APPENDIX
Supplemental Materials:
Animal Model:
Animals were given heparin (100-200 IU/kg) before the start of occlusion. Lidocaine (2-4 mg/Kg
bolus followed by continuous infusion of 50μg/kg/min) was administered to the animal to control
episodes of arrhythmias. X-ray coronary angiograms were made, and a coronary guide wire was
then advanced into the LAD. A coronary angioplasty balloon sized to match the diameter of the
mid LAD was delivered to that segment and inflated at the lowest pressure necessary to occlude
distal flow. After 150-minutes, the balloon was deflated and removed to allow reperfusion. Postprocedural pain was managed with Buprenorphine (0.003 mg/Kg) and transdermal Fentanyl
patch (25-75μg/Kg/h) for 72 hours. The pig was recovered and the myocardium was allowed to
heal for 90 days.
Supplemental Table 1: Timeline for evaluation of efficacy and safety.
Time Point
BSN
1M 3M Post Immediately >24h- 1 Wk 1M
2M 3M
(MI) Post
MI
post TESI 84h Post post Post Post Post
MI
(TESI)
TESI
TESI TESI TESI TESI
Infarct Creation
X
Myocardial Biopsy
X
Bone Marrow
X
Biopsy
Injection
X
Cardiac Imaging
X
X
X
X
X
X
Sacrifice and
X
Necropsy
Blood work
X
X
X
X
X
X
X
X
X
Body weight
X
X
X
X
X
X
X
X
Continuous
X
X
X
X
X
X
Arrhythmia
monitoring
Daily Assessment of clinical status by a member of laboratory, weekly health assessment by
Veterinarian BSN, baseline; MI, myocardial infarction; M, month; TESI, transendocardial stem cell
injection
2
Supplemental Table 2. Pig Study Demographics
Placebo
MSC
CSC/ MSC
P-value
Age (months)
MI
12.7±0.4
13.5±0.9
13.0±0.3
0.7
TESI
16.0±0.4
16.7±0.9
16.1±0.4
0.7
Endpoint
19.0±0.4
19.5±1.3
19.1±0.4
0.7
Body weight
MI
27.5±1.2
26.2±1.1
27.3±1.4
0.7
(kg)
TESI
30.3±0.9
31.5±1.6
30.4±0.5
0.7
Endpoint
35.3±2.1
36.5±3.3
34.4±2.5
0.9
Sample size (n) TESI
6
6
8
n/a
Endpoint
6
4*
7†
Follow up
After TESI
3.0±0.0
3.0±.0.1*
3.0±0.0
0.5
(months)
MSC, mesenchymal stem cell; CSC, cardiac stem cell; MI, myocardial infarction;
TESI, transendocardial stem cell injection.* Mean of 4 animals, 2 animals have 4-5
weeks of follow up as explained in the Supplementary results section. **1 animal
excluded for delivery error.
Cell Manufacturing Process
C-Kit+ CSCs (CSCs) were isolated from five to eight 1-2 mm endomyocardial biopsies, obtained
from the septal wall of the right ventricle immediately following MI/reperfusion, as previously
described(1). Biopsies were grown in Ham’s F12 medium (Life Technologies, Grand Island, NY)
containing 10% fetal bovine serum (FBS), 0.2mM L-glutathione, 5 mU/mL erythropoietin, 10
ng/mL human fibroblast growth factor (Recombinant Human FGF-basic PeproTech), and 1%
glutamine/penicillin/streptomycin (GPS; Life Technologies). CSCs that migrated out of biopsies
were isolated using anti-c-kit (anti-CD117) antibodies (Anti-Mouse CD117 APC eBioscience, and
CD117 Alexa Fluor 488 AbD Serotec) conjugated with magnetic microbeads and processed
through the VarioMACSTM separation system (Miltenyi). MSCs from the bone marrow aspirate
were isolated by Ficoll-density centrifugation and plastic adherence. Cells were cultured and
amplified in alpha-MEM medium supplemented with 20% FBS and 1% GPS. MSCs were harvested
and cryopreserved(2). Immediately prior to cryopreservation, MSC samples were sent for sterility
testing.
Immunohistochemistry
Swine hearts were fixed for >72 hours in 10% buffered formalin and sliced transversely into 4mm-thick slices from the apex to the base. Slices were embedded in paraffin and processed for
microscopy. Twelve slides (four from each zone; IZ, BZ and RZ) were randomly chosen from each
3
animal for quantification of phospho-histone H3 (PHH3) positive nuclei. Each slide was incubated
with the primary antibody against PHH3 (Abcam, Cambridge, MA, #14955), followed by a goat
secondary anti-mouse IgG antibody (Life Technologies #A11032). DAPI (4', 6-diamidino-2phenylindole) was added to stain nuclei.
Endothelial function: Flow mediated dilation (FMD)
Endothelial function was measured by the brachial artery FMD method as previously
described(3). All FMD (%) measurements were performed on fasted, anesthetized animals with
the animal supine and rested for 10 minutes. A limb was comfortably immobilized in the extended
position, allowing for ultrasound scanning of the brachial artery. Each vessel image was recorded
twice; first at rest, followed by inflation of a cuff to supra-systolic pressure (≈50 mmHg above
systolic pressure) for 5 minutes. The second arterial diameter was imaged and recorded after the
cuff was deflated. FMD (%) was calculated as [(arterial diameter after 5 min of cuff inflation −
arterial diameter at rest)/basal arterial diameter]. Ultrasound measurements of brachial artery
diameter were performed at baseline, post-MI and monthly post-TESI.
Cardiac MRI
Cardiac magnetic resonance imaging (CMR) studies were conducted using a Siemens TIM Trio 3T
(Erlangen, Germany) scanner running with Numaris 4 / software version MR B19 using a 4channel body coil with ECG gating and short breath-hold acquisitions. CMR was done at baseline,
1 month post MI, 3 months post MI (pretreatment), 1 month post TESI, 2 months post TESI, and 3
months post TESI (sacrifice).
All animals were sedated with ketamine, induced with sodium pentobarbital, intubated and
anesthesia was maintained with isoflurane. ECG leads were placed on the limbs and chest for
gating image acquisition. The animals were placed in the supine position and a phased-array
surface coil was placed over the heart.
Steady-state free precession (SSFP) cine images in 2-chamber, 4 chamber and short axis planes
(slice thickness 4mm, field of view [FOV] 280mm, matrix 256x80, repetition time (TR) 41msec,
echo time (TE) 1.5 msec, number of averages 2, band width [BW] 840kHz, flip angle 80 degrees)
were obtained.
4
At end-diastole and end-systole, user defined epicardial and endocardial borders were drawn in
contiguous short axis cine images covering the apex to mitral valve plane using Qmass MR 7.2 and
7.5 (Medis Inc) Software to calculate global function.
Perfusion imaging was performed after two intravenous bolus injections of gadolinium
(Magnevist, Bayer Healthcare) 0.15mmol/kg with an ECG-gated interleaved saturation recovery
gradient echo planar imaging pulse sequence (EFGRET-ET). An entire short-axis stack was
acquired every 2–4 heartbeats. Imaging parameters were as follows: TR/TE = 7.2 and 1.8 ms; flip
angle = 20o; 128 x 128 matrix; 8 mm slice thickness/no gap; bandwidth 125 kHz; 28-com FOV;
and 0.5–1 NSA. Using the 16-segment model, infarct zone, border zone, and remote zone was
defined as segment 8 and 9, segment 10 and 13, and segment 2 and 4, respectively.
Short axis and 2-chamber long axis delayed enhancement (DE) images (slice thickness of 4mm
with no gap, FOV 240-340mm, matrix 256x80, TR/TE 450ms/3ms, BW 150kHz, and a flip angle of
25 degrees) were acquired 10 minutes following intravenous infusion of gadolinium. Infarct scar
size was calculated from the short axis delayed myocardial enhancement images covering the apex
to the mitral valve plane. Epicardial and endocardial contours of the LV were drawn with a semiautomated tool. The intensity of a normal region of myocardium was calculated and scar tissue
was determined by using an intensity threshold two standard deviations above normal
myocardium.
Regional function was measured by tagged CMR images using Diagnosoft 2.71 (Diagnosoft Inc).
Three contiguous short axis tagged images encompassing the scar were selected for analysis. User
defined epicardial and endocardial contours were drawn to create a 24-segment mesh for each
slice, and Eulerian circumferential strain (Ecc) for each segment at each time point of the cardiac
cycle was measured. Using the RV insertion as a reference point, corresponding tagged images
were classified as either transmural infarction segments (scar ≥ 50% of the respective segment),
margin of infarcted area (scar < 50% of the respective segment) and remote area (remote
segments to the infarction area). The peak Ecc for each area was calculated by averaging the peak
Ecc (more negative is greater contractility) from each individual segments of the specified zone.
The peak diastolic strain rate was calculated by averaging the most positive point in the early
diastolic portion of the strain rate curve from each individual segments of the specified zone. The
same slices and areas were used between all-time points.
5
Perfusion imaging was performed after two intravenous bolus injections of gadolinium
(Magnevist, Bayer Healthcare) 0.15mmol/kg with an ECG-gated interleaved saturation recovery
gradient echo planar imaging pulse sequence (EFGRET-ET). An entire short-axis stack was
acquired every 2–4 heartbeats. Imaging parameters were as follows: TR/TE = 7.2 and 1.8 ms; flip
angle = 20o; 128 x 128 matrix; 8 mm slice thickness/no gap; bandwidth 125 kHz; 28-com FOV;
and 0.5–1 NSA. Using the 16-segment model, infarct zone, border zone, and remote zone was
defined as segment 8 and 9, segment 10 and 13, and segment 2 and 4, respectively.
Short axis and 2-chamber long axis delayed enhancement (DE) images (slice thickness of 4mm
with no gap, FOV 240-340mm, matrix 256x80, TR/TE 450ms/3ms, BW 150kHz, and a flip angle of
25 degrees) were acquired 10 minutes following intravenous infusion of gadolinium. Infarct scar
size was calculated from the short axis delayed myocardial enhancement images covering the apex
to the mitral valve plane. Epicardial and endocardial contours of the LV were drawn with a semiautomated tool. The intensity of a normal region of myocardium was calculated and scar tissue
was determined by using an intensity threshold two standard deviations above normal
myocardium.
Vascular density
Immunohistochemical staining for von Willebrand factor-related antigen was used to assess the
vascular density in samples of infarct zone, border zone, and remote zone for each subject. All
images were obtained using 10x magnification and the settings were constant for the entire study.
All slides were photographed using a digital camera (Nikon ECLIPSE TS100). The vessel stains
were assessed in five randomly selected fields per section, and quantified using Image J software,
version 1.44p (National Institutes of Health, USA).
Supplemental Results
Survival
One animal was euthanized 5 weeks after TESI due to respiratory distress less than 24 hours after
a surgical intervention (cardiac rhythm monitoring device replacement). Following veterinarian
consultation, euthanasia was recommended and performed. A second animal was euthanized 4
weeks after TESI when she was found with respiratory distress associated with a large hematoma
adjacent to the trachea. The hematoma was mixed in consistency and was at the same site of the
surgical incision on the neck. The veterinarian was consulted for management, efforts to manage
6
the hematoma with drainage and compression were instituted, but finally euthanasia was
performed.
Transendocardial delivery success rate
Transendocardial product was delivered to 20 animals. All injections were successfully delivered
to the infarct-border zone with similar unipolar voltage in the injection site (CSCs/MSCs: 6.3±0.2;
MSCs: 6.4±0.3; placebo: 6.4±0.2, P=NS). One animal from the CSC/MSC cell group was excluded
from the analysis due to delivery error caused by the inability to acquire a complete anatomical
map during TESI.
Arrhythmia monitoring
Swine from all groups were monitored for cardiac arrhythmias after TESI until the end of the
study (CSC/MSC n=5; MSCs n=4; Placebo n=4) using an implantable continuous monitoring
REVEAL device (Medtronic Inc.). All EKG recordings were read by an interpreter blinded to
treatment and verified by an expert Cardiologist.
In summary, all arrhythmias were non-
sustained Ventricular Tachycardia (duration <30s), self-terminated, and did not require any
intervention. Supplementary Table 5 and 6 summarize the EKG findings.
Supplemental Table 3. Arrhythmia episodes
Duration Number of
Animal
BPM
(sec)
beats
Episode 1
7
16
286
Pig 1 Episode 2
8
18
273
Episode 3
4
10
261
Episode 1
4
15
333
Pig 2
Episode 2
4
21
333
Episode 1
4
13
231
Pig 3 Episode 2
10
38
353
Episode 3
8
27
250
Pig 4 Episode 1
13
41
286
Date
~9w Post INJ
~9w Post INJ
~9w Post INJ
~3w Post INJ
4w Post INJ
~10w Post INJ
~10w Post INJ
~10w Post INJ
8w Post INJ
7
Supplemental Table 4. Continuous cardiac monitoring after TESI
Electrocardiogram Analysis
Animal ID
No
Bradycardia
Ventricular
Ventricular
Arrhythmia
Tachycardia Fibrillation
(# of episodes)
Placebo (Plasmalyte A)
Pig 1
X
Absent
Absent
Absent
Pig 2
X
Absent
Absent
Absent
Pig 3
X
Absent
Absent
Absent
Pig 4
Absent
1*
Absent
MSC (200M) alone
Pig 5
X
Absent
Absent
Absent
Pig 6
Absent
3*
Absent
Pig 7
X
Absent
Absent
Absent
Pig 8
X
Absent
Absent
Absent
CSC (1M) / MSC (200M)
Pig 9
X
Absent
Absent
Absent
Pig 10
X
Absent
Absent
Absent
Pig 11
X
Absent
Absent
Absent
Pig 12
Absent
3*
Absent
Pig 13
Absent
2*
Absent
*No sustained ventricular tachycardia
PEA / Asystole
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Absent
Necropsy
Whole body (brain, liver, spleen, kidney, lung and ileum) necropsies were performed to assess
ectopic tissue formation. There were no signs of tumorigenicity or ectopic tissue formation on
whole body necropsy 3 months following administration of either cell product. One MSC-treated
animal had evidence of a mild pneumonia that was not clinically evident while alive and one
placebo treated animal had generalized vasculitis, which was considered a pre-existing condition.
8
Supplemental Table 5. Cardiac MRI myocardial infarct size and viability before and after
TESI
3m post-MI 3 months Between group % Change
Between
(before
after TESI comparison
group
TESI)
(absolute
comparison
values)
(% change)
MI size
Placebo
11.8±2.0
10.6±2.2
P=0.02
−12.9±4.2
P<0.001
(% of LV
MSC
15.1±1.4
8.7±0.6*
−44.1±6.8*
mass)
CSC/MSC
11.8±1.4
7.2±0.6*
−37.2±5.4*
MI size
Placebo
6.8±1.2
6.8±1.2
P=0.02
−1.4±7.2
P<0.001
(g)
MSC
7.8±0.6
5.8±0.5*
−30.4.1±4.2*
CSC/MSC
6.4±0.7
4.8±0.4*
−22.9±4.4*
Placebo
52.0±2.9
59.1±4.7
13.5±5.9
Viable
MSC
41.1±4.6
63.6±8.5*
P=0.6
43.7±13.3*
P=0.0002
Tissue (g)
CSC/MSC
48.4±2.8
62.7±2.9*
30.9±7.0*
LV mass
Placebo
58.1±2.6
65.7±4.2
P=0.04
13.0±5.2
P=0.2
(g)
MSC
52.2±2.6
67.7±8.7
28.0±10.2
CSC/MSC
54.4±2.9
67.4±3.0
24.8±5.6
9
Supplemental Table 6. Global and Regional Function
Baseline 3m post- 3 months Between group % Change
Between
MI
after
comparison
group
(before
TESI
(absolute
comparison
TESI)
values)
(% change)
Ejection
Placebo
65.0±3.6 43.0±3.5 44.4±3.1
P=0.02
6.7±4.1
P=0.01
Fraction
MSC
64.0±3.3 43.6±1.5 46.1±3.4
6.4±3.2
(%)
CSC/MSC
62.0±3.6 45.6±1.8 52.4±4.4*
13.9±6.2*
Stroke
Placebo
29.9±3.2 27.5±3.2 32.3±2.9
P=0.5
22.4±12.0
P=0.008
Volume
MSC
26.3±3.7 27.8±2.5 32.1±3.4
21.2±4.7
(mL)
CSC/MSC
26.9±3.5 24.8±2.6 36.5±4.0*
47.2±11.1*
Cardiac
Placebo
3.2±0.3 2.8±0.3 3.2±0.2
P=0.3
15.5±9.5
P=0.1
Output
MSC
2.7±0.2 2.7±0.2 3.6±0.2
27.8±13.6
(L/min)
CSC/MSC
2.5±0.2 2.4±0.2 3.5±0.3*
50.5±11.3*
Diastolic
Placebo
Not done 0.3 ±0.1 0.3±0.1
P=0.9
−9.0±13.4
P=0.3
Strain
MSC
0.3±0.1 0.4±0.1
6.9±6.5
Rate
CSC/MSC
0.3±0.1 0.4±0.1
18.9±8.6
MI, myocardial infarction; TESI, transendocardial stem cell injection; MSC, mesenchymal stem cell;
CSC, cardiac stem cell. *P<0.05 compared to 3 months after MI
References for Supplementary Materials
1.
Johnston PV, Sasano T, Mills K et al. Engraftment, differentiation, and functional benefits of
autologous cardiosphere-derived cells in porcine ischemic cardiomyopathy. Circulation
2009;120:1075-83, 7 p following 1083.
2.
Williams AR, Hare JM. Mesenchymal stem cells: biology, pathophysiology, translational
findings, and therapeutic implications for cardiac disease. Circ Res 2011;109:923-40.
3.
Liu Y, Xiong Y, Liu D et al. The effect of enhanced external counterpulsation on C-reactive
protein and flow-mediated dilation in porcine model of hypercholesterolaemia. Clin Physiol
Funct Imaging 2012; 32:262-7.
10
Supplemental Figure 1. Study Design
MI
RANDOMIZATION
BMA
RV Biopsy
0
TESI
6 months
3 months
Cell manufacturing (n=20)
Placebo vs MSCs alone vs. CSC/MSC
n=6
n=6
n=8*
1, 2 and 3 months post TESI MRI scan
NOGA system
MRI
PV-loops
MI
(n=28)
10 injec on sites (5mL)
in the infarct-border zone
MRI
PV-loops
(n=20)
MRI
PV-loops
Sac
(n=18*)
MI, myocardial infarction; TESI, transendocardial stem cell injection; MRI, magnetic resonance
imaging; BMA, bone marrow aspiration RV, right ventricle; MSC, mesenchymal stem cell; CSC,
cardiac stem cell; * One animal was excluded from this group because of delivery error.
11
Supplemental Figure 2. Cardiac serum markers and white blood cell counts changes after
myocardial infarction and transendocardial stem cell injection.
1.6
15
Placebo
MSCs
CSC/MSC
1.2
WBC (103/m/L)
Troponin I (ng/mL)
1.4
Troponin-I
A
1.0
0.8
0.6
WBC count
B
10
5
0.4
0.2
0.0
3m
2m
1m
h
1w
h
48
24
TE
SI
3m
I
1m
CPK
C
400
CK-MB (ng/mL)
10000
1m
3m
SI
TE
h
24
h
48
1w
1m
2m
3m
CK-MB
D
300
200
100
5000
1m
2m
3m
3m
1w
2m
h
48
1m
h
24
1w
SI
TE
48
h
3m
24
h
1m
TE
SI
MI
3m
0
0
1m
0
MI
MI
CPK (U/L)
15000
0
0
20000
M
0
0
(A) Troponin I, (B) white blood cells (WBC) count, (C) creatinine phosphokinase (CPK) and (D)
creatinine kinase MB (CK-MB) isoenzyme exhibit no significant difference between CSC/MSC, MSC
alone and placebo groups at baseline, immediately post-myocardial infarction and 3 months after
transendocardial stem cell injection (TESI).
12
Supplemental Figure 3. DE-MRI scar size correlates with gross pathology scar size.
The Pearson correlation coefficient between scar size assessed by DE-MRI and the scar size
measured by gross pathology was r=0.93 (95% CI, 0.80 to 0.98; p<0.0001).
Supplemental Figure 4. An example of vascular (von Willebrand) staining in the infarct zone of
an animal from each of the treatment groups. A. Placebo, B. MSC, C. CSC/MSC. Bar = 100 µm for all
panels.
A
B
C