Mesenchymal Stem Cell Implantation in a Swine

Myocardial Infarct Model: Engraftment and

Functional Effects
Jay G. Shake, MD, Peter J. Gruber, MD, PhD, William A. Baumgartner, MD,
Guylaine Senechal, MS, Jennifer Meyers, BS, J. Mark Redmond, MD,
Mark F. Pittenger, PhD, and Bradley J. Martin, PhD
The Johns Hopkins Medical Institutions, and Osiris Therapeutics, Inc, Baltimore, Maryland

Background. A novel therapeutic option for the treat- Results. Microscopic analysis showed robust engraft-
ment of acute myocardial infarction involves the use of ment of MSCs in all treated animals. Expression of
mesenchymal stem cells (MSCs). The purpose of this muscle-specific proteins was seen as early as 2 weeks and
study was to investigate whether implantation of autol- could be identified in all animals at sacrifice. The degree
ogous MSCs results in sustained engraftment, myogenic of contractile dysfunction was significantly attenuated at
differentiation, and improved cardiac function in a swine 4 weeks in animals implanted with MSCs (5.4% ⴞ 2.2%
myocardial infarct model. versus ⴚ3.37% ⴞ 2.7% in control). In addition, the extent
Methods. MSCs were isolated and expanded from bone of wall thinning after myocardial infarction was mark-
marrow aspirates of 14 domestic swine. A 60-minute left edly reduced in treated animals.
anterior descending artery occlusion was used to produce Conclusions. Mesenchymal stem cells are capable of
anterior wall infarction. Piezoelectric crystals were engraftment in host myocardium, demonstrate expres-
placed within the ischemic region for measurement of sion of muscle specific proteins, and may attenuate
regional wall thickness and contractile function. Two contractile dysfunction and pathologic thinning in this
weeks later animals autologous, Di-I–labeled MSCs (6 ⴛ model of left ventricular wall infarction. MSC cardiomy-
107) were implanted into the infarct by direct injection. oplasty may have significant clinical potential in attenu-
Hemodynamic and functional measurements were ob- ating the pathology associated with myocardial
tained weekly until the time of sacrifice. Immunohisto- infarction.
chemistry was used to assess MSC engraftment and (Ann Thorac Surg 2002;73:1919 –26)
myogenic differentiation. © 2002 by The Society of Thoracic Surgeons

U nfortunately for persons who suffer large myocar-

dial infarctions the adult heart lacks regenerative
capabilities. In these patients congestive heart failure
tion in small animal preparations [2– 4]. The purpose of
this study was to investigate whether autologous bone
marrow derived stem cells are capable of engraftment
develops when a critical threshold of myocardium dies into host myocardium and differentiation into myogenic
and compensatory mechanisms become overwhelmed. A lineages, and could contribute to improved regional wall
novel therapeutic option to prevent the progression to- function in a large animal myocardial infarct model.
ward heart failure involves the introduction of healthy
cells into the infarct in an effort to repopulate the region, Material and Methods
commonly referred to as cellular cardiomyoplasty. This Marrow Harvest
new population of cells may potentially prevent the
Fourteen female swine (25 to 35 kg) were sedated with
typical wall thinning, scar formation, and decline in
ketamine (1,000 mg intramuscular [IM]) and brought into
systolic function seen in these patients. the laboratory. Intravenous access was established through
One particularly promising cell transplantation alter- an ear vein and the animals were anesthetized with thio-
native involves the use of mesenchymal stem cells pental (10 mL of 5% intravenous [IV]). They were then
(MSC). These cells possess pluripotential capabilities [1] placed in a ventral recumbency position with the hind legs
and have been shown to undergo myogenic differentia- tucked under the body. The iliac crest area was prepared
and draped using sterile technique and approximately

Presented at the Thirty-seventh Annual Meeting of The Society of Doctors Senechal, Meyers, Pittenger, and Martin dis-
Thoracic Surgeons, New Orleans, LA, Jan 29 –31, 2001. close that they have a financial relationship with Osiris
Address reprint requests to Dr Martin, Osiris Therapeutics, Inc, 2001 Therapeutics, Inc.
Aliceanna St, Baltimore, MD 21231; e-mail: [email protected].

© 2002 by The Society of Thoracic Surgeons 0003-4975/02/$22.00

Published by Elsevier Science Inc PII S0003-4975(02)03517-8
1920 SHAKE ET AL Ann Thorac Surg
STEM CELL GRAFTING IN SWINE INFARCT MODEL 2002;73:1919 –26

14 mL of bone marrow was aspirated into a syringe con- ered with a polytetrafluoroethylene (PTFE) sheet (Pre-
taining 6,000 units of heparin. All animals received bu- clude Pericardial Membrane; Goretex, Flagstaff, AZ). The
prenorphine (0.3 mg IM) as an analgesic before returning to chest was closed and a single mediastinal tube (18F) was
the animal facility. The marrow aspirates were then sent to placed to reestablish a negative intrapleural pressure and
the Osiris Pre-Clinical/Animal Tissue Core Facility for MSC evacuate any remaining blood or irrigation solution. The
isolation and expansion. inhaled anesthetic was then turned off, the animal extu-
bated when appropriate, and allowed to recover. The
MSC Isolation, Expansion, and Labeling
chest tube was removed when there was no visible air
Mesenchymal stem cell isolation, and confirmation of a leak or blood accumulation. Animals received postoper-
uniform population following colony expansion, was per- ative antimicrobial therapy (amoxicillin 500 mg orally
formed at Osiris as previously described [1]. Briefly, the
twice daily for 3 days) and buprenorphine (0.3 mg IM
bone marrow aspirates were passed through a density
twice daily for 3 days) for postoperative pain.
gradient to eliminate unwanted cell types. When plated,
a small number of cells developed in visible symmetric
colonies by days 5 to 7. Hematopoietic cells, fibroblasts, Sonomicrometry
and other nonadherant cells were washed away during Four piezoelectric crystals (Sonometrics Corp, London, On-
medium changes. The remaining purified mesenchymal tario, Canada) were placed in the infarct region: epicar-
stem cell population was further expanded in culture dium, endocardium, and two midmyocardium. The crystals
until approximately 60 million MSCs were present. On were placed in a spatial orientation in order to measure
the day of implantation, the MSCs were labeled with a myocardial wall thickness and segment length throughout
cross-linkable membrane dye, CM-DiI (Molecular the cardiac cycle (see Fig 1). Barbed piezoelectric crystals
Probes, Inc, Eugene, OR) following the manufacturer’s were used to prevent migration during the remodeling
protocol. Prior to injection the cells were thoroughly phase. Also all crystals were further fixed in place with
washed and held on ice at an approximate concentration sutures as the leads exited the epicardium. Confirmation of
of 20 million cells/mL. proper crystal arrangement was possible through online
Myocardial Infarction and Instrumentation analysis during placement. Two dimensional analysis of
epicardial and endocardial crystals (LV thickness) ex-
Swine were sedated with ketamine (1,000 mg IM), in-
pressed on an X-Y axis allowed immediate calculation of LV
duced with sodium pentobarbital (30 mg/kg IV), endo-
wall thickness and crystal excursion. Transducer wires were
tracheally intubated, and maintained on isoflurane (0.8%
externalized as previously described and connected to a
to 1.5%) through a Narkomed ventilator. Electrocardio-
graphic monitoring and rectal temperatures were contin- Series 5001 Digital Sonomicrometer (Sonometrics Corp,
uously displayed. Animals were then prepared and London, Ontario, Canada) for data acquisition. Ventilation
draped in routine sterile fashion. Benzathine penicillin was suspended during sonomicrometry measurements if
(1.2 ␮ IM) and gentamycin (80 mg IV) were given as cyclical variations were noted. Off-line analysis was per-
preoperative antimicrobial therapy. A midline sternot- formed on a desktop computer using software from Sono-
omy was performed and the heart suspended in a peri- metrics. End-systole was defined as 20 milliseconds prior to
cardial cradle. A tygon catheter (0.06 in inner diameter) the peak dp/dt, while end-diastole was taken at the upward
was placed in the apex of the left ventricle, sutured in deflection of the dp/dt trace.
place, and used for left ventricular (LV) pressure and
blood gas monitoring. The left anterior descending (LAD) MSC Implantation
coronary artery was dissected free just distal to the first Two weeks after the production of a myocardial infarc-
diagonal branch and isolated with a vessel loop. A brief tion the animals were returned to the operating room for
occlusion of the coronary artery was performed to iden- a second procedure. The animals were anesthetized and
tify the region to be rendered ischemic. Four piezoelectric monitored as usual and the chest reentered with great
crystals were then placed in the area destined for infarc-
caution to prevent disruption of the instrumentation.
tion to measure regional wall motion (segment shorten-
Prior to implantation, the MSCs were Di-I labeled,
ing and wall thickening). At completion of the surgical
washed two times, and resuspended in 3 mL of vehicle
instrumentation, a 15-minute stabilization period was
with 2000 units of heparin. Once the area of infarction
allowed prior to baseline recordings of hemodynamics
was clearly visualized six injections containing either
and regional cardiac function. Lidocaine (2 mg/kg IV
bolus, then 0.5 mg/min IV) was started and continued autologous MSCs in 0.5 mL (total 6.0 ⫻ 107 cells in 3 mL;
until the time of reperfusion. A 60-minute occlusion of n ⫽ 7), or a comparable volume of vehicle (n ⫽ 5), was
the LAD was used to produce myocardial infarction. At given using a 30-gauge needle. The injections were given
the end of 60 minutes and after baseline recordings, the in a manner such that the regions between the piezoelec-
snare was released and reperfusion was visually con- tric crystals received all of the cells. This technique was
firmed. At this time the piezoelectric crystal leads and the used to optimize the likelihood of demonstrating re-
LV catheter were brought together below the xyphoid gional functional changes after MSC cardiomyoplasty.
process, tunneled posteriorly, and externalized near the After the MSC/vehicle injection, the chest was closed and
scapula. The pericardium was approximated as well as the animals recovered as previously described. Every
possible and any remaining epicardial surface was cov- week for the first month and biweekly thereafter the ani-
Ann Thorac Surg SHAKE ET AL 1921
2002;73:1919 –26 STEM CELL GRAFTING IN SWINE INFARCT MODEL

Fig 1. Location of the left ventricle

infarct (A) and spatial orientation of
the piezoelectric crystals (B). An en-
docardial and epicardial pair of
crystals measured the wall thickness
and active wall thickening during
the cardiac cycle. Two mid-myocar-
dial crystals measured the segment
shortening. Mesenchymal stem cells
were directly injected in the localized
region between the piezoelectric
crystals.

mals were sedated and brought to the laboratory for serial Statistical Analysis
hemodynamic measurements and sonomicrometry. All data in the text are presented as mean ⫾ SEM.
Between-group comparison of piezoelectric crystal (end
Histology and Microscopy diastolic wall thickness and systolic wall thickening) and
At the end of the experimental protocol the animals were hemodynamic measurements (LV pressures, heart rate,
anesthetized with ketamine and euthanized with saturated dp/dt) were made using a repeated measures analysis of
solution of potassium chloride and the heart excised with variance (ANOVA). Infarct size was analyzed using an
the instrumentation intact. The appropriate location and unpaired Student’s t test. Differences were deemed sig-
alignment of the piezoelectric crystals was verified in the nificant when p values ⬍ 0.05 were obtained.
explanted heart. If the endocardial and epicardial crystals
were found to be misaligned (not perpendicular to the LV Animal Care
circumference) dimensional data from that animal were not All work was preapproved by the Johns Hopkins School
included in the final data analysis. Subsequently, any ad- of Medicine Animal Care and Use Committee and per-
hesions fixed to the heart’s epicardial surface were sharply formed according to the guidelines outlined in the
dissected free and the atria and great vessels were removed “Guide for the Care and Use of Laboratory Animals”
at their origin. The left ventricle was then sliced into 6 or 7 published by the National Institutes of Health (NIH
rings parallel to the minor axis of the left ventricle. All rings publication 85-23, revised 1985).
were then weighed for subsequent determination of infarct
size. Both surfaces of the ventricular slices were photo-
graphed while immersed in water to eliminate highlights Results
and digital images were obtained. The percentage of each Of the 14 animals entered into the protocol, 12 were used
slice that was infarcted was determined using the Scion in the final data analysis. The initial two animals were
Image Analysis program (Beta version 4.0.2). First the in- excluded from analysis after discovery of life-threatening
farcted area of the ventricular ring was calculated and infections involving the externalized instrumentation.
reported in pixels. Then the regional area of the entire left The infections were discovered prior to randomization
ventricular ring was determined. Based on these measure- into either the MSC or control groups. After the addition
ments (in pixels), the percent infarction for the ring could be of short-term postoperative antibiotics no further ani-
calculated. These percentages were multiplied by the mals suffered serious infection. Two additional animals
weight of the rings and summed, converting the percentage (both MSC) while included in the histology and micros-
values to grams, and enabled a reconstruction of infarct size copy evaluation were omitted from regional functional
of the left ventricle. The tissue was then prepared for analysis owing to malfunctioning piezoelectric crystals or
immunohistochemistry and confocal microscopy by freez- waveforms incompatible with proper placement.
ing in O.C.T. and serial sectioning at 5 ␮m. Antibodies
examined included ␣-actinin (Sigma, St. Louis, MO), tropo- Hemodynamics
nin-T (Sigma), tropomyosin (Sigma), myosin heavy chain- There was no significant difference in any of the moni-
MHC (DSHB), and phospholamban (Affinity Bioreagents, tored hemodynamic variables (heart rate, left ventricular
Golden, CO). The confocal microscope (PCM 2000; Nikon, pressure, and dp/dt) noted between groups. It should be
Melville, NY) was equipped with two excitation lasers of noted that this model resulted in severe cardiac injury
different wavelengths to simultaneously identify two sepa- with reductions in contractile function noted in both
rate molecular markers. groups after infarction. Of these 12 animals presented,
1922 SHAKE ET AL Ann Thorac Surg
STEM CELL GRAFTING IN SWINE INFARCT MODEL 2002;73:1919 –26

significant augmentation in systolic wall thickening at 4

weeks (5.4% ⫾ 2.2%) when compared with control ani-
mals (F value ⫽ 7.72, p value ⫽ 0.04). These data do not
demonstrate a global restoration of contractile function to
preinfarct levels but rather a significant attenuation in
the degree of dysfunction after infarction in the localized
region of MSC implantation.

Histology and Microscopy

Fig 2. End diastolic wall thickness as measured by sonomicrometry. The hearts of all animals had large anterior/septal infarc-
Note the extent of left ventricular wall thinning in the control group tions and showed signs of enlargement at sacrifice. The
after production of the left ventricular infarct. Mesenchymal stem size of the infarctions was not significantly different
cell (MSC) implantation resulted in an attenuation of wall thinning between groups (12.45% ⫾ 2.07% of LV in control versus
when compared with control animals at all time points. Data are 14.69% ⫾ 1.65% in MSC). Cross-sections revealed trans-
group averages and presented as mean ⫾ SEM. mural infarctions in all animals and compensatory hy-
pertrophy events in several. Upon gross examination
clinical signs of congestive heart failure (exercise intoler-
control hearts had marked LV wall thinning in the region
ance and pulmonary edema) developed in 8 by the time
of vehicle injection consistent with the sonomicrometry
they were euthanized. This symptomatic reduction in
results. In contrast the animals receiving MSC treatment
cardiac function was characteristically observed between
had apparently normal gross LV geometry and mainte-
6 and 8 weeks and the incidence was not significantly
nance of wall thickness at the site of MSC implantation.
different between groups. Figure 4 illustrates these findings in photographs of LV
sections from the site of sonomicrometery crystal place-
Sonomicrometry
ment and MSC injection. Histologic examination of
In the 2-week period after myocardial infarction, wall MSC-treated myocardium revealed collagen-rich scar
thickness and systolic function were markedly reduced in tissue with an open, lattice work appearance as opposed
all animals. Prior to implantation of either MSCs or saline to the dense, tightly pack extracellular matrix found in
vehicle wall thickness had been reduced approximately the thinned walls of control animals.
25% (from 9.9 ⫾ 0.8 mm to 7.9 ⫾ 1.6 mm) and systolic wall There was successful engraftment of MSCs in all
thickening had been replaced by paradoxic wall thinning. treated animals as shown by the large number of Di-I–
The MSC and vehicle injections (3 mL) were within the labeled cells seen on microscopy. Labeled MSCs could be
region bordered by the piezoelectric crystals. This fact is identified at the site of implantation for as long as 6
demonstrated by the acute change in end diastolic wall months. Implanted MSCs were not found in regions of
thickness (EDWT) measured immediately after injection myocardium remote from the injection site or systemi-
(7.9 ⫾ 1.6 mm preinjection versus 9.4 ⫾ 2.1 mm postin- cally. Cells appear to preferentially engraft in regions of
jection, p ⬍ 0.05). This increase in EDWT persisted in the necrotic tissue and adhere to the collagen rich matrix.
MSC group and by 2 weeks measured 11.3 ⫾ 2.4 mm. In Clusters of cells forming a continuum could occasionally
contrast, the wall thickness in control animals returned to be found in proximity to viable myocardium (Fig 5A and
preinjection valves by 2 weeks and the region of infarc- B). In addition to the impressive number of cells found in
tion continued to thin throughout the duration of the the infarct region colocalization of immunofluorescence
study. There was a sustained improvement in EDWT in indicates that a significant portion of the surviving cells
the MSC-treated animals when compared with control at had begun to express several muscle specific proteins
all time points postinjection (see Fig 2). This augmenta- (␣-actinin, tropomyosin, toponin-T, myosin heavy chain,
tion of wall thickness neared statistical significance at 4
weeks (F ⫽ 4.36, p ⫽ 0.07).
In addition to preventing pathological wall thinning in
the region of treatment, MSC implantation also appeared
to locally augment systolic function. More precisely MSC
implantation attenuated the degree of contractile dys-
function observed after infarction. These data are sum-
marized in Figure 3. Percent systolic wall thickening was
defined as EDWT ⫺ ESWT / EDWT ⫻ 100, where
EDWT ⫽ end diastolic wall thickness and ESWT ⫽ end
systolic wall thickness. Both control and treated animals Fig 3. Systolic wall thickening as measured via sonomicrometry and
displayed profound reductions in regional function im- expressed as a percent of baseline values. Myocardial infarction re-
mediately after infarction (15.7% ⫾ 6.7% at baseline sulted in a marked reduction in contractile function in both control
and treated animals. While paradoxic wall thinning was observed in
versus ⫺0.65% ⫾ 3.1% preimplant). The decline in func-
control animals (negative thickening values), mesenchymal stem cell
tion progressed over time in the vehicle-treated animals (MSC) implantation significantly reduced the degree of systolic dys-
(⫺7.4% ⫾ 3.8% at 6 weeks). Systolic wall thickening was function. Differences between the two groups were evident by 2
often replaced by paradoxic wall thinning such that weeks and became statistically significant by 4 weeks. Data are
diastolic wall thickness values were larger than systolic group averages and presented as mean ⫾ SEM. *p ⬍ 0.05 versus
values. In contrast, MSC-treated animals exhibited a control.
Ann Thorac Surg SHAKE ET AL 1923
2002;73:1919 –26 STEM CELL GRAFTING IN SWINE INFARCT MODEL

Fig 4. Gross left ventricular (LV) cross-sections at the level of mesenchymal stem cell (MSC) injection and piezoelectric crystal placement.
Large transmural infarctions were observed in all animals. Control animals (top row) had wall thinning and loss of normal ventricle chamber
geometry associated with remodeled infarctions. MSC-treated animals (bottom row) had preserved wall thickness and LV geometry. (IFN ⫽
interferon.)

and phospholamban) not present in MSCs prior to im- MSCs in host myocardium yet quantification of cell
plantation (Fig 5C–F). Muscle-specific protein expression number remained technically difficult. In an effort to offer
was observed as early as 2 weeks postimplantation and the MSCs an optimal survival advantage they were
persisted as all subsequent time points examined (up to introduced at 2 weeks postinfarction to avoid the peak of
6 months). No ectopic tissues were observed and no the postinfarct inflammatory phase.
evidence for MSC differentiation to adipose, bone, carti- By 2 weeks postimplantation there was expression of
lage, tendon, or any tissue other than muscle was seen in muscle-specific proteins identifiable by immunofluores-
any animal. Similarly, no significant inflammatory infil- cence. The lack of true “cardiac-specific” markers for
trates were identified at the site of MSC implantation. immunohistochemistry has limited our ability to conclu-
sively state that MSCs implanted in the heart differenti-
ate to a cardiac phenotype (as opposed to skeletal mus-
Comment cle). The only muscle marker that gives some degree of
insight into this question is phospholamban. The expres-
The primary findings of the present study were (1)
sion of phospholamban indicates that if these MSCs are
mesenchymal stem cells engraft into host myocardium
differentiating to either a cardiac or a slow-twitch skeletal
when implanted by direct injection; (2) the MSCs ex- phenotype. Theoretically either may be beneficial in the
pressed muscle-specific proteins after implantation into heart due to their nonfatigueable nature. It should also
areas of injured myocardium; (3) local milieu dependent be noted that no nonmuscle tissue types (ie, bone, carti-
factors as opposed to exogenous substances influence lage, adipose, parenchyma) were observed in any animal,
differentiation toward a myogenic lineage; and (4) im- consistent with the notion of milieu-specific
planted cells had a beneficial impact on cardiac remod- differentiation.
eling when implanted 2 weeks postinfarction. Perhaps the most significant observation in this study
Mesenchymal stem cells are pluripotent cells capable is the maintenance of wall thickness seen in the MSC-
of differentiation into many cell types [1]. Furthermore, treated group. This conclusion is supported by both the
muscle-specific differentiation has been seen [3]. What piezoelectric crystal data and the gross morphology of
has made MSCs particularly appealing to investigators in the ventricle at the time of sacrifice. This MSC-induced
cellular cardiomyoplasty is that these cells are easily maintenance of wall thickness could be due either to the
obtained from bone marrow, can be expanded in culture, addition of tissue or to an inhibition of remodeling
and can be cryopreserved for future use. process observed after infarction. Potentially MSCs could
In this study we observed significant engraftment of alter collagenase activity or other enzymatic pathways
1924 SHAKE ET AL Ann Thorac Surg
STEM CELL GRAFTING IN SWINE INFARCT MODEL 2002;73:1919 –26

Fig 5. Mesenchymal stem cell (MSC) engraftment and muscle-specific protein-expression using immunohistochemistry. The Di-I labeled mes-
enchymal stem cells (red) and muscle protein-specific antibodies (green) fluoresce under confocal microscopy. Colocalization (yellow) con-
firms cells having both probes present. Specifically this represents specific muscle protein expression within the MSCs being imaged. (B) This
hematoxylin and eosin (H&E)–stained serial section corresponds to (A) illustrating a cluster of MSC in proximity to host myocardium. Several
muscle-specific proteins were identified in treated myocardium. Examples include ␣-actinin (␣-act) (A and D), troponin T (TnT) (C), phos-
pholamban (PLB) (E), and tropomyosin (TM) (F) expression at 4 weeks postimplantation.

responsible for the pathologic wall thinning observed in mined, it appears that extracellular matrix alterations are
remodeled myocardium. Although the precise mecha- likely involved.
nism by which MSC implantation limits the extent of The implantation of MSCs had a beneficial effect on
myocardial thinning after infarction has yet to be deter- systolic wall thickening despite the absence organized
Ann Thorac Surg SHAKE ET AL 1925
2002;73:1919 –26 STEM CELL GRAFTING IN SWINE INFARCT MODEL

contractile elements (5.4% ⫾ 2.2% in treated animals at 4 In conclusion, mesenchymal stem cell implantation
weeks versus ⫺3.4% ⫾ 2.7% in control). It is apparent that may attenuate contractile dysfunction and pathologic
the mechanism of improvement does not lie in systolic remodeling of the ventricular wall after infarction. MSCs
contraction of implanted MSCs but rather in the attenu- show sustained engraftment within infracted myocar-
ation of paradoxical systolic wall thinning. Regional con- dium and exhibit myogenic differentiation as evidenced
tractile function is severely impaired in both groups after by expression of muscle-specific proteins. Further study
infarction. In other words the degree of dysfunction is needed to fully characterize the engrafted MSCs, their
appears exaggerated in the control group when com- integration with host myocardium, optimal cell number
pared with MSC treatment. Taken together the data and delivery technique, and the degree to which they
suggest that the implantation of MSCs may result in a contribute to global changes in cardiac function.
more compliant, less stiff region of ventricle with im-
prove diastolic filling properties. Improvements in dia-
stolic dysfunction have also been reported in other mod- Osiris Therapeutics, Inc, received a research award from the
Department of Commerce, National Institute of Standards and
els of cellular cardiomyoplasty [5]. Technology/Advanced Technologies Program (NIST/ATP). The
Currently one of the more popular cell types used by Cardiac Surgery Laboratory is partially funded by generous gifts
cardiomyoplasty investigators has been immature skele- from the Mildred and Carmont Blitz Cardiac Research Fund.
tal myoblasts, commonly referred to as “satellite cells.” JGS is currently the Irene Piccinini Investigator in Cardiac
Surgery established to recognize outstanding research trainees
Murry and coworkers [6] have shown conversion of
in Cardiac Surgery at The Johns Hopkins Medical Institutions.
satellite grafts to fatigue-resistant, slow-twitch fibers bet- The authors would like to thank the scientists at Osiris for their
ter suited for cardiac workloads. Taylor and associates [5] contributions and continued enthusiasm. Also, they wish to
have shown improved myocardial performance in a rab- thank Mr Jeffrey Brawn and Ms Melissa Haggerty for their
bit model where myoblasts were incorporated into outstanding technical assistance. This project could not have
been completed without the participation of all of them.
cryoinfarcts. Furthermore, Chiu and associates [7] have
demonstrated “milieu-influenced” differentiation of sat-
ellite cells into cardiac-like muscle cells. Lastly, xenoge-
neic and allogeneic myoblast cell transplantation has References
been successful in a large animal model. Additionally
1. Pittenger MF, Mackay AM, Beck SC, et al. Multilineage
these transplanted cells were thought to initiate prolifer- potential of adult human mesenchymal stem cells. Science
ation of surrounding microvasculature by a presumed 1999;284:143– 6.
angiogenic factor [8]. 2. Wang JS, Shum-Tim D, Galipeau J, et al. Marrow stromal
Cellular cardiomyoplasty has also been done using cells for cellular cardiomyoplasty: feasibility and potential
clinical advantages. J Thorac Cardiovasc Surg 2000;120:999 –
cells other than skeletal myoblasts. Li and colleagues [9] 1006.
have shown improved cardiac function in a model of 3. Wakitani S, Saito T, Caplan AI. Myogenic cells derived from
smooth muscle cell transplantation. Despite the socially rat bone marrow mesenchymal stem cells exposed to 5-aza-
and politically controversial use of fetal tissue for re- cytidine. Musc Nerve 1995;18:1417–26.
search several groups have investigated the use of fetal 4. Ferrari G, Cusella-DeAngelis G, Coletta M, et al. Muscle
regeneration by bone marrow derived myogenic progeni-
cardiomyocytes as a method to alleviate heart failure tors. Science 1998;279:1528–30.
[10 –12]. Others have used bone marrow derived cells to 5. Taylor DA, Atkins BZ, Hungspreugs P, et al. Regenerating
repair or replace damaged muscle fibers [2, 4]. These cells functional myocardium: improved performance after skele-
were exposed to an appropriate microenvironment in tal myoblast transplantation. Nat Med 1998;4:929–33.
6. Murry CE, Wiseman RW, Schwartz SM, et al. Skeletal
order to direct the cells toward a myogenic lineage. myoblast transplantation for repair of myocardial necrosis.
Preliminary evidence suggests that exogenous agents J Clin Invest 1996;98:2512–23.
such as 5-azacytidine can induce cultured bone marrow 7. Chiu RC, Zibaitis A, Kao RL. Cellular cardiomyoplasty:
cells to differentiate into cardiac-like muscle cells prior to myocardial regeneration with satellite cell implantation.
injection into myocardial infarcts [3, 13]. Ann Thorac Surg 1995;60:12–18.
8. Van Meeter CH, Claycomb WC, Delcarpio JB, et al. Myoblast
A major limitation of the current model was the fact transplantation in the porcine model: a potential technique
that all animals progressed to failure by 6 to 8 weeks. As for myocardial repair. J Thorac Cardiovasc Surg 1995;110:
MSCs were only injected cells in a relatively small area 1442– 8.
within infarct, the overall morbidity associated with such 9. Li RK, Jia ZQ, Weisel RD, et al. Smooth muscle cell trans-
plantation into myocardial scar tissue improves heart func-
an injury was not altered. This pilot study was done to tion. J Mol Cell Cardiol 1999;31:513–22.
demonstrate MSC engraftment and differentiation and to 10. Scorsin M, Hagege AA, Dolizy I, et al. Can cellular trans-
examine what effect MSC implantation may have on plantation improve function in doxorubicin-induced heart
regional function following infarction. Because of the failure? Circ 1998;98(suppl):151–5.
premature progression toward failure, adequate time 11. Scorsin M, Hagege AA, Vilquin JT, et al. Comparison of the
effects of fetal cardiomyocytes and skeletal myoblast trans-
may not have been allowed for the cells to synthesize and plantation on post-infarction left ventricular function. J Tho-
assemble a functional contractile apparatuses. Future rac Cardiovasc Surg 2000;119:1169–75.
work will determine if injection of a larger number of 12. Koh GY, Soonpaa MH, Klug MG, et al. Stable fetal cardio-
cells throughout the entire region infarct improves global myocyte grafts in the hearts of dystrophic mice and dogs.
J Clin Invest 1995;96:2034– 42.
function and prolongs survival. Such a model may allow 13. Tomita S, Li RK, Weisel RD, et al. Autologous transplanta-
time for more complete myogenic differentiation of im- tion of bone marrow cells improves damaged heart function.
planted MSCs. Circulation 1999;100(supple II):247–56.
1926 SHAKE ET AL Ann Thorac Surg
STEM CELL GRAFTING IN SWINE INFARCT MODEL 2002;73:1919 –26

DISCUSSION
DR CONSTANTINE MAVROUDIS (Chicago, IL): That was a DR ZUND: The second question is you got the cells from the
very nice study. I have a couple of questions. Did you note or bone marrow. Do you think you can get these cells as well from
measure the orientation of the transplanted cells and whether the blood?
they aligned with the intercalated disks and whether in fact they
took their place alongside the other myocytes? And was the DR SHAKE: Potentially. I know Dr. Meyer’s group has been
thickening that you demonstrated due to the fact that you doing some work with that. We have very good success with
injected something into the area, or in fact do you think it was bone marrow aspiration. It works very well. Perhaps the most
due to the potential cells or the differentiating cells that you interesting thing that we have found is that we can use alloge-
placed there? I didn’t catch it well—Your groups were one group neic bone marrow derived stem cells. They seem to be totally
that you injected yourselves and another group that you injected immunoprivileged and we are currently doing a study where we
something other than cells, or you did not? are injecting cells from one animal into another with no immu-
nosuppression and we are seeing no rejection.
DR SHAKE: The vehicle was injected.
DR CHARLES R. BRIDGES (Philadelphia, PA): I was wondering
DR MAVROUDIS: The vehicle. So then in many respects you if you looked at remodeling in terms of measuring other param-
have answered one part of my question. But I wonder if you had eters, in terms of, for example, ventricular dilatation. Your
injected some other cells that might have done something to results with respect to systolic thickening are fascinating and
cause the thickness and maybe not have caused the potential for outstanding. But in terms of overall ventricular remodeling, I
contractility that you’re seeking. This is a very nice study. This know that with just the crystals you put there, you really could
has a potential for revolutionizing this field and I hope you only assess local changes. But did you look globally? Did you
continue and find all the other answers. Congratulations. echo any of these animals, for example, to see if you saw changes
there?
DR SHAKE: Thank you. As far as your first question about
orientation, we did not see the MSCs orienting themselves in the
DR SHAKE: Basically all of those questions are hopefully going
center of the infarct. Because, remember, we inject them in the
to be able to be answered in our next group of animals which are
middle of the infarct where the crystals are placed. We did see
ongoing. We have nearly finished the series. We have injected
some stray cells that made their way to the edge of the infarct,
the entire infarct with allogeneic stem cells, and we are echoing
very close to the host myocardium. Those cells were orienting
all animals.
themselves more appropriately and appeared to be larger and
And as far as remodeling, this experiment was really a pilot
elongated. Obviously there is some communication between the
study to determine the mode of delivery as well as the volume
host myocardium and these cells. In our next study, hopefully
and the number of cells to deliver. So we were not trying to look
we’ll be able to answer the question as far as what would be the
at any of the specific questions that you are asking. But certainly
orientation of the cells adjacent to host myocardium.
And second of all, we were pleased when we saw increased those are great questions and great things to pursue in our
wall thickness and thickening at the site of the piezoelectric follow-up studies.
crystals. We do not know why it is thickened there. It could be
twofold. We do see a large number of cells present. We also see DR JOHN E. MAYER, JR (Boston, MA): If I might ask a question.
a lot of colocalization demonstrating MSCs with contractile I am interested in your speculations on what you think are the
proteins. But we also believe there is, perhaps, a paracrine-like local factors, be they either physical or biochemical, cytokine,
effect ongoing, where there is less remodeling that is occurring whatever, signals that you think are inducing these relatively
in the heart. Perhaps there is actually prevention of the collag- uncommitted cells down a myogenic lineage.
enases or other similar kind of enzymes that are breaking down
that area. These MSCs may be communicating with the cells DR SHAKE: That is a great question. We would like to know
secreting these enzymes. what the physical or chemical signal is that tells the MSCs to
pass down a given lineage.
DR GREGOR ZUND (Zurich, Switzerland): Congratulations on
your paper. What do you think about the pretreatment of such DR MAYER: I was hoping you could tell me.
cells before implantation? And a second question: do you think
you would get such cells as well from the blood? DR SHAKE: We do not know what that is, but we think that
there is a soluble factor. At Osiris they have seen that there is
DR SHAKE: Well, we have seen that we do not need to pretreat. some evidence for a soluble factor that actually sends MSCs
As I mentioned, other investigators are certainly trying to send down a lineage. What it is, we do not know, and it is not purified
skeletal myoblasts or MSCs down lineages. We have seen that if at this point.
just injected into the right environment, the MSCs differentiate
into what they should be in that environment. Our collaborators DR MAYER: The second question would be whether or not you
at Osiris are using human MSCs for replacement of collagen in have tried to deliver any physical signals to these cells in vitro,
osteoarthritic knees, and we have also seen appropriate differ- in culture, and observed what effects that might have?
entiation when injected into hearts. So I am not sure if we need
to send them down the lineage. DR SHAKE: We have not but it would be interesting.