Cell 127, December 15, 2006 ©2006 Elsevier Inc. 1101
Congenital and acquired heart diseases lead to abnor-
malities in cardiac growth, contractility, and vascularity
and represent the principle noninfectious cause of human
morbidity and mortality. The inability of postnatal car-
diac muscle cells to proliferate poses a major barrier to
the functional restoration of the diseased heart. This has
spawned intense interest—and equally intense contro-
versy—regarding the therapeutic potential of cardiac stem
cells. Regenerative stem cell therapies for heart disease
necessitate an understanding of the molecular mecha-
nisms that govern the fates, differentiation, and morpho-
genesis of the myriad cell types that comprise the heart
and an ability to manipulate these mechanisms. Three
new studies provide evidence for cardiac progenitor cells
that have the potential to differentiate into all three of the
major cell types of the heart: cardiac myocytes, smooth
muscle cells, and endothelial cells (Kattman et al., 2006;
Moretti et al., 2006; Wu et al., 2006). These studies begin to
address the plasticity and developmental potential of car-
diac progenitor cells. At the same time, they raise interest-
ing questions about the embryonic origins of cardiac cell
lineages and the criteria used to define stem cells and their
descendants. Here, we attempt to place these recent stud-
ies in perspective with the complex and often contentious
conclusions in the field and to highlight important issues to
be resolved in the future.
The Embryonic Origins of Cardiac Cell Lineages
Formation of the mature multichambered heart requires
contributions of diverse cell types with specialized func-
tions. Cardiac muscle cells become specialized to form
ventricular and atrial myocytes, as well as cells of the con-
duction system. Smooth muscle cells form the venous
and arterial vasculature, endothelial cells give rise to the
endocardium and valves, and the epicardium serves as a
source of precursors for the coronary vasculature.
Recent studies have demonstrated that the heart
forms from two separate progenitor cell populations or
“heart fields” that segregate from a common progenitor
at gastrulation (Kelly et al., 2001; Cai et al., 2003; Abu-
Issa et al., 2004). The earliest population of cardiac
progenitors, referred to as the primary heart field, origi-
nates in the anterior splanchnic mesoderm and gives
rise to the cardiac crescent, which ultimately contrib-
utes to the left ventricle and atria (Figure 1). Cells from
the cardiac crescent migrate medially and form the lin-
ear heart tube, which consists of an inner endocardial
layer and an outer myocardial layer. Rightward loop-
ing and differential growth along the outer curvatures
of the heart tube ultimately generates the multicham-
bered heart (reviewed in Buckingham et al., 2005). An
additional source of cardiac precursors, referred to as
the secondary or anterior heart field, is derived from
the pharyngeal mesoderm located medial to the car-
diac crescent (Figure 1). Lineage tracing experiments
have shown that the secondary heart field contributes
primarily to the right ventricle and outflow tract (Kelly et
al., 2001; Cai et al., 2003; Laugwitz et al., 2005).
The primary and secondary heart fields can be distin-
guished by the expression of specific transcription factors
and signaling molecules. The T-box transcription factor
Tbx5 and the bHLH transcription factor Hand1, for exam-
ple, mark the primary heart field, whereas Hand2, the
LIM-homeodomain transcription factor Isl1, and Fgf10,
among others, mark the secondary heart field (Kelly et
al., 2001; Cai et al., 2003). Mouse embryos with muta-
tions in these and other genes display abnormalities in the
corresponding regions of the heart (Buckingham et al.,
2005). Other cardiac regulatory genes, such as the home-
obox gene Nkx2-5, are expressed in both heart fields and
rely on different cis-regulatory elements for expression
in these progenitor populations, demonstrating that the
A Common Progenitor at the Heart of
Daniel J. Garry1,2,3 and Eric N. Olson2,3
1Department of Internal Medicine
2Department of Molecular Biology
3Donald W. Reynolds Cardiovascular Clinical Research Center
University of Texas Southwestern Medical Center, Dallas, TX 75390, USA
Formation of the heart requires the coordinated functions of cardiac myocytes, smooth
muscle cells, endothelial cells, and connective tissue elements. Several recent studies now
reveal that these different cell types arise from a common progenitor (Kattman et al., 2006;
Moretti et al., 2006; Wu et al., 2006). These findings raise interesting questions about the
lineage relationships of cardiovascular progenitor cell populations and suggest possibili-
ties for cardiac repair in both congenital and acquired heart disease.
1102 Cell 127, December 15, 2006 ©2006 Elsevier Inc.
transcriptional networks that drive cardiogenesis in differ-
ent regions of the developing heart are distinct (reviewed
in Schwartz and Olson, 1999).
Clonality, Self Renewal, and Multipotentiality
Several studies, including recent papers in Cell and
Developmental Cell, describe multipotent mesodermal
progenitors that give rise to diverse cells types within the
heart (Kattman et al., 2006; Moretti et al., 2006; Wu et
al., 2006). Before considering these results, it is worth-
while to consider the defining properties of stem cells:
clonality, self renewal, and multipotentiality. Embryonic
stem (ES) cells derived from the blastocyst represent the
prototypic stem cell population capable of parenting the
more than 200 somatic cell types. As development pro-
ceeds, stem cell progeny become progressively special-
ized and restricted in their developmental potential.
Lineage tracing with reporter genes is commonly
used to demonstrate that a progenitor population can
give rise to different cell types in the embryo. However,
only a clonal analysis can show unequivocally that a
single cell is capable of generating progeny of multi-
ple lineages. There have been numerous recent claims
regarding the existence of cardiac stem cells with
remarkably different characteristics. Part of the con-
fusion derives from the fact that, in many cases, the
“stem cell” populations are almost certainly heteroge-
neous populations. The confounding
and often contradictory properties of
different cardiac stem cells raise ques-
tions about the criteria used to define
them. These controversies illustrate
the need for careful documentation of
clonality, self renewal and the extent
of multipotentiality in studies of car-
diac stem cells.
Multipotent Cardiac Progenitors
Numerous studies have provided evi-
dence for multipotent progenitor cells
that give rise to the different cell types
of the heart (Ema et al., 2006). Such
multipotentiality of muscle progenitor
populations is emerging as a common
theme, as demonstrated by the genesis of skeletal mus-
cle and aortic smooth muscle cells from a common pro-
genitor (Esner et al., 2006).
Perhaps the best example of a multipotent progeni-
tor is the hemangioblast—a mesodermal progenitor of
hematopoietic and vascular lineages marked by the
expression of brachyury (Bry) and the VEGF receptor
Flk1. Evidence for the existence of hemangioblasts capa-
ble of parenting both endothelial and hematopoietic cells
was recently obtained using a mouse ES cell/embryoid
body differentiation assay (Huber et al., 2004) and a fate
mapping strategy in the zebrafish (Vogeli et al., 2006).
Analysis of Flk1-lacZ knockin embryos has also revealed
that Flk1 is expressed in progenitors with the potential to
contribute to a broad range of mesodermal derivatives,
including cells migrating from the primitive streak and
progenitors located in the cardiac crescent, somites,
and extraembryonic mesoderm (Ema et al., 2006). Flk1
is indispensable for the formation of endothelial and
hematopoietic lineages, but it is not absolutely required
for the development of other mesodermal lineages such
as the cardiac muscle lineage (Ema et al., 2006).
It is particularly interesting that multipotent cardiovas-
cular stem cells can also be isolated from the adult heart.
Beltrami et al. (2003) have described Lin−/ckit+ cells iso-
lated from the adult rat heart that display the properties
of stem cells and are able to give rise to cardiac myo-
Figure 1. Cardiac Stem Cells
The heart forms from two heart fields. Scanning
electron micrographs of representative stages of
mouse development are shown with derivatives of
the primary and secondary heart fields shown in
color (left column; adapted from Kaufman, 1992).
The individual panels show the characteristics and
lineage relationships of cardiac progenitor cells as
described in Kattman et al. (2006) (top), and Moret-
ti et al. (2006) (middle), Wu et al. (2006) (bottom).
Abbreviations designate the following: a, atrium;
lv, left ventricle; oft, outflow tract; rv, right ventri-
cle, CV-CFC, cardiovascular colony forming cells;
MICPs, multipotent isl1+ cardiac progenitors.
Cell 127, December 15, 2006 ©2006 Elsevier Inc. 1103
cytes, smooth muscle cells, and endothelial cells when
cultured in vitro. Moreover, injection of these cells into
the infarcted, ischemic heart improves myocardial func-
tion. Whether these cells represent remnants of embry-
onic cardiogenic precursors or a separate cell popula-
tion remains to be determined. The observation that
cardiospheres (spherical, cellular aggregates formed
from a clonal cell population of unknown origin) can be
isolated and exponentially expanded from adult human
and mouse hearts further illustrates the proliferative and
developmental potential of adult cardiac progenitors
(Messina et al., 2004).
Using ES cells bearing a green fluorescent protein
(GFP) reporter targeted to the Bry gene, Kattman et al.
(2006) showed that a Flk1−/GFP-Bry+ progenitor can give
rise sequentially, in culture, to two distinct populations of
Flk1+/GFP-Bry+ progenitors. The first corresponds to the
hemangioblast, which parents endothelial and hemat-
opoietic lineages, and the second is capable of generat-
ing cardiac myocytes, vascular smooth muscle cells, and
endothelial cells (Figure 1). The latter cell, which is the
cardiovascular equivalent of the hemangioblast (referred
to as a cardiovascular colony-forming cell), expresses
Flk1 prior to cardiac markers, such as Nkx2-5, GATA4,
and MEF2C. Cardiovascular colony-forming cells appear
to be capable of generating both the primary and second-
ary heart fields, as shown by the nearly mutually exclusive
expression of Tbx5 and Isl1, respectively, in clonal colo-
nies. These findings are consistent with conclusions that
the two heart fields develop from a common progenitor
with endothelial potential and segregate early in develop-
ment (Meilhac et al., 2004).
Cardiovascular colony-forming cells could also be iso-
lated from head-fold stage mouse embryos and clonally
expanded in culture to generate colonies that expressed
endothelial, vascular smooth muscle, and cardiac mark-
ers. This study, together with previous studies (Huber
et al., 2004; Ema et al., 2006), collectively support the
hypothesis that Flk1 expression identifies multilineage
mesodermal progenitors that can acquire a cardiac mus-
cle fate. A future challenge will be to determine whether
such Flk1+ progenitors are fully committed to cardio-
vascular lineages during embryogenesis or whether
they possess broader developmental capacity following
injury or when placed in a permissive environment.
Using transgenic mice bearing an enhanced yel-
low fluorescent protein (eYFP) reporter controlled by
the Nkx2-5 enhancer as an early cardiac marker, we
isolated cardiac progenitors from cardiac crescent,
heart tube, and looped heart of mouse embryos by
flow cytometry (Masino et al., 2004). Transcriptome
analysis revealed stage-specific molecular signatures
of cardiac progenitors from each of these develop-
mental periods (Masino et al., 2004). Notably, cardiac
progenitors isolated from the cardiac crescent were
enriched in transcripts commonly expressed in car-
diac, endothelial, and hematopoietic lineages (Masino
et al., 2004).
Wu et al. (2006) used a similar approach to isolate
cardiac progenitors from mouse embryos and differenti-
ated ES cells expressing Nkx2-5-GFP. Nkx2-5-GFP+cells
obtained from ES cultures displayed high proliferative
capacity and were capable of generating clonal colonies
with cardiac and smooth muscle phenotypes in vitro.
These cells expressed modest levels of the stem cell mark-
ers ckit and Sca1 but did not express endothelial markers,
suggesting that endothelial and myogenic lineages had
already segregated by the time the Nkx2-5-GFP transgene
was activated. Nkx2-5-GFP+ cells isolated from hearts of
E9.5 mouse embryos were also able to generate cardiac
and smooth muscle lineages when allowed to differenti-
ate en mass in vitro. Under these conditions, Nkx2-5-GFP
expression segregated with cardiac markers, whereas
smooth muscle markers were restricted to GFP− cells,
suggesting that maintenance of Nkx2-5 expression char-
acterizes the cardiac phenotype. Attempts to isolate and
expand a clonal Nkx2-5-GFP+/ckit+ cell population from
the embryo were unsuccessful, as were efforts to direct
Nkx2-5-GFP+/ckit+ cells to other fates (i.e., neural, skel-
etal muscle, hematopoietic lineages), which leaves open
the question of whether multipotentiality is unique to ES-
derived progenitors. Thus, Nkx2-5-GFP+/ckit+ bipotential
cardiovascular progenitors appear to be more restricted
in developmental potential than the progenitors of cardio-
vascular colony-forming cells described by Kattman et al.
(2006). Consistent with this notion, the progenitors of car-
diovascular-colony forming cells did not express Nkx2-5
or other cardiac markers, suggesting they exist earlier in a
developmental pathway than the Nkx2-5-GFP+/ckit+ pro-
genitors (Kattman et al., 2006).
Additional evidence for multipotent embryonic car-
diac progenitors emerges from the studies of Laugwitz et
al. (2005) and Moretti et al. (2006), who showed by line-
age tracing that Isl1+ cells, referred to as multipotent isl1+
cardiac progenitors (MICPs), contribute to endothelial,
endocardial, smooth muscle, conduction system, right
ventricular, and atrial myogenic lineages (Figure 1). Using
flow cytometry, Isl1 expressing cells were isolated from the
developing heart tube of E8.0–8.5 mouse embryos and,
following expansion in a mesenchymal feeder layer sys-
tem, were observed to spontaneously express endothelial,
cardiac, and smooth muscle markers. ES cell-derived Isl1+
cardiac progenitors marked by expression of GFP from the
Nkx2-5 locus were also able to give rise to more restricted
progenitor cell populations that coexpressed Isl1 and
Nkx2-5 (displaying cardiac and smooth muscle potential)
or Isl1 and Flk1 (displaying endothelial potential).
Moretti et al. (2006) liken the multipotent isl1+ cardiac
progenitor to the hematopoietic stem cell, which is able
to reconstitute all blood cell lineages in vivo. Although this
is an interesting analogy, the inability, thus far, to clonally
generate the full spectrum of cardiac cells from a single
Isl1+progenitor in vivo, the restriction of Isl1 expression to
the secondary heart field, and the existence of numerous
other progenitor populations that contribute to the heart,
suggest that heart development is not so simple.
1104 Cell 127, December 15, 2006 ©2006 Elsevier Inc.
The realization that diverse cell types of the heart can
arise from common progenitors raises numerous ques-
tions for the future. For example, what dictates the deci-
sion of cardiovascular progenitors to adopt one fate ver-
sus another? Bone morphogenetic proteins (BMPs) and
Wnts play positive and negative roles in specification of
cardiac cell fates. How these and other signals impinge
on the transcriptional machinery that directs cardiac
muscle, smooth muscle, and endothelial cell fates
remains to be determined. The relative contributions of
various progenitor populations (e.g., primary and sec-
ondary heart fields, epicardium, etc.) to the diverse cell
types of the heart also remain to be resolved.
Why do the cardiac progenitors described by different
groups display different characteristics, and what are the
relationships between different multipotent cardiac pro-
genitors? Given that the path from a progenitor to a fully
differentiated cell type, be it cardiac myocyte, smooth
muscle, or endothelial, represents a continuum of develop-
ment, it is likely that numerous variables, such as the devel-
opmental time of isolation, the anatomical source of cells,
as well as culture conditions affect the phenotypes of cells
along the pathway. In addition, different markers used for
isolation and identification of progenitor cells undoubtedly
select for specific subsets of cells. Cardiac progenitors iso-
lated from the adult heart also display distinct and overlap-
ping traits based on their multilineage diversification, cell
surface markers, proliferative capacity, and location within
the heart (Beltrami et al., 2003; Oh et al., 2003; Martin et
al., 2004). Limited knowledge exists regarding the relation-
ships of these progenitors to one another or the molecular
networks that regulate their proliferation, self renewal, or
lineage differentiation states. Studies of isolated stem cells
that use ex vivo marker expression data to define plastic-
ity and fate potential of putative cardiac progenitors tell us
what is possible in a tissue culture dish but not necessar-
ily what actually happens in vivo during development and
disease. This issue is especially relevant to studies in which
putative stem cells are being used in humans to treat myo-
cardial disease. There is clearly room for strengthening the
link between basic science and clinical studies.
The propensity of ES-derived cells to generate ter-
atomas in vivo limits their use in human regenerative
therapies. Multipotent cardiac progenitors, therefore,
represent a potentially attractive alternative for cardiac
repair. Clinical trials are already underway using hemat-
opoietic and mesenchymal stem cells for restoration of
function in the infarcted and failing heart, and initial suc-
cesses have been touted. Nevertheless, numerous hur-
dles remain, and caution is warranted. The timing of cell
delivery following injury, the identity and number of cells
delivered, the mode of delivery and the long-term ben-
efits are issues that remain to be addressed. Moreover,
it is unclear whether cell-based therapies could limit the
progression of heart failure or whether autologous stem
cell populations are capable of transdifferentiating to a
cardiac fate and repopulating the injured human heart.
Further analysis of embryonic as well as adult cardiac
progenitors with respect to their derivation, molecular
regulation, and relationships to one another promises to
enhance our understanding of cardiac development and
disease and may serve as a platform for developing ther-
apies to treat congenital and acquired heart disease.
The authors are funded by the NIH and the Donald W. Reynolds Foun-
dation. D.J.G. is an Established Investigator of the American Heart
Abu-Issa, R., Waldo, K., and Kirby, M.L. (2004). Dev. Biol. 272, 281–285.
Beltrami, A.P., Barlcchi, L., Torella, D., Baker, M., Limana, F., Chimenti,
S., Kasahara, H., Rota, M., Musso, E., Urbanek, K., et al. (2003). Cell
Buckingham, M., Meilhac, S., and Zaffran, S. (2005). Nat. Rev. Genet.
Cai, C.-L., Liang, X., Shi, Y., Chu, P.-H., Pfaff, S.L., Chen, J., and Evans,
S. (2003). Dev. Cell 5, 877–889.
Ema, M., Takahashi, S., and Rossant, J. (2006). Blood 107, 111–117.
Esner, M., Meilhac, S.M., Relaix, F., Nicolas, J.-F., Cossu, G., and Buck-
ingham, M.E. (2006). Development 133, 737–749.
Huber, T.L., Kouskoff, V., Fehling, H.J., Palis, J., and Keller, G. (2004).
Nature 432, 625–630.
Kattman, S.J., Huber, T.L., and Keller, G.M. (2006). Dev. Cell 11, 723–
Kaufman, M.H. (1992). The atlas of mouse development. pp. 442. Aca-
demic Press, London.
Kelly, R.G., Brown, N.A., and Buckingham, M.E. (2001). Dev. Cell 1,
Laugwitz, K.-L., Moretti, A., Lam, J., Gruber, P., Chen, Y., Woodard, S., Lin,
L.-Z., Cai, C.-L., Lu, M.M., Reth, M., et al. (2005). Nature 433, 647–653.
Martin, C.M., Meeson, A., Robertson, S.M., Hawke, T.J., Richardson,
J.A., Bates, S., Goetsch, S.C., Gallardo, T.D., and Garry, D.J. (2004). Dev.
Biol. 265, 262–275.
Masino, A.M., Gallardo, T.D., Wilcox, C.A., Olson, E.N., Williams, R.S.,
and Garry, D.J. (2004). Circ. Res. 95, 389–397.
Meilhac, S.M., Esner, M., Kelly, R.G., Nicolas, J.F., and Buckingham, M.E.
(2004). Dev. Cell 6, 685–698.
Messina, E., De Angelis, L., Frati, G., Morrone, S., Chimenti, S., Fiordal-
iso, F., Salio, M., Battaglia, M., Latronico, M.V.G., Coletta, M., et al. (2004).
Circ. Res. 95, 911–921.
Moretti, A., Caron, L., Nakano, A., Lam, J.T., Bernshausen, A., Chen, Y.,
Qyang, Y., Bu, L., Sasaki, M., Yunfu, S., et al. (2006). Cell, this issue.
Oh, H., Bradfute, S.B., Gallardo, T.D., Nakamura, T., Gaussin, V., Mishina,
Y., Pocius, J., Michael, L.H., Behringer, R.R., Garry, D.J., et al. (2003).
Proc. Natl. Acad. Sci. USA 100, 12313–12318.
Schwartz, R.J., and Olson, E.N. (1999). Development 126, 4187–4192.
Vogeli, K.M., Jin, S.-W., Martin, G.R., and Stainier, D.Y.R. (2006). Nature
Wu, S.M., Fujiwara, Y., Cibulsky, S.M., Clapham, D.E., Lien, C.-L., Schul-
theiss, T.M., and Orkin, S.H. (2006). Cell, this issue.