transplantation of human embryonic stem cell-derived ... · we injected 3 (larova, teltow,...

Ttmparfs

FRSDbR

a

Journal of the American College of Cardiology Vol. 50, No. 19, 2007© 2007 by the American College of Cardiology Foundation ISSN 0735-1097/07/$32.00P

PRECLINICAL STUDY

Transplantation of Human EmbryonicStem Cell-Derived Cardiomyocytes ImprovesMyocardial Performance in Infarcted Rat Hearts

Oren Caspi, MD,* Irit Huber, PHD,* Izhak Kehat, MD, PHD,*† Manhal Habib, MD,*Gil Arbel, MSC,* Amira Gepstein, PHD,* Lior Yankelson, MD,* Doron Aronson, MD,†Rafael Beyar, MD, PHD,† Lior Gepstein, MD, PHD*†

Haifa, Israel

Objectives We evaluated the ability of human embryonic stem cells (hESCs) and their cardiomyocyte derivatives (hESC-CMs) to engraft and improve myocardial performance in the rat chronic infarction model.

Background Cell therapy is emerging as a novel therapy for myocardial repair but is hampered by the lack of sources for hu-man cardiomyocytes.

Methods Immunosuppressed healthy and infarcted (7 to 10 days after coronary ligation) rat hearts were randomized toinjection of undifferentiated hESCs, hESC-CMs, noncardiomyocyte hESC derivatives, or saline. Detailed histologi-cal analysis and sequential echocardiography were used to determine the structural and functional conse-quences of cell grafting.

Results Transplantation of undifferentiated hESCs resulted in the formation of teratoma-like structures. This phenome-non was prevented by grafting of ex vivo pre-differentiated hESC-CMs. The grafted cardiomyocytes survived, pro-liferated, matured, aligned, and formed gap junctions with host cardiac tissue. Functionally, animals injectedwith saline or nonmyocyte hESC derivatives demonstrated significant left ventricular (LV) dilatation and func-tional deterioration, whereas grafting of hESC-CMs attenuated this remodeling process. Hence, post-injury base-line fractional shortening deteriorated by 50% (from 20 � 2% to 10 � 2%) and by 30% (20 � 2% to 14 � 2%)in the saline and nonmyocyte groups while improving by 22% (21 � 2% to 25 � 3%) in the hESC-CM group.Similarly, wall motion score index and LV diastolic dimensions were significantly lower in the hESC-CM animals.

Conclusions Transplantation of hESC-CMs after extensive myocardial infarction in rats results in the formation of stable car-diomyocyte grafts, attenuation of the remodeling process, and functional benefit. These findings highlight thepotential of hESCs for myocardial cell therapy strategies. (J Am Coll Cardiol 2007;50:1884–93) © 2007 by theAmerican College of Cardiology Foundation

ublished by Elsevier Inc. doi:10.1016/j.jacc.2007.07.054

viol

psctdba

he adult heart lacks effective repair mechanisms, and,herefore, any significant cell loss, such as occurs duringyocardial infarction, may lead to the development of

rogressive heart failure. Cell therapy and tissue engineeringre emerging as novel therapeutic paradigms for myocardialepair (1,2). Cardiomyocyte transplantation (derived frometal or neonatal rat heart tissue or from mouse embryonictem cells) was associated with smaller infarcts (3,4), pre-

rom the *Sohnis Family Research Laboratory for Cardiac Electrophysiology andegenerative Medicine, the Rappaport Family Institute for Research in the Medicalciences, Technion-Israel Institute of Technology, Haifa, Israel; and the †Cardiologyepartment, Rambam Medical Center, Haifa, Israel. This study was partially funded

y the Israel Science Foundation (grant no. 1078/04), by the American Cell Therapyesearch Foundation, and by the Grand Family research grant.

tManuscript received March 5, 2007; revised manuscript received July 26, 2007,

ccepted July 30, 2007.

ented cardiac dilatation and remodeling (5), and evenmproved myocardial performance (6). Clinical translationf these promising strategies is hampered, however, by theack of cell sources for human cardiac tissue.

See page 1894

A possible solution to the aforementioned cell-sourcingroblem may be the recently described human embryonictem cell (hESC) lines (7). These unique cells have theapacity for prolonged undifferentiated proliferation in cul-ure while maintaining the potential to differentiate intoerivatives of all 3 germ layers. Recently, using the embryoidody (EB) differentiating system, we and others establishedreproducible cardiomyocyte differentiation system from

he hESC (8–11). Cells isolated from the contracting EBs

dtrditpd

rhweddwrm

M

IeeTcCdpcisw

lteIga(raEdAFvp5P

wud

CS3avodwwaiTc�gcwwirtcmc�gcmH3cfacGBClF(IwasPoP�f(dr(G

1885JACC Vol. 50, No. 19, 2007 Caspi et al.November 6, 2007:1884–93 Myocardial Regeneration With hESC

emonstrated molecular, structural, and functional proper-ies of early-stage human cardiomyocytes (8–14). Moreecently the ability of the human embryonic stem cell–erived cardiomyocytes (hESC-CMs) to survive (15) and

ntegrate structurally and functionally with healthy host cardiacissue, both in vitro (in coculturing studies) and in vivo (byacing the heart in an animal model of slow heart rate), wasemonstrated (16,17).In the current study, we continued to explore the possible

ole of this unique tissue for myocardial repair in diseasedearts. Specifically, our aims were 3-fold: 1) to determinehether existing heart tissue can provide the appropriate

nvironment to augment cardiomyocyte differentiation of un-ifferentiated hESC; 2) to determine the ability of ex vivo-ifferentiated hESC-CMs to survive, proliferate, and integrateith host tissue after grafting into the uninjured and infarcted

at heart; and 3) to assess the ability of cell grafting to improveyocardial performance in this model.

ethods

n vitro hESC cardiomyocyte differentiation. Undiffer-ntiated hESCs (H9.2-clone) were propagated on mousembryonic fibroblast feeder layer as previously described (8).o induce differentiation, hESCs were dispersed to small

lumps using collagenase-IV 1 mg/ml (Gibco, Invitrogen,arlsbad, California) and cultured in suspension for 7 to 10ays. The generated EBs were plated on gelatin-coatedlates and observed for the appearance of spontaneousontractions. For the transplantation studies, the contract-ng areas (30 to 45 days of differentiation) were microdis-ected and dissociated into small clusters (20 to 100 cells)ith collagenase-B (1 mg/ml) (Roche, Basel, Switzerland).To identify the transplanted cells, we used a number of

abeling, staining, and tracking techniques: 1) labeling withhe fluorescent cell tracer Vybrant-carboxyfluorescein diac-tate, succinimidyl ester (25 �mol/l, Molecular Probes,nvitrogen); 2) tagging with genetic markers: enhancedreen fluorescent protein (eGFP) and nLacZ expression waschieved using lentiviral transduction and electroporationNucleofector program-A23 and V-transfection solution),espectively; and 3) immunostaining for human-specificntigens.ffect of pre-existing cardiac environment on hESCifferentiation. All animal studies were approved by thenimal Board and Safety Committee of the Technionaculty of Medicine. Primary cultures of neonatal ratentricular myocytes (density 1.5 � 106 cells/ml) wererepared as previously described (16). Early-stage EBs (2 todays in suspension) were labeled with CM-DiI (Molecularrobes), added to the rat cultures, and followed for 4 weeks.To assess for the effect of the in vivo cardiac environment,

e injected 3 � 106 undifferentiated hESCs into theninjured and infarcted rat left ventricular (LV) myocar-
ium. The hearts were harvested 4 weeks later. v
ell transplantation. Maleprague-Dawley rats (250 to00 g) were anesthetized (ket-mine/xylasine), intubated, andentilated. After left thoracot-my, the proximal left anteriorescending coronary artery (LAD)as ligated. A second thoracotomyas performed 7 to 10 days later,

nd cells were transplanted to thenfarcted area at 4 different sites.hree groups were studied: 1) a

ontrol group in which saline (300l) was injected (n � 8); 2) aroup in which 1.5 � 106 non-ardiomyocyte hESC derivativesere grafted (n � 9); these cellsere derived from noncontract-

ng differentiating EBs and rep-esent a heterogeneous popula-ion of early stage differentiatingells of different lineages (18),ost of which are epithelial-like

ells; and 3) a group in which 1.5106 hESC-CMs (n � 8) were transplanted. To prevent

raft rejection, animals from all groups were treated withyclosporine A (15 mg/kg/day) and methylprednisolone (2g/kg/day).istological examination. The hearts were harvested at

6 h to 8 weeks after grafting, frozen in liquid nitrogen, andryo-sectioned (8-�m sections). Immunostaining was per-ormed using monoclonal antibodies to human-mitochondriantigens, troponin I (TnI), Ryanodine, Cx43 (all from Chemi-ion, Temecula, California), humanKi-67, (DakoCytomation,lostrup, Denmark), TRA1-60 and TRA1-81 (Hybridomaank, Iowa City, Iowa), Oct-4 (Santa Cruz, Santa Cruz,alifornia), sarcomeric �-actinin (Sigma), and human-human

eukocyte antigen (HLA)-A,B,C antigens (BD-Pharmingen,ranklin Lakes, New Jersey), or polyclonal antibodies to eGFP

MBL, Woods Hole, Massachusetts) and collagen types I andII (Southern Biotech, Birmingham, Alabama). Preparationsere incubated with secondary antibodies at 1:100 dilutions

nd analyzed by confocal microscopy (Nikon and Bio-Radcanning system, Hercules, California).olymerase chain reaction (PCR) analysis. The presencef human cells within the rat hearts was evaluated usingCR-based deoxyribonucleic acid (DNA) amplification of the-satellite region of the human chromosome 17. Hearts were

rozen in liquid nitrogen and homogenized with PolytronKinematica, Newark, New Jersey). Genomic DNA was pro-uced using Easy-DNA-Kit (Invitrogen). Polymerase chaineaction was carried out using Red-Load-Taq-Master-MixLAROVA, Teltow, Germany). The forward primer wasGGATAATTTCAGCTGACTAAACAG, and the re-

Abbreviationsand Acronyms

eGFP � enhanced greenfluorescent protein

FS � fractional shortening

hESC � human embryonicstem cell

hESC-CM � humanembryonic stemcell–derived cardiomyocyte

HLA � human leukocyteantigen

LAD � left anteriordescending coronary artery

LV � leftventricle/ventricular

LVDd � left ventricular end-diastolic diameter

MLC-2a � myosin lightchain-2a

PCR � polymerase chainreaction

Tn � troponin

erse primer was GTGTTTCATAG
CTGCTCTTTCCA.

LatCep(wntap(rumGGEpdMluwaa�aeSdTml(metBprw(w

R

Cmmdwec

flcmwecs(t

cthHnetsaiibltmlIhiahodbmgnftfass

atattncam3

1886 Caspi et al. JACC Vol. 50, No. 19, 2007Myocardial Regeneration With hESC November 6, 2007:1884–93

aser microdissection and reverse transcription PCRnalysis. Cryo-sectioned slices (20 �m) containing theransplanted nLacZ-expressing cells were stained with 1%resylviolet, dehydrated, and frozen in �80°C. Slides were

valuated for beta-Galactosidase activity using LacZ re-orter kit (Invitrogen). Laser capture microdissectionP.A.L.M. Microlaser Technologies, Bernreid, Germany)as performed to isolate the positively stained cells. Ribo-ucleic acid (RNA) was isolated from these cells usinghe RNA easy microkit (Qiagen, Hilden, Germany),nd reverse transcription into complementary DNA waserformed using Reverse-iT-1st-Strand-Synthesis KitABgene, Epsom, United Kingdom). Polymerase chaineaction for myosin light chain-2a (MLC-2a) was executedsing 1U Taq-DNA polymerase (PeqLab, Erlangen, Ger-any). The forward primer was AAGGTGAGTGTCCCA-AGG, and the reverse primer was ACAGAGTTTATT-AGGTGCCC.chocardiography. Transthoracic echocardiography waserformed 5 to 7 days after coronary ligation and 30 to 60ays after cell grafting, using the GE-Vivid3 system (10-Hz transducer, GE Healthcare, Haifa, Israel). The fol-

owing parameters were measured: 1) maximal left ventric-lar end-diastolic (LVDd) and end-systolic dimensions; 2)all motion score (1 � normal, 2 � hypokinesia, 3 �

kinesia, 4 � dyskinesia, and 5 � aneurismal dyskinesis);nd 3) fractional shortening (FS) was calculated as: FS (%)

[(LVDd�)/LVDd] � 100. All measurements wereveraged for 3 cardiac cycles and performed by an experi-nced operator blinded to the treatment group.tatistical analysis. Data are expressed as mean � stan-ard error of the mean or median and interquartile range.he Kolmogorov-Smirnov procedure was applied to deter-ine whether the data were normally distributed. A general

inear model 2-way repeated-measures analysis of varianceANOVA) was used to test the hypothesis that changes ineasures of LV function over time varied among the 3

xperimental groups. The model included the effects ofreatment, time, and treatment-by-time interaction. Theonferroni correction was used to assess significance ofredefined comparisons at specific time points. Lung-tibiaatio data were not normally distributed. Therefore, groupsere compared with the nonparametric 1-way ANOVA

Kruskal-Wallis test) followed by the Mann-Whitney testith Bonferroni correction.

esults

ardiac environment does not enhance hESC cardio-yogenesis. One of the obstacles in using hESC foryocardial repair is the relatively low cardiomyocyte yield

uring spontaneous in vitro differentiation (19). Therefore,e tested the hypothesis that the existing cardiac tissue

nvironment may provide the necessary signals to augment
ardiomyocyte differentiation. To test this assumption, 726 fl
uorescently labeled early stage differentiating EBs wereocultured with primary cultures of neonatal rat ventricularyocytes (n � 5). Identical populations of early stage EBsere plated on gelatin-covered plates (spontaneous differ-

ntiation). In contrast to our initial assumption, the per-entage of EBs containing contracting areas was found to beignificantly lower (p � 0.01) in the coculturing group0.7% of the 726 EBs studied) when compared with that inhe control group (9.9% of 621 EBs).

We next evaluated the possible effects of the in vivoardiac environment on hESC differentiation. Undifferen-iated hESCs were grafted into the LV myocardium ofealthy and infarcted cyclosporine-immunosuppressed rats.istological examination, 4 weeks later, demonstrated that

ot only that the in vivo cardiac environment did notnhance hESC cardiomyogenesis (as assessed by immunos-ainings for cardiac-specific markers), but also that it re-ulted in almost no differentiation into the cardiac lineage inll hearts studied. Rather, in 6 of 10 healthy and in 3 of 6nfarcted rats, injection of undifferentiated hESCs resultedn the formation of teratoma-like structures, characterizedy the presence of advanced cell derivatives of all 3 germayers (Fig. 1A). In the rest of the hearts, either theransplanted cells could not be detected or they showed aore limited differentiation pattern, mostly to epithelial-

ike cell derivatives (data not shown).n vivo grafting of the hESC-CMs into the uninjuredeart. Because grafting of undifferentiated hESCs resulted

n the formation of teratoma-like structures, we pursued anlternative track of initial ex vivo differentiation of theESC into the cardiac lineage followed by in vivo graftingf the hESC-CMs. Using our previously described EBifferentiating system (8), we mechanically dissected theeating areas within the EBs (each containing approxi-ately 6,000 cells), dispersed them into small clusters, and

rafted them into the LV myocardium. To characterize theature of the grafted cells before transplantation, we per-ormed in vitro immunocytostaining studies of samples ofhese beating clusters (Figs. 1B and 1C). Immunostainingor undifferentiated hESC markers (Tra-1-60, Tra-1-81,nd Oct-4 [not shown]) failed to show any positivelytained cells (Fig. 1B). In contrast, 71 � 4% of the cellstained positively for TnI (Fig. 1C).

Histological evaluation, conducted at 36 h and 4 weeksfter hESC-CM transplantation revealed the presence ofhe grafted cells and lack of teratoma formation (Fig. 2). Tossure the accurate identification of the grafted cellshroughout this study, we used a number of labeling,racking, and staining techniques. These included immu-ostaining for human-specific markers (using antimito-hondrial [Fig. 2A], anti-HLA-A,B,C [Fig. 3C], andnti-Ki67 [Fig. 2D] antibodies, introduction of geneticarkers eGFP [Figs. 2B, 2C, and 2E and Figs. 3A and

B] or nLacZ [Figs. 4B and 4C]), and prelabeling with
uorescent cell tracers (Figs. 2F and 3D).

cmtentcofh

immmcoppc

(ijIodnhpToo(wOa

mt(


As shown in Figure 2, the cell grafts consisted ofonfluent masses of human cells interspersed within hostyocardium. Histological examination, at 36 h post-

ransplantation (Figs. 2A to 2D), demonstrated that thengrafted hESC-CMs were relatively small, had a highuclear/cytoplasmatic ratio, displayed an early striated pat-ern, and were arranged isotropically within the host myo-ardium. In 10 � 2% of these human cardiomyocytes (251f 2,404 cells examined in 4 hearts), we could find evidenceor cell proliferation, as assessed by coimmunostaining foruman Ki67 and TnI (Fig. 2D).Four weeks after cell transplantation, we could still

dentify the grafted hESC-CMs. Although the cells re-ained relatively immature, the cell grafts appeared to beore organized and have undergone some ultrastructuralaturation. The hESC-CMs were larger, their nucleus/

ytoplasm ratio was lower, they demonstrated a morerganized striated pattern (Fig. 2E), and they were stainedositively with anti-Ryandine antibodies (indicating theresence of sarcoplasmatic reticulum) (Fig. 2F). This was

A

BhESC colonies Contracting EB

Tra 1-60

Tra 1-81

Figure 1 In Vivo Transplantation of Undifferentiated hESC Resu

(A) Hematoxilin and eosin staining of a teratoma within the infarcted area. Note thnar epithelium (right, bottom). (bars: left 500 �m, middle 200 �m, right 100 �m(bottom). Note the positive staining of undifferentiated human embryonic stem ceing embryoid bodies (EBs) (right) (bar: 100 �m). (C) Troponin I immunostaining o

oupled with a complete withdrawal from the cell cycle I

negative Ki67 staining, data not shown). Similarly, Cx43mmunostaining demonstrated the development of gapunction by the engrafted cells (Fig. 2E).n vivo engraftment in the infarcted rat heart. The abilityf the hESC-CMs to engraft also in the infarcted myocar-ium (7 to 10 days after coronary ligation) was assessedext. Histological evaluation of the hearts injected with theESC-CMs (at 30 or 60 days) (Fig. 3) demonstrated theresence of the grafted cells and lack of teratoma formation.he grafted hESC-CMs were identified as confluent clustersf human cells that could be located occasionally in the centerf the scar (Fig. 3A) but more commonly in the border zoneFigs. 3B to 3D). The cell grafts were demonstrated to alignith host cells with no significant inflammatory encapsulation.ccasionally, gap junctions could be identified between host

nd grafted cells (Fig. 3D).We next quantified the histological information to deter-ine the relative size of the hESC-CM cell grafts within

he infarcted area. This analysis was performed in 8 heartsin each heart between 3 to 5 sections were examined).

Teratoma-Like Structures

ence of early stage hyaline cartilage (right, top) and gastrointestinal-like colum-Immunostaining for the pluripotent markers Tra-1-60 (top) and Tra-1-81C) colonies (left) and the absence of staining in cells isolated from the contract-isolated from the contracting EBs (bar: 60 �m).

Cs

lts in

e pres). (B)ll (hESf cells

nitially, the infarcted area was identified using anticollagen

awthuusvf

ttemhot

4ichsp

hIscpei(h


ntibodies. The viable cardiomyocyte area within this regionas then determined by immunostaining for TnI. Finally,

he fraction of this cardiomyocyte cellular area that was ofuman origin was quantified in coimmunostaining studiessing the aforementioned labeling and staining techniques,sed to identify the transplanted human cells. Our resultshow that 25 � 6% of the infarct area was comprised ofiable cardiac tissue of which 45 � 12% was stained positiveor human markers.

To verify and provide additional information regardinghe survival and maintenance of the cardiomyocyte pheno-ype of the grafted hESC-CMs, we developed 2 assays. Tovaluate for the presence of the human cells within the ratyocardium, we established a PCR assay for detecting

uman-specific DNA. Our results demonstrate the presencef the human cells immediately after cell grafting (1 h), and

C

A

B

R

E

Figure 2 Transplantation of hESC-CMs in the Healthy Heart

(A) The transplanted area was localized by coinjection of 2-�m fluorescent beadschondria antibodies (red) (bar: 50 �m). (B) (Left) Hematoxilin and eosin staining(hESC-CMs) (arrows). (Right) Immunostaining of the boxed area. In this example,fied as yellow cells containing both troponin I (red) and enhanced green fluorescen�m). (C) Immunostaining with antisarcomeric �-actinin (red) and antienhanced grehESC-CMs (arrows) displayed an immature, striated pattern (bar: 12 �m). (D) Asshuman Ki-67 (green) and antitroponin I (red) antibodies (bar: 10 �m). (E) (Left) I�-actinin (red) and anti-Cx43 (white) antibodies. (Right) Superposition of the immNote the relatively organized sarcomeric pattern (bar: 20 �m). (F) Immunostainingbodies. Nuclei are counterstained with To-Pro3 (blue) in all immunofluorescent ima

heir survival at 1 and 8 weeks after cell transplantation (Fig. (

A). While these PCR studies do not provide quantitativenformation regarding the size of the graft, our preliminaryalibration study demonstrated that a significant number ofuman cells (�100,000) are required for such a positiveignal. In contrast, hearts injected with saline failed to showositive PCR signal (Fig. 4A).We next performed additional experiments, in which the

ESC-derived cell grafts were transfected to express nLacZ.mmunostaining analysis of the grafted area confirmed theurvival of the nLacZ-expressing cells and their humanardiac phenotype (Fig. 4B). To establish the cardiomyocytehenotype of the engrafted cells, beyond the histologicalxamination, the nLacZ-expressing cells were carefullysolated for RNA analysis using laser microdissectionFig. 4C). Reverse transcription PCR studies using auman-specific primer sequence for the MLC-2a gene

B

D

HC

R

HC

F

), and the grafted cells were identified by immunostaining with antihuman mito-ing a cluster of grafted human embryonic stem cell-derived cardiomyocytesced green fluorescent protein-expressing hESC-CMs were used and were identi-ein (green) immunosignals. HC � human cardiomyocytes; R � rat (bar: 100orescent protein (green, right) antibodies. At this stage (36 h), the graftednt of the proliferation capacity of the hESC-CMs (36 h post-grafting) using anti-staining of the grafted area (30 days post-transplantation) using antisarcomeric

aining results with antienhanced green fluorescent protein (green) antibodies.transplanted fluorescently labeled (yellow) hESC-CMs with anti-Ryanodine anti-

(greendepictenhant proten fluessmemmunounostof theges.

shown to be robustly expressed by the hESC-CMs in

vchFtsacjst

ctwc0

cT4i


itro but not detected in the rat LV myocardium)onfirmed the expression of this gene by the transplanteduman cells (Fig. 4D).unctional assessment of cell grafting. We next assessed

he functional consequences of cell grafting. To this end,erial echocardiography measurements were performed innimals transplanted with hESC-CMs (n � 8) and wereompared with those in control groups of animals in-ected with nonmyocyte hESC derivatives (n � 9) oraline (n � 8) (Figs. 5 and 6). These studies revealed that

Figure 3 Transplantation of hESC-CMs in the Infarcted Heart

(A) Identification of the grafted hESC-CMs at the scar’s center using antienhancedbodies. (Right) Superposition of both images. The scar was identified using anticoCMs at the infarct border zone using antienhanced green fluorescent protein (greegrafted hESC-CMs with antihuman-human leukocyte antigen antibodies. (Top) Immantihuman-human leukocyte antigen (green) and antitroponin I (red) antibodies (bthe grafted cells (prelabeled with Vybrant-CFDA) (green, left) and host cardiomyocydle) (bar: 60 �m). Abbreviations as in Figure 2.

he 3 experimental groups differed with respect to 2

hanges in measures of LV function and remodeling overime (Fig. 6). The treatment-by-time interaction effectas significant for all echocardiographic parameters in-

luding FS, LVDd, and wall motion score index (all p �.001).The saline injection control group displayed a typical

ourse of post-infarction LV remodeling (Figs. 5 and 6).he FS of healthy rats (before LAD ligation) was around5% to 50%. A few days after coronary occlusion (post-njury baseline), we noted an initial decrease in FS to 20 �

fluorescent protein (green, left) and antisarcomeric �-actinin (red, middle) anti-antibodies (blue) (bar: 80 �m). (B) Immunostainings of the transplanted hESC-antitroponin I (red, right) antibodies (bar: 75 �m). (C) Identification of the

tochemistry results (bar: 100 �m). (Bottom) Immunofluorescent staining using�m). (D) Development of gap junctions (Cx43 immunostaining, white) betweenrrows). Cardiomyocytes were identified using antitroponin I antibodies (red, mid-

greenllagenn) andunohisar: 75tes (a

%. Fractional shortening continued to deteriorate to 13 �

2aid

dFti5tLnps

cr

Fhdih0f(cm

ws(i


% and 10 � 2% at 30 and 60 days, respectively (Figs. 6And 6B). This deterioration in FS and wall motion scorendex (Fig. 6C) was coupled with progressive LV diastolicilatation (Fig. 6D).Transplantation of noncardiomyocyte hESC derivatives

id not change this remodeling process (Figs. 6A to 6D).igure 5 depicts examples of serial M-mode images ob-

ained over time in the 3 groups. Note in the animalsnjected with saline and noncardiomyocytes (Figs. 5A andB) that the typical akinesis of the anterior wall persistedhroughout the study, and this was coupled with significantV dilation and functional deterioration. Consequentially,o significant changes were noted in the echocardiographicarameters in the nonmyocyte group when compared withaline-injected group (Figs. 6A to 6D).

Grafting of hESC-CMs attenuated this remodelingourse of scar expansion, LV dilation, and functional dete-

Figure 4 Long-Term Survival and Maintenance ofCardiac Phenotype of the Transplanted hESC-CMs

(A) Polymerase chain reaction-based deoxyribonucleic acid amplification usinghuman-specific primers at various time points after transplantation. (B) Coim-munostaining with anti-LacZ (red) and antitroponin I (green) antibodies (bar:100 �m). (C) The nLacZ-expressing cells (blue) were isolated using lasermicrodissection (shown before [left] and after [right] laser microdissection)(bar: 200 �m). (D) Myosin light chain-2a expression in the microdissectednLacZ-expressing cells, in a remote left ventricular (LV) site, and in a saline-injected heart. Contracting embryoid bodies served as positive controls. d �

days; hESC-CMs � human embryonic stem cell-derived cardiomyocytes.

ioration (Figs. 5 and 6). Note in the M-mode examples in

igure 5, the lack of LV dilation at 30 and 60 days in theESC-CM transplanted animal. Consequentially, at 60ays, hESC-CM cell grafting was associated with a signif-cant improvement in all measures of LV function includingigher FS (Bonferroni-adjusted for 3 comparisons, p �.001), lower wall motion score index (Bonferroni-adjustedor 3 comparisons, p � 0.008), and lower LVDdBonferroni-adjusted for 3 comparisons, p � 0.03) whenompared with animals injected with saline or noncardio-yocyte hESC derivatives (Figs. 6B to 6D).The positive effect of hESC-CMs on LV functional status

as translated to a clinical benefit. This was manifested by aignificant reduction (p � 0.05) in the lung/tibia weight ratio0.39 � 0.03 g/cm) when compared with that in the saline-njected group (0.67 � 0.11 g/cm, Bonferroni-adjusted for 3

Baseline 30 days 60 days

Salin

eN

on C

ardi

ac H

uman

E

SC D

eriv

ativ

esH

uman

ESC

Der

ived

C

ardi

omyo

cyte

s

Figure 5 Functional Results After Cell Transplantation

(Top and middle) M-mode echocardiographic images demonstrating post-infarction remodeling with left ventricular dilatation and functional deteriorationin the saline- and nonmyocyte transplantation groups, respectively. (Bottom)Similar images in the human embryonic stem cell (ESC)-derived cardiomyocytegrafting group. Note the absence of significant left ventricular dilatation(bar: 0.5 cm).

crcsba

D

Imt


omparisons, p � 0.04) (Fig. 6E). The latter parameterepresents a quantitative measure for the degree of pulmonaryongestion. Assessment of blood pressure, heart rate, and bodyurface electrocardiogram did not reveal significant changesetween the groups, and we did not note any sustained

Fractional Shortening (Means)

Time Following Transplantation

Fra

ctio

nal S

hort

enin

g(%

)

0

10

20

30

40

Post MI Baseline

30 days 60 days

Human ESC derived cardiomyocytes


Fra

ctio

nal S

hort

enin

g (%

)

0

10

20

30

40

Non Cardia

Time Fol

Fra

ctio

nal S

hort

enin

g (%

)

0

10

20

30

40

Post MIBaseline

30 days 60 days Post MBaseli

B

A

†

‡

D Left Ventricle Diastolic Diameter


Dia

met

er (

cm)

0.70

0.75

0.80

0.85

0.90

0.95

1.00

1.05

1.10

Human ES Derived CardiomyocytesNon Cardiac Human ES derivativesSaline

**

Post MIBaseline

30 days 60 days

Figure 6 Functional Assessment of Cell Grafting

(A) Changes in fractional shortening in individual animals before (post-injury baselshortening (B), wall motion score index (C), and left ventricular end-diastolic diam(green), nonmyocyte (purple), and saline-injection (red) groups. (E) Comparison oBonferroni-adjusted for 3 comparisons: *p � 0.05; †p � 0.01; ‡p � 0.005. MI �

rrhythmias in the transplanted animals. n

iscussion

n this study, we assessed the potential role of hESC foryocardial repair. The main findings of the study are: 1)

he adult and neonatal cardiac tissue does not provide the

n ESC derivatives

ransplantation

Saline


Fra

ctio

nal S

hort

enin

g (%

)

0

10

20

30

40

ys 60 days Post MI Baseline

30 days 60 days

Wall Motion Score Index (Means)


Wal

l Mot

ion

Scor

e In

dex

1.0

1.5

2.0

2.5

3.0

3.5

Post MIBaseline

30 days 60 days

†*

Lung Weight to Tibia Length Ratio

Lun

g / T

ibia

(G

r/C

m)

0.0

0.2

0.4

0.6

0.8

1.0

Human ESC Derived

Cardiomyocytes

Non CardiacHuman ESCDerivatives

Saline

*

d 30 and 60 days after cell grafting. (B to D) Changes in the average fractional) values in the human embryonic stem cell (ESC)-derived cardiomyocytesng weight to tibia length ratio between the groups. The p values were

ardial infarction.

c Huma

lowing T

Ine

30 da

C

E

ine) aneter (Df the lumyoc

ecessary environment to guide the differentiation of

hutvirptrma

ieacTecatreelt

um(tac

dctdatswwactt

rorcita

oi

pCotdiftaan

po(ms(cmtapampalacpease

RnE

R


ESC into the cardiac lineage; moreover, grafting ofndifferentiated hESCs into the in vivo rat heart resulted inhe formation of teratoma-like structures; 2) grafting of exivo pre-differentiated hESC-CMs into the uninjured andnfarcted heart did not result in the generation of teratomas,ather the engrafted cardiomyocytes were shown to survive,roliferate, and integrate with host cardiac tissue; and 3)ransplantation of hESC-CMs can favorably affect theemodeling process and even improve myocardial perfor-ance in this rat model of permanent coronary occlusion

nd extensive myocardial infarction.Tissue microenvironment is thought to play a major role

n stem cell differentiation toward specific lineages. How-ver, in contrast to our initial hypothesis, the neonatal anddult rat cardiac environment did not provide the necessaryues for hESC cardiomyogenesis both in vitro and in vivo.hese results may suggest that a much earlier embryonic

nvironment may be required for directing embryonic stemells into the cardiac lineage. Experimental evidence from

number of developmental model organisms suggestshat the primitive visceral endoderm may play a pivotalole in promoting cardiac fate induction in the adjacentarly mesoderm (20). This concept was further strength-ned by a recent study showing that a visceral endoderm-ike cell line (END-2) can promote the cardiac differen-iation of hESC (9).

Importantly, the fact that intramyocardial grafting ofndifferentiated hESC resulted in the formation of terato-as, as also shown in the mouse embryonic stem cell model

21), underscores the need to establish cardiomyocyte selec-ion protocols (22,23) or other strategies (3) that wouldscertain the lack of any undifferentiated hESCs in futurelinical cell transplantation procedures.

Because loss of cardiomyocytes plays a critical role in theevelopment of heart failure, cardiomyogenesis has been theentral aim of cardiac regenerative therapy (1). Our long-erm histological data as well as the results of the assayseveloped to detect human DNA within the rat heart and tossess the continuous expression of cardiac-specific genes byhe grafted cells demonstrate the long-term engraftment,urvival, and integration of the hESC-CMs. Interestingly,e noted superior engraftment results at the scar’s peripheryhen compared with its center. This finding may be

ttributed to the lower initial number of injected cells at theenter, to the greater technical difficulty in injecting cells inhis much thinner region, and to the improved vasculariza-ion of the border zone.

Grafting of hESC-CMs ameliorated the typical infarctemodeling process of LV expansion and functional deteri-ration. Similar attenuation of the remodeling process inodents was previously reported after delivery of a variety ofell types (1,2). The mechanism underlying the functionalmprovement in these studies may not necessarily involvehe generation of new contractile elements (cardiomyocytes)
nd may be related to changes in the mechanical properties
f the scar, recruitment of endogenous stem cells, andnduction of angiogenesis.

While all of the aforementioned mechanisms probablylay a role in the functional benefit observed after hESC-Ms transplantation, it is also possible that some of thebserved effect may stem from direct contribution to con-ractility by the transplanted cardiomyocytes. Although ourata do not provide direct evidence for such a mechanism, its supported by a number of findings: 1) the structural andunctional cardiomyocyte phenotype of the grafted cells; 2)heir ability to functionally integrate with host cardiac tissues previously demonstrated both in vitro and in vivo (16,17);nd 3) the lack of significant improvement after grafting ofoncardiomyocyte hESC derivatives.Despite these encouraging results and the enormous

otential of the hESC-CMs, several obstacles need to bevercome before this strategy can become a clinical reality19). These include the need to generate a directed andore efficient differentiating system, the need to establish

election protocols to derive pure population of cardiac cells22,23), and the need to scale up the entire process to derivelinically relevant number of cells. In addition, to achieveaximal contractile benefit by the grafted cells, the rela-

ively “immature” hESC-CMs should undergo structuralnd functional maturation towards an adult-like ventricularhenotype. This issue is also important to reduce the risk forrrhythmias. A beginning of such structural and functionalaturation process was already noted in vitro, during

rolonged hESC-CM culturing (12,24), and when cultureds 3-dimensional engineered tissue constructs (25). Simi-arly, the long-term engraftment studies, presented here,lso demonstrated some structural maturation of the graftedells. Finally, strategies to counter immune rejection wouldrobably be required (26). These strategies may includestablishing “banks” of major histocompatibility complexntigen-typed hESCs, genetically modifying the hESCs touppress the expected immune response, induction of tol-rance, and somatic nuclear transfer.

eprint requests and correspondence: Dr. Lior Gepstein, Tech-ion’s Faculty of Medicine, P.O. Box 9649, Haifa, 31096, Israel.-mail: [email protected].

EFERENCES

1. Laflamme MA, Murray CE. Regenerating the heart. Nat Biotechnol2005;23:845–56.

2. Dimmeler S, Zeiher AM, Schneider MD. Unchain my heart: thescientific foundations of cardiac repair. J Clin Invest 2005;115:572–83.

3. Menard C, Hagege AA, Agbulut O, et al. Transplantation ofcardiac-committed mouse embryonic stem cells to infarcted sheepmyocardium: a preclinical study. Lancet 2005;366:1005–12.

4. Li RK, Mickle DA, Weisel RD, et al. Natural history of fetal ratcardiomyocytes transplanted into adult rat myocardial scar tissue.Circulation 1997;96:II179–86; discussion 186–7.

5. Etzion S, Battler A, Barbash IM, et al. Influence of embryoniccardiomyocyte transplantation on the progression of heart failure in a
rat model of extensive myocardial infarction. J Mol Cell Cardiol2001;33:1321–30.

1

1

1

1

1

1

1

1

1

1

2

2

2

2

2

2

2


6. Muller-Ehmsen J, Peterson KL, Kedes L, et al. Rebuilding a damagedheart: long-term survival of transplanted neonatal rat cardiomyocytesafter myocardial infarction and effect on cardiac function. Circulation2002;105:1720–6.

7. Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al. Embryonic stem celllines derived from human blastocysts. Science 1998;282:1145–7.

8. Kehat I, Kenyagin-Karsenti D, Snir M, et al. Human embryonic stemcells can differentiate into myocytes with structural and functionalproperties of cardiomyocytes. J Clin Invest 2001;108:407–14.

9. Mummery C, Ward-van Oostwaard D, Doevendans P, et al. Differ-entiation of human embryonic stem cells to cardiomyocytes: role ofcoculture with visceral endoderm-like cells. Circulation 2003;107:2733–40.

0. Xu C, Police S, Rao N, Carpenter MK. Characterization andenrichment of cardiomyocytes derived from human embryonic stemcells. Circ Res 2002;91:501–8.

1. He JQ, Ma Y, Lee Y, Thomson JA, Kamp TJ. Human embryonicstem cells develop into multiple types of cardiac myocytes: actionpotential characterization. Circ Res 2003;93:32–9.

2. Snir M, Kehat I, Gepstein A, et al. Assessment of the ultrastructuraland proliferative properties of human embryonic stem cell-derivedcardiomyocytes. Am J Physiol Heart Circ Physiol 2003;285:H2355–63.

3. Satin J, Kehat I, Caspi O, et al. Mechanism of spontaneous excitabilityin human embryonic stem cell derived cardiomyocytes. J Physiol2004;559:479–96.

4. Kehat I, Gepstein A, Spira A, Itskovitz-Eldor J, Gepstein L. High-resolution electrophysiological assessment of human embryonic stemcell-derived cardiomyocytes: a novel in vitro model for the study ofconduction. Circ Res 2002;91:659–61.

5. Laflamme MA, Gold J, Xu C, et al. Formation of human myocardiumin the rat heart from human embryonic stem cells. Am J Pathol
2005;167:663–71.
6. Kehat I, Khimovich L, Caspi O, et al. Electromechanical integrationof cardiomyocytes derived from human embryonic stem cells. NatBiotechnol 2004;22:1282–9.

7. Xue T, Cho HC, Akar FG, et al. Functional integration of electricallyactive cardiac derivatives from genetically engineered human embry-onic stem cells with quiescent recipient ventricular cardiomyocytes:insights into the development of cell-based pacemakers. Circulation2005;111:11–20.

8. Itskovitz-Eldor J, Schuldiner M, Karsenti D, et al. Differentiation ofhuman embryonic stem cells into embryoid bodies compromising thethree embryonic germ layers. Mol Med 2000;6:88–95.

9. Gepstein L. Derivation and potential applications of human embry-onic stem cells. Circ Res 2002;91:866–76.

0. Nascone N, Mercola M. An inductive role for the endoderm inXenopus cardiogenesis. Development 1995;121:515–23.

1. Nussbaum J, Minami E, Laflamme MA, et al. Transplantation ofundifferentiated murine embryonic stem cells in the heart: teratomaformation and immune response. FASEB J 2007;21:1345–57.

2. Huber I, Itzhaki I, Caspi O, et al. Identification and selection ofcardiomyocytes during human embryonic stem cell differentiation.FASEB J 2007;21:2551–63.

3. Klug MG, Soonpaa MH, Koh GY, Field LJ. Genetically selectedcardiomyocytes from differentiating embronic stem cells form stableintracardiac grafts. J Clin Invest 1996;98:216–24.

4. Sartiani L, Bettiol E, Stillitano F, Mugelli A, Cerbai E, Jaconi ME.Developmental changes in cardiomyocytes differentiated from humanembryonic stem cells: a molecular and electrophysiological approach.Stem Cells 2007;25:1136–44.

5. Caspi O, Lesman A, Basevitch Y, et al. Tissue engineering ofvascularized cardiac muscle from human embryonic stem cells. CircRes 2007;100:263–72.

6. Bradley JA, Bolton EM, Pedersen RA. Stem cell medicine encounters
the immune system. Nat Rev Immunol 2002;2:859–71.

transplantation of human embryonic stem cell-derived ... · we injected 3 (larova, teltow,...

Documents