 . After 3- or 24- were decreased in Isolated-Mitos incubated with
h of incubation, D15 M-EVs or Isolated-Mitos were buffers containing 100 mmol/l H 2O 2 at 3 and 24 h after
labeled with MTDR and analyzed by FCM. M-EVs treatment (Supplemental Figure 10B). These data
retained functional mitochondria up to concentration suggested that mitochondria in M-EV were more
of 2 mmol/l Ca2þ within 24 h, whereas 24-h of incu- resistant to Ca 2þ overload and oxidative stress than
2þ
bation with Ca at 0.2 and 2 mmol/l significantly mitochondria in Isolated-Mitos were.
1084 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

F I G U R E 7 D15 M-EV Therapy Enhances Bioenergetics and PGC-1 a –Mediated Mitochondrial Biogenesis in the Peri-infarct Region

A B
Functional Mitochondria Troponin | Hoechst ATP Levels (Fold)
1.5

*
1.0 * * PBS
D15 M-EVs
Isolated-Mito
0.5

YZ 0.0
Sham Day-3 Day-28
XZ MI

C
1.5
*
mRNA Levels (Fold)

* * * *
1.0 * Sham + PBS
MI + PBS
MI + D15 M-EVs
0.5 MI + Isolated-Mito

0.0
PGC-1α ERRγ NRF-1 TFAM NDUFAB1 SDHB CYTC COX4I1 ATP5ME
Biogenesis ETC

D Sham MI 1.5

PBS PBS D15 M-EVs Isolated-Mito * * * *

PGC-1α 1.0 Sham + PBS

MI + PBS
VDAC
MI + D15 M-EVs
COX IV 0.5 MI + Isolated-Mito

GAPDH

0
PGC-1α VDAC COXIV

(A) MTDRþ-mitochondria were detected inside the troponin Iþ-cardiomyocytes, highlighted with white arrows. Bar ¼ 50 mm. (B) Tissue ATP levels, normalized by tissue
protein levels (n ¼ 4 to 6 per group). (C) Mitochondrial biogenesis-related gene expression levels in the peri-infarct region, normalized to glyceraldehyde-3-phosphate
dehydrogenase (GAPDH), measured by reverse transcription polymerase chain reaction (n ¼ 4 to 5 per group). (D) Representative immunoblots and quantitative
analyses of PGC-1a, voltage-dependent anion channels (VDAC), and cytochrome c oxidase subunit 4 (COX IV), normalized to GAPDH (n ¼ 4 to 5 per group). *p < 0.05
by a 1-way analysis of variance followed by Tukey test. ATP5ME ¼ ATP synthase subunit-e; COX4I1 ¼ cytochrome c oxidase subunit 4 isoform 1; CYTC ¼ cytochrome c;
NDUFAB1 ¼ nicotinamide adenine dinucleotide ubiquinone oxidoreductase subunit AB1; other abbreviations as in Figures 1, 3, and 5.

M-EVs IMPROVE MITOCHONDRIAL FUNCTION, survival. iCMs were exposed to hypoxia for 24 h and
CONTRACTILE PROFILE, AND CELL SURVIVAL OF treated with 3.0
10 8/ml D15 M-EVs at the time of
HYPOXIA-INJURED iCMs. Appropriate level of intra- reoxygenation. Double-labeling of injured iCMs with
cellular bioenergetics is essential for mitochondrial MTG and MTDR revealed that M-EV treatment
quality, contractile activity, and, ultimately, cell significantly increased the proportion of functional
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1085
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

mitochondria (MTGþMTDR þ) at 24 h after treatment higher human mtDNA when compared with Isolated-
(Figure 5A). We found that 3.0
108 /ml M-EVs Mito group at 3 days after MI (Supplemental
significantly improved the contractile profiles, Figure 12A). However, human mtDNA was not
including D peak and maximum contractile rate at detected in either group at day 28. Intramyocardial
24 h after treatment (Figure 5B). M-EV treatment injections of D15 M-EVs, but not Isolated-Mitos,
increased viable cells by reducing early phase significantly increased tissue ATP levels at days 3
apoptotic cell death at 24 h after treatment and 28 (Figure 7B). M-EV treatment induced up-
(Figure 5C). In contrast, Isolated-Mitos that contained regulation of mitochondrial biogenesis-related
the same amount of mitochondrial proteins as 3.0 mRNAs, including PGC-1a , ERR g , succinate dehy-
10 8/ml M-EVs showed no effect on the mitochondrial drogenase complex iron sulfur subunit B, SDHB,
and cellular phenotypes. cytochrome c, CYTC, and cytochrome c oxidase
M-EV THERAPY PREVENTS POST-MI CARDIAC subunit 4 isoform 1, COX4I1, at day 3 (Figure 7C). The
REMODELING. In an in vivo mouse model of MI, protein expressions of PGC-1 a and COX IV were up-
intramyocardial injection of D15 M-EVs (1.0
108 ) regulated at day 28 (Figure 7D). Finally, D15 M-EV
into the peri-infarct region significantly improved left injection increased mtDNA copy numbers at day 28
ventricular (LV) ejection fraction as compared with (Supplemental Figure 12B). These data suggested
control group (Figures 6A and 6B, Supplemental that M-EV therapy restored bioenergetics and facili-
Table 3). Furthermore, cardiac magnetic resonance tated mitochondrial biogenesis in the peri-
(CMR) imaging measurements, including LV end- infarct region.
diastolic volume, LV end-systolic volume, and LV
DISCUSSION
mass demonstrated significantly reduced LV remod-
eling in D15 M-EV–treated animals (Figures 6C to 6E).
The novel findings of the present study are the
Manganese-enhanced CMR imaging detected signifi-
following: 1) D15 M-EVs contained more functional
cantly greater viable myocardium in D15 M-EVs
mitochondria than D50 M-EVs did; 2) EV-mediated
(Figure 6F). D15 Ld-EVs improved LV size and LV
transfer of mitochondrial and nonmitochondrial
ejection fraction significantly less when compared
cargo restore intracellular bioenergetics and contrac-
with D15 M-EVs. D50 M-EVs or D50 Ld-EVs prevented
tile property in hypoxia-injured iCMs; 3) therapeutic
post-MI LV dilatation, whereas no effect was
benefit from nonmitochondrial cargos in D15 M-EVs
observed in LV ejection fraction (Supplemental
contents depends on PGC-1a –related mRNA/proteins;
Figures 11A to 11E). Isolated-Mitos that contained
and 4) intramyocardial injection of 1.0
10 8 D15 M-
the same amount of mitochondrial proteins as D15 M-
EVs improves post-MI cardiac function in mice.
EVs (1.0
108 ) showed no effect. At week 4, we
The effectiveness of iCMs transplantation has been
measured lung weight and LV weight, showing that
reported in preclinical MI models (35). Our data sug-
the treatment with D15 M-EVs had trends toward
gest that immature iCMs can be a promising candi-
reduced lung and LV weight when indexed by tibia
date for cell therapy because they may supply
length (Supplemental Figures 11F and 11G). These
mitochondria and/or activate mitochondrial biogen-
data suggested the superiority of D15 M-EV therapy in
esis and enhance dynamics. However, immature iCMs
preventing post-MI LV remodeling.
exhibit less developed conduction and Ca2þ handling
M-EV THERAPY RESTORES BIOENERGETICS AND (Supplemental Figure 2), which may cause
ACTIVATES PGC-1 a –MEDIATED MITOCHONDRIAL automaticity-induced arrhythmias in the trans-
BIOGENESIS IN THE FAILING HEART. We next planted myocardium, as previously reported (35). To
examined the mechanisms by which D15 M-EVs address this, we explored the therapeutic effects of
prevented progression of post-MI HF. MTDR-labeled paracrine factors. Although the arrhythmogenicity of
M-EVs (1.0
10 8) were injected into myocardium in M-EV injection needs to be examined using large
the peri-infarct region. Immunohistochemistry at 8 h animal models, this cell-free approach is expected to
after injection demonstrated that MTDR þ-functional confer cardioprotection without inducing ventricular
mitochondria were detected inside the troponin I þ- arrhythmia.
cardiomyocytes in the peri-infarct region (Figure 7A). Intramyocardial injection of isolated mitochondria
To trace the transferred mitochondria, we measured reduced infarct size in animal models of ischemia/
human mtDNA in the peri-infarct myocardium. D15 reperfusion injury (36). However, clinical application
M-EVs injected in the myocardium showed 62
of this technique is extremely challenging (37) due to
1086 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

the following constraints of isolated mitochondria: 1) the cardiomyocytes harbor the most active mito-
effective function in the extracellular environment; chondria and produce the most effective M-EVs.
2) transmembrane transfer into the recipient car- However, we cannot exclude other cell types, which
diomyocytes; and 3) sufficient and sustainable ATP could yield equally effective M-EVs. Second, we used
generation for contractile performance in the car- differential ultracentrifugation to isolate the M-EVs.
diomyocytes. M-EV therapy addresses these chal- However, it is unknown what other methods may
lenges. M-EV mitochondria were more resistant to isolate the M-EVs more efficiently. Future studies are
high extracellular Ca2þ concentrations than isolated warranted to clarify the most effective EV cell source
mitochondria were (Supplemental Figure 10), sug- and optimal purification methods. Finally, we have
gesting that the lipid bilayer of the EVs prevented only investigated mitochondrial biogenesis-related
Ca 2þ diffusion to damage the mitochondria and the mRNA profiles as nonmitochondrial therapeutic fac-
related cargo. EVs facilitated quick transfer of their tors. It is, however, reported that transfer of protein
mitochondrial cargo into the recipient iCMs in clusters or noncoding RNAs inside the EVs also initi-
contrast to the Isolated-Mitos that are rarely inter- ates various intracellular signaling. Comprehensive
nalized by the iCMs (Figure 2, Supplemental Figure 9). analysis of EV-related proteins/RNAs is required to
These findings suggested that the lipid bilayer of EVs identify nonmitochondrial factors that activate
is able to fuse with the cell membrane, resulting in mitochondrial biogenesis.
direct release of their cargo into the cytoplasm of the
recipient cells. In addition, the nonmitochondrial CONCLUSIONS
cargo of D15 M-EVs facilitated PGC-1 a –mediated
mitochondrial biogenesis in the recipient iCMs This study demonstrated the feasibility of M-EV–
(Figure 3B), suggesting that mitochondrial and non- mediated transfer of mitochondria and their related
mitochondrial cargos may contribute to sufficient and bioenergetics cargo for the treatment of failing hearts.
persistent restoration of intracellular energetics in Our data provide an important proof-of-concept of
hypoxia-injured iCMs and post-MI murine heart. successful in vitro augmentation of cellular bio-
M-EVs are smaller than the expected size range of energetics and effective in vivo intramyocardial
intracellular mitochondria, suggesting that the frag- delivery of M-EVs in a murine model of MI. This
mented mitochondria via mitochondrial fission ma- iCM-based autologous source of bioenergetics may
chinery are engulfed by EVs. Consistent with this provide an effective therapy for mitochondria-related
finding, D15s that exhibited higher expression level of diseases including advanced HF patients.
mitochondrial dynamics-related genes secreted more
ACKNOWLEDGMENTS The authors thank Meredith
mitochondria-containing EVs than D50s did
Weglarz (Stanford Shared FACS Facility) and John
(Figure 1B). These mitochondria are not “dysfunc-
Perrino (Stanford Cell Science Imaging Facility).
tional mitochondria” that are destined to be elimi-
nated by autophagy because they retain structural
and functional integrity (Figures 1B to 1E). FUNDING SUPPORT AND AUTHOR DISCLOSURES
Our data showed that human-derived M-EVs
Dr. Ikeda has received funding support through the Stanford Dean’s
restored bioenergetics in mouse heart tissues. It is
Postdoctoral Fellowship, Japan Heart Foundation/Bayer Yakuhin,
reported that mitochondrial fusion proteins such as and an American Heart Association postdoctoral fellowship. Ms.
Mfn1/2 and Opa1 in mouse and MFN1/2 and OPA1 in Santoso has received funding support through the Alpha Omega
Alpha Carolyn B. Kuckein Student Research Fellowship. Dr. Yang has
human cells have a high degree of homologies.
received funding support through National Institutes of Health, Na-
Human-derived mitochondria efficiently fuse to the tional Heart, Lung, and Blood Institute grants 1 K24 HL130553K and
mitochondrial networks in mice cells and mixing of UM1 L12026; and funding support from SPARK Stanford University
the matrix protein content is completed within 4 h funding support through Stanford Cardiovascular Institute Seed
Grant. All other authors have reported that they have no relationships
(38). These data suggest that human-derived mito-
relevant to the contents of this paper to disclose.
chondria inside the M-EVs were transferred into
mouse cardiomyocytes and fuse to native mitochon-
ADDRESS FOR CORRESPONDENCE: Dr. Phillip C.
drial networks, resulting in enhanced mitochondrial
Yang, Biomedical Innovations Building, 240 Pasteur
function.
Drive, 3053, Stanford, California 94304, USA. E-mail:
STUDY LIMITATIONS. First, we have only examined [email protected]. Twitter: @YangLab10, @gentaro_
the EVs secreted from iCMs on the assumption that ikeda, @meeshsantoso.
JACC VOL. 77, NO. 8, 2021 Ikeda et al. 1087
MARCH 2, 2021:1073–88 Vesicle-Mediated Mitochondria Transfer

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: M-EVs TRANSLATIONAL OUTLOOK: Studies of transendo-

transfer mitochondrial and nonmitochondrial substances cardial administration of M-EVs in larger animal models
into cardiomyocytes, stimulating mitochondrial biogen- would facilitate assessment of the therapeutic potential
esis and enhancing energy metabolism in failing of this strategy for patients with HF.
myocardium.

REFERENCES

1. Yancy CW, Jessup M, Bozkurt B, et al. 2013 survival in severe chronic heart failure. N Engl J IFN and TNF responses in endothelial cells. Circ
ACCF/AHA guideline for the management of heart Med 2001;344:1651–8. Res 2019;125:43–52.
failure: a report of the American College of Car-
12. Fox K, Ford I, Steg PG, Tendera M, Ferrari R. 22. Garcia-Martinez I, Santoro N, Chen Y, et al.
diology Foundation/American Heart Association
Ivabradine for patients with stable coronary artery Hepatocyte mitochondrial DNA drives nonalco-
Task Force on Practice Guidelines. J Am Coll Car-
disease and left-ventricular systolic dysfunction holic steatohepatitis by activation of TLR9. J Clin
diol 2013;62:e147–239.
(BEAUTIFUL): a randomised, double-blind, pla- Invest 2016;126:859–64.
2. Taylor CJ, Ordóñez-Mena JM, Roalfe AK, et al. cebo-controlled trial. Lancet 2008;372:807–16.
Trends in survival after a diagnosis of heart failure 23. Hayakawa K, Esposito E, Wang X, et al.
13. Swedberg K, Komajda M, Böhm M, et al., for Transfer of mitochondria from astrocytes to neu-
in the United Kingdom 2000-2017: population
the SHIFT Investigators. Ivabradine and outcomes rons after stroke. Nature 2016;535:551–5.
based cohort study. BMJ 2019;364:l223.
in chronic heart failure (SHIFT): a randomised
3. Neubauer S. The failing heart: an engine out of 24. Takahashi K, Tanabe K, Ohnuki M, et al. In-
placebo-controlled study. Lancet 2010;376:
fuel. N Engl J Med 2007;356:1140–51. duction of pluripotent stem cells from adult hu-
875–85.
man fibroblasts by defined factors. Cell 2007;131:
4. Sekiguchi M. Electron microscopical observa-
14. CONSENSUS Trial Study Group. Effects of 861–72.
tions of the myocardium in patients with idiopathic
enalapril on mortality in severe congestive heart
cardiomyopathy using endomyocardial biopsy. 25. National Research Council. Guide for the Care
failure: results of the Cooperative North Scandi-
J Mol Cell Cardiol 1974;6:111–22. and Use of Laboratory Animals. Washington, DC:
navian Enalapril Survival Study (CONSENSUS).
5. Starling RC, Hammer DF, Altschuld RA. Human N Engl J Med 1987;316:1429–35. The National Academies Press, 1996. Available
myocardial ATP content and in vivo contractile at: www.nap.edu/catalog/5140.html. Accessed
15. Pitt B, Zannad F, Remme WJ, et al., for the
function. Mol Cell Biochem 1998;180:171–7. December 10, 2019.
Randomized Aldactone Evaluation Study In-
6. Beer M, Seyfarth T, Sandstede J, et al. Absolute vestigators. The effect of spironolactone on 26. Yang X, Pabon L, Murry CE. Engineering
concentrations of high-energy phosphate metab- morbidity and mortality in patients with severe adolescence: maturation of human pluripotent
olites in normal, hypertrophied, and failing human heart failure. N Engl J Med 1999;341:709–17. stem cell-derived cardiomyocytes. Circ Res 2014;
myocardium measured noninvasively with (31)P- 114:511–23.
16. Brown DA, Perry JB, Allen ME, et al. Expert
SLOOP magnetic resonance spectroscopy. J Am consensus document: mitochondrial function as a 27. Lundy SD, Zhu W-Z, Regnier M,
Coll Cardiol 2002;40:1267–74. therapeutic target in heart failure. Nat Rev Cardiol Laflamme MA. Structural and functional matura-
7. Sharov VG, Todor AV, Silverman N, Goldstein S, 2017;14:238–50. tion of cardiomyocytes derived from human
Sabbah HN. Abnormal mitochondrial respiration in pluripotent stem cells. Stem Cells Dev 2013;22:
17. Murry CE, Soonpaa MH, Reinecke H, et al.
failed human myocardium. J Mol Cell Cardiol 1991–2002.
Haematopoietic stem cells do not trans-
2000;32:2361–7. differentiate into cardiac myocytes in myocardial 28. Zhou R, Yazdi AS, Menu P, Tschopp J. A role
8. Sebastiani M, Giordano C, Nediani C, et al. In- infarcts. Nature 2004;428:664–8. for mitochondria in NLRP3 inflammasome activa-
duction of mitochondrial biogenesis is a mal- 18. Jung J-H, Fu X, Yang PC. Exosomes generated tion. Nature 2011;469:221–5.
adaptive mechanism in mitochondrial from iPSC-derivatives: new direction for stem cell 29. Vega RB, Horton JL, Kelly DP. Maintain-
cardiomyopathies. J Am Coll Cardiol 2007;50: therapy in human heart diseases. Circ Res 2017; ing ancient organelles: mitochondrial biogen-
1362–9. 120:407–17. esis and maturation. Circ Res 2015;116:
9. Garnier A, Zoll J, Fortin D, et al. Control by 19. de Couto G, Gallet R, Cambier L, et al. Exoso- 1820–34.
circulating factors of mitochondrial function and mal MicroRNA transfer into macrophages medi- 30. Lehman JJ, Barger PM, Kovacs A, Saffitz JE,
transcription cascade in heart failure: a role for ates cellular postconditioning. Circulation 2017; Medeiros DM, Kelly DP. Peroxisome proliferator-
endothelin-1 and angiotensin-II. Circ Heart Fail 136:200–14. activated receptor gamma coactivator-1 pro-
2009;2:342–50.
20. Jang SC, Crescitelli R, Cvjetkovic A, et al. motes cardiac mitochondrial biogenesis. J Clin
10. Waagstein F, Hjalmarson A, Varnauskas E, Mitochondrial protein enriched extracellular vesi- Invest 2000;106:847–56.
Wallentin I. Effect of chronic beta-adrenergic re- cles discovered in human melanoma tissues can be
31. Xu R, Greening DW, Rai A, Ji H, Simpson RJ.
ceptor blockade in congestive cardiomyopathy. Br detected in patient plasma. J Extracell Vesicles
Highly-purified exosomes and shed microvesicles
Heart J 1975;37:1022–36. 2019;8:1635420.
isolated from the human colon cancer cell line
11. Packer M, Coats AJ, Fowler MB, et al. for the 21. Puhm F, Afonyushkin T, Resch U, et al. Mito- LIM1863 by sequential centrifugal ultrafiltration
Carvedilol Prospective Randomized Cumulative chondria are a subset of extracellular vesicles are biochemically and functionally distinct.
Survival Study Group. Effect of carvedilol on released by activated monocytes and induce type I Methods 2015;87:11–25.
1088 Ikeda et al. JACC VOL. 77, NO. 8, 2021

Vesicle-Mediated Mitochondria Transfer MARCH 2, 2021:1073–88

32. Tachibana A, Santoso MR, Mahmoudi M, et al. 35. Shiba Y, Gomibuchi T, Seto T, et al. Allogeneic 38. Yoon YG, Haug CL, Koob MD. Interspecies
Paracrine effects of the pluripotent stem cell- transplantation of iPS cell-derived cardiomyocytes mitochondrial fusion between mouse and human
derived cardiac myocytes salvage the injured regenerates primate hearts. Nature 2016;538: mitochondria is rapid and efficient. Mitochondrion
myocardium. Circ Res 2017;121:e22–36. 388–91. 2007;7:223–9.

33. Kim MJ, Hwang JW, Yun C-K, Lee Y, Choi Y-S. 36. Kaza AK, Wamala I, Friehs I, et al. Myocardial
Delivery of exogenous mitochondria via centrifu- rescue with autologous mitochondrial trans-
KEY WORDS bioenergetics, heart failure,
gation enhances cellular metabolic function. Sci plantation in a porcine model of ischemia/
human stem cells, mitochondria, myocardial
Rep 2018;8:3330. reperfusion. J Thorac Cardiovasc Surg 2017;153:
infarction
934–43.
34. Adamiak M, Cheng G, Bobis-Wozowicz S, et al.
Induced pluripotent stem cell (iPSC)-derived 37. Bertero E, Maack C, O’Rourke B. Mitochon-
extracellular vesicles are safer and more effective drial transplantation in humans: “magical” cure A PPE NDI X For supplemental methods,
for cardiac repair than iPSCs. Circ Res 2018;122: or cause for concern? J Clin Invest 2018;128: figures, tables, and references, please see the
296–309. 5191–4. online version of this paper.