| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Physiological Reviews, Vol. 83, No. 1, January 2003, pp. 59-115; 10.1152/physrev.00017.2002.
Copyright ©2003 by the American Physiological Society
University of Antwerp, Antwerp, Belgium
I. INTRODUCTION
II. ENDOTHELIAL ORGANIZATION IN THE HEART
A. Cardiac Endothelium Versus Coronary Vascular Endothelium
B. Cardiac Endothelial Morphology
III. CARDIAC ENDOTHELIAL-MYOCARDIAL INTERACTION
A. Cardiac Growth
B. Cardiac Contractile Performance
C. Cardiac Rhythmicity
IV. CARDIAC ENDOTHELIAL DYSFUNCTION: ROLE IN THE PATHOGENESIS OF CARDIAC FAILURE
A. Endothelial Activation and Dysfunction
B. Peripheral Vascular Endothelial Dysfunction in Cardiac Failure
C. Coronary Endothelial Dysfunction in Cardiac Failure
D. Cardiac Endothelial Dysfunction in Cardiac Failure
E. Role of NO
F. Role of ET
G. Role of Other Myocardial-Endothelial Signaling Pathways
V. THE ENDOTHELIAL SYSTEM: CONJECTURAL ROLE OF CARDIAC ENDOTHELIUM
VI. SUMMARY
 |
ABSTRACT |
|---|
 Top NextReferences
|
|---|
Brutsaert, Dirk L.
Cardiac Endothelial-Myocardial Signaling: Its Role in
Cardiac Growth, Contractile Performance, and Rhythmicity. Physiol. Rev. 83: 59-115, 2003; 10.1152/physrev.00017.2002.
Experimental work during the past 15 years
has demonstrated that endothelial cells in the heart play an obligatory
role in regulating and maintaining cardiac function, in particular, at the endocardium and in the myocardial capillaries where endothelial cells directly interact with adjacent cardiomyocytes. The emerging field of targeted gene manipulation has led to the contention that
cardiac endothelial-cardiomyocytal interaction is a prerequisite for normal cardiac development and growth. Some of the
molecular mechanisms and cellular signals governing this interaction,
such as neuregulin, vascular endothelial growth factor, and
angiopoietin, continue to maintain phenotype and survival of
cardiomyocytes in the adult heart. Cardiac endothelial cells, like
vascular endothelial cells, also express and release a variety of auto-
and paracrine agents, such as nitric oxide, endothelin, prostaglandin
I2, and angiotensin II, which directly influence cardiac
metabolism, growth, contractile performance, and rhythmicity of the
adult heart. The synthesis, secretion, and, most importantly, the
activities of these endothelium-derived substances in the heart are
closely linked, interrelated, and interactive. It may therefore be
simplistic to try and define their properties independently from one
another. Moreover, in relation specifically to the endocardial
endothelium, an active transendothelial physicochemical gradient for
various ions, or blood-heart barrier, has been demonstrated.
Linkage of this blood-heart barrier to the various other
endothelium-mediated signaling pathways or to the putative vascular
endothelium-derived hyperpolarizing factors remains to be
determined. At the early stages of cardiac failure, all major
cardiovascular risk factors may cause cardiac endothelial activation as
an adaptive response often followed by cardiac endothelial dysfunction.
Because of the interdependency of all endothelial signaling pathways,
activation or disturbance of any will necessarily affect the others
leading to a disturbance of their normal balance, leading to further
progression of cardiac failure.
| |
I. INTRODUCTION |
|---|
|
TopPreviousNextReferences
|
|---|
This review is about the interactions beween endothelial cells in the heart and adjacent cardiomyocytes and about how this cross-talk determines cardiac growth, contractile performance, and rhythmicity.
After the discovery in 1980 by Furchgott and Zawadski (186) of the obligatory role of the vascular endothelium in vasomotor tone, the endothelium has emerged as an essential structural and functional element of the cardiovascular system. Subsequent contributions by numerous eminent investigators have over the past 20 years profoundly altered our thinking about the cardiovascular system and redefined many of our working concepts (31, 32, 101, 114, 187, 198, 204, 234, 235, 396, 402, 454, 465, 598, 600-602, 604, 642). In 1998, Furchgott was awarded the Nobel Prize, together with Louis Ignarro and Ferid Murad, for their contributions to the discovery of the endothelium-derived relaxing factor, nitric oxide (NO) (396, 432). Furchgott's theorem about the central, obligatory role of the endothelial cell in cardiovascular regulation has, however, been of much more fundamental and conceptually of far greater importance in biomedical sciences (185).
The endothelium contributes to cardiovascular homeostasis not only by regulating vascular permeability but also by adjusting the caliber of blood vessels to hemodynamic and hormonal demands and by maintaining blood fluidity. Endothelial cells perform these functions by the expression, activation, and release of powerful vasoactive substances as well as of numerous other bioactive molecules. These substances include vasoconstricting and vasodilating factors, pro- and anticoagulant factors, pro- and antithrombotic factors, growth and antigrowth factors, factors that contribute to angiogenesis and tissue remodeling, as well as to immune reactions and tissue inflammation. The endothelium thus creates a complex and finely tuned balance of interactions with the immediate environment. These interactions are common to all endothelial cells. Regional and species differences in the set point of these balances may, however, result in substantial organ- or vascular bed-specific phenotypic diversity of endothelial cell function (31, 198, 236, 295, 313, 320, 355, 453, 478, 484).
In 1986, we proposed that endothelial cells in the heart might similarly play an obligatory role in regulating and maintaining cardiac function. Cardiomyocytes represent the bulk of cardiac tissue mass constituting ~75% of total tissue volume in the heart, but it has been estimated that their number accounts for <40% of all cardiac cells, the majority of cells being highly adaptive nonmyocyte cells, such as fibroblasts, macrophages, circulating blood cells, vascular smooth muscle cells, and endothelial cells. We observed that the contractile performance of isolated cardiac muscle could be profoundly modified by the presence of intact endocardial endothelial cells (55, 62, 63). Modulation of the contractile state of subjacent cardiomyocytes by endocardial endothelial cells was subsequently extended to include the contribution of the endothelium in the myocardial capillaries (323, 366, 414, 470). Further observations both in vitro and in vivo on aspects of the effects of cardiac endothelium and of related autocrine and paracrine factors and cytokines on cardiac function were made by many investigators in various animal species, including in humans (24, 48, 67, 81-83, 104, 146, 205, 206, 326, 372, 442, 446, 447, 511, 539, 543, 611) (Fig. 1).
|
As a result, our insights into cardiac function have evolved from an autoregulatory muscular pump, with coronary perfusion and neurohormonal control as the sole important modulators, to a pluricellular, multifunctional organ, in which the endothelial cell plays an essential role with respect to cardiac metabolism, growth, contractile performance, and rhythmicity. Although it is inconceivable to regard the control of blood vessels and the peripheral circulation without considering the obligatory role of the vascular endothelial cell, it is surprising that many continue to approach cardiac function merely in terms of its cardiomyocytes, coronary blood supply, and neurohormonal drive.
The present review focuses on the interactions between the cardiomyocytes and the cardiac endothelial cells of the endocardium and the intramyocardial capillaries, where the endothelial cell is closely apposed to the cardiomyocytes.
The main objectives of this review are 1) to discuss how cardiac endothelial cells interact with their immediately sub- and adjacent cardiomyocytes to control and maintain cardiac growth, contractile performance, and rhythmicity; and 2) to review current knowledge of cardiac endothelial activation and dysfunction and, in particular, how phenotypic changes may contribute to the pathogenesis of cardiac disease leading to cardiac failure.
These objectives are approached from the point of view of a traditional clinical physiologist observing global cardiac function. The emphasis is on the selective integration of all cardiac endothelium-myocardium-mediated signaling pathways within global organ structure and function. Preference has intentionally been given to whole organ integration, while at the same time selectively incorporating some of the molecular biology and genetics of more recently discovered signaling pathways which have emerged from the growing field of targeted gene manipulation.
Focusing merely on cardiac endothelial-myocardial interactions may at first seem to create a rather restrictive view on cardiac function. While avoiding wherever possible the ever-growing number of intracellular signaling pathways omnipresent in all cardiac cells but redundant, i.e., not indispensable, for the strict endothelial-myocardial reciprocal interactions, such apparent "narrow window" approach may, moreover, overlook strictly myocardium-related features of the heart (171). It also precludes other cellular interactions which may be equally important or indispensable for the maintenance of the cardiac phenotype, such as between fibroblasts or monocytes/macrophages or epicardial mesothelial cells, with either cardiomyocytes or with cardiac endothelial cells, among others (137, 209, 210, 374, 617). This approach has, however, provided novel insights. Focusing on cardiac endothelial-myocardial interactions has led to an innovative conceptual paradigm of cardiac function which challenges the centripetal, myocardium-oriented conception of cardiac structure and function to which we have become accustomed.
Review articles, published over the past decade (6, 22, 23, 57-59, 263, 281, 325, 442, 519, 522, 625), offer more comprehensive coverage of selected aspects and may provide a valuable supplement to this review.
| |
II. ENDOTHELIAL ORGANIZATION IN THE HEART |
|---|
|
TopPreviousNextReferences
|
|---|
A. Cardiac Endothelium Versus Coronary Vascular Endothelium
In the postnatal and adult heart, endothelial cells can affect cardiac function in different ways. It is indeed important to distinguish between the contribution of the cardiac endothelial cells in the myocardial capillaries and at the endocardium and the contribution of the coronary vascular endothelium in the major epicardial and smaller intramyocardial coronary arteries and veins (Fig. 2). The vascular endothelium in the coronary conduit and resistance vessels controls coronary artery function as in any other vascular bed in the body, but its contribution to cardiac function is indirect by controlling coronary blood supply to the myocardium. Cardiac endothelial cells in the myocardial capillaries (MyoCapE) and in the endocardial endothelium (EE), in contrast, are in close proximity to adjacent cardiomyocytes, allowing for direct cellular communication and signaling between both cell types.
|
Crucially important for endothelial cell-to-cardiomyocyte interactions at these two sites is their intercellular distance and their cell number ratio. The latter depends on the capillary-to-cardiomyocyte ratio and intercapillary distance, which will vary with species and cardiac sampling site. In left ventricular wall and papillary muscle of adult rat (13, 469, 591) and neonatal mice (72), capillary-to-cardiomyocyte number ratio may on cross-sectional view vary from 0.91 to 1.12. The lower figure of 0.5 observed in human "endomyocardial" biopsies from right ventricular trabeculae (401) can be ascribed to the close proximity of endocardial endothelial cells that are the dominant endothelial cells in this zone. In fact, cardiac endothelial cells outnumber cardiomyocytes by 3:1, although cardiac endothelial cell-to-cardiomyocytal volume (or mass) ratio is of the order of only 0.04-0.05 (13). The intercapillary distance was reported to be 20.2 µm in the ventricular wall (591) and 15.6 µm in papillary muscle (13) of normal rat heart and 6 ± 0.4 µm in normal neonatal mice heart (72). With a distance of ~1 µm between the capillary endothelial cell and the nearest cardiomyocyte, this provides for an action (diffusion) radius of ~3-12 µm for each capillary endothelial cell into the neighboring cardiomyocytes (72, 468, 469, 583). This would suffice for an efficient endothelial-myocardial signaling agent, even for those with short half-lives, such as endothelium-derived nitric oxide (NO) with its biological half-life of 20 s and whose liposolubility will further aid its diffusion. NO diffuses from its source over a range of 150-300 µm within 4-15 s (303); diffusion distance may even reach 600 µm (276).
The closest distance from endocardial endothelial cells to subjacent cardiomyocytes, depending on animal species and cardiac site, varies from <1 µm in small mammals to 10-30 µm in the ventricle and >50 µm in atria of larger mammals, including humans. In the subendocardial space, additional interactions of the endocardial endothelial cells with the subendocardial terminal Purkinje fiber network and with the extensive subendocardial neural plexus (95, 96, 330, 359, 644) may be equally important.
It appeared from our original experimental observations that the actions of EE and MyoCapE on myocardial contractile performance were additive, and analogous if not identical. Although EE and MyoCapE share common features in their effects on subjacent cardiomyocytes, these two cardiac endothelial cell types are, however, not identical. They differ with regard to developmental, morphological, and functional properties, the major difference probably resulting from the way in which they perceive and transmit signals and from their different hierarchic position and contribution within the overall endothelial system.
B. Cardiac Endothelial Morphology
In vertebrates, the luminal surface of the mature heart is structurally complex, comprising furrows, cylinder- and sheetlike trabeculae, and papillary muscles. In the human heart, the architecture of the ventricular wall is more elaborate in the right than in the left ventricle. In the left ventricle, trabeculations are more densely organized at the apex and posterior wall than on the septum whose surface is usually smooth; in the right ventricle, intense trabeculation may constitute over 50% of wall thickness. The complex cavitary surface of the cardiac wall is completely lined by the EE. The morphology of the EE has been well described (6). It is recognizable as a sheet of endothelial cells with a central nuclear bulge and distinct, extensive intercellular junctions. EE cells are somewhat larger than endothelial cells in most other portions of the circulatory system. The luminal surface of most of these cells possesses a scattered variety of microappendages, or microvilli, projecting into the cardiac cavity. It can be estimated that the labyrinth of trabeculae and furrows, together with the numerous microvilli on the luminal surface of the EE cells, may augment the surface area by a factor of 100 or more. This surprisingly large contact surface area of the EE offers an astonishingly high ratio of cavitary surface area to ventricular volume, particularly in the right ventricle. This suggests an important sensor role for the EE (55).
The intercellular clefts between EE cells are three to five times deeper than in MyoCapE (Fig. 3) and are often highly tortuous with two or three cells imbricated. Most clefts contain one or two tight junctions. Only in a few areas, depending on age and on species, are EE cells separated by simpler clefts that are shallower and lacking tight junctions (compare Fig. 3, top and bottom left). In MyoCapE, in contrast, most clefts are simple, shallow, and exhibit few tight junctions with many interruptions (66, 614). The degree of cellular overlap in the EE thus exceeds that in MyoCapE three- to fivefold as confirmed by experiments in which cardiac endothelial cells were stained with an antibody recognizing an endothelial marker (platelet-endothelial cell adhesion molecule; PECAM) (Fig. 4, top).
|
|
PECAM belongs to the immunoglobulin superfamily and participates in the establishment and maintenance of barrier function and permeability in many endothelia (105a, 162, 211).
In the EE, PECAM-1 staining is typically confined to the border zone of
the EE cells corresponding to the zone of cellular overlap and
intercellular clefts (8). In MyoCapE, in contrast, PECAM-1
stained more diffusely over the entire cell surface. The pronounced
difference in distribution of PECAM-1 suggests profound differences in
transendothelial permeability between these two cardiac sites.
Transendothelial transport is believed to be mediated by diffusion
through the intercellular clefts, or by nondiffusional vesicular
transport, or through cell-matrix junctions, i.e., focal adhesion
contacts (340). The paucity in most cavitary EE cells of
central actin stress fibers, which normally limit the adhesion of
endothelial cells to substrate matrix proteins, suggests that transport
through focal adhesion contacts is probably of less importance in EE.
The abundance of vesicles in MyoCapE compared with the scarcity of
vesicles in EE similarly suggests that vesicular transport would be
less prominent in EE than in MyoCapE. It would thus appear that
trans-EE-permeability is controlled predominantly through
the intercellular clefts, as mediated through 1) the extent and complex structure of the clefts which are, moreover, lined by a
dense electrically charged glycocalyx; 2) the presence of one or two tight junctions (zonula occludens); and 3) the
presence of a well-organized zonula adherens with complex
interactions between a circumferential actin filament band and several
connecting proteins, e.g., vinculin. Costaining of EE for actin and for
nonmuscle myosin (6) has suggested that changes in
trans-EE-permeability could be mediated not only by stretch
(431) or by passive retraction of the EE cells, e.g., by
phosphorylation of actin-linking proteins, such as vinculin and
-catenin, but also by activation of actin-myosin interactions.
EE and MyoCapE differ also in the way by which they communicate with adjacent endothelial as well as subjacent nonendothelial cells (59). The abundance of gap junctions between EE cells (Fig. 4, bottom left), as evident from transmission electron microscopy and immunostaining with several connexins (Cx43, Cx40, Cx37), and their absence in MyoCapE (Fig. 4, bottom right) suggest that EE displays an intimate electrochemical linkage among neighboring EE cells. This type of communication would allow for rapid intercellular electrochemical spreading of the functional properties of EE after activation of even a single EE cell (39). The ensuing amplification, moreover, would be in accordance with their suggested sensor function. The lack of gap junctions in MyoCapE, in contrast, suggests a more local control of myocardial function. Major differences were also observed in relation to the potential for communicating with nonendothelial cell types. Intercellular adhesion molecule-1 (ICAM-1) and antigens of the major histocompatibility complex (MCHI, MCHII and, DR) were more prominent in MyoCapE than in EE, PECAM was differently distributed, and the endothelial markers PAL-E and von Willebrand's factor were more abundant in EE than in MyoCapE (7).
| |
III. CARDIAC ENDOTHELIAL-MYOCARDIAL INTERACTION |
|---|
|
TopPreviousNextReferences
|
|---|
EE and MyoCapE share nevertheless many common features, typical for cardiac endothelial cells, and likely to be critical for normal cardiac growth, contractile performance, and rhythmicity. They reflect direct endothelial-cardiomyocytal interactions, e.g., through reciprocal signaling by peptide growth factors, through auto- and paracrine mechanisms, and, at least for the endocardial endothelium, through an active transendothelial blood-heart barrier. As noted above, the involvement of EE and MyoCapE in cardiac growth, contractile performance, and rhythmicity must be clearly distinguished from the functional role of coronary vascular endothelial cells in the heart. The latter endothelial cells act only indirectly through changes in coronary vasomotor tone, hence in myocardial blood supply. Because coronary vascular endothelial cells fall outside the scope of the present review, they are not considered further. The distinct features of EE and MyoCapE on cardiac growth, contractile performance, and rhythmicity will now be dealt with in greater detail (Fig. 5).
|
A. Cardiac Growth
The modulating role of the vascular endothelial cell on vascular, including coronary, growth is widely recognized. The development of a vascular supply through either vasculogenesis (in situ formation of vessels) or angiogenesis (vessel formation through outgrowth or branching) is a fundamental requirement for embryonic organ development as for wound healing and tissue regeneration in adult life. The search for potential angiogenic growth factors or regulators has yielded numerous candidates. A pivotal role in the regulation of normal and abnormal vascularization of all, including coronary vessels, has been ascribed to the specific vascular endothelial growth factor (VEGF) and the fibroblast growth factor (FGF) (439, 581, 582) among other more recently discovered molecules. We would refer the reader to the considerable literature on the subject (70-72, 145, 158-160, 167, 214, 529).
Much less is known about the impact on myocardial growth by cardiac endothelial cells (MyoCapE and EE). Whether and to what extent cardiac endothelial cells control cardiogenesis in the embryo or play a role in cardiac remodeling in the adult is under intensive investigation. Much work has been done on growth-modulating properties of the various auto- and paracrine factors, such as NO, prostacyclin, endothelin, and angiotensin, released or activated by cardiac endothelium and their possible role in hypertrophic responses or "remodeling" in the adult heart. Most of our present knowledge about the obligatory link between cardiac endothelial cells and cardiomyocytes with respect to growth comes from the rapidly expanding field of developmental cardiology, particularly from experiments on embryonic development in transgenic animal models (19, 164, 558).
1. Cardiac development
In the vertebrate embryo, the heart develops from the cardiogenic mesoderm to form the double-walled primary heart tube. This tube consists of only two cell types, i.e., endocardial endothelial cells of the inner layer, which is separated by an extracellular, amorph elastic matrix, the cardiac jelly, from the outer layer which consists of a mesh of cardiomyocytes. Endocardial endothelial cells and cardiomyocytes appear at about the same time during development. Based on morphological heterogeneity within the cardiogenic mesoderm, the two cell lineages were at first thought to develop rather independently from one another. There still remains some controversy, however, as to whether myocardial and endocardial lineages have a common or a separate origin (140, 339, 381, 382, 384). In other words, although cardiomyogenic and cardiac endothelial progenitors are both derived from lateral plate mesoderm, the relationship of these cell lineages in the embryo is unresolved. It is clear, however, that these two cell lineages diversify before the primary heart is formed. The cardiomyocytes express muscle-specific genes and proteins long before initiation of heart beat (215, 647). Endocardial endothelial cells first express an endothelial marker (the QH-1 antigen) as they segregate from the epithelialized cardiogenic mesoderm (329, 560). In an immortalized cardiogenic mesodermal cell line, the QCE-6 cell, both cardiac myogenic and endothelial cell lineages could be commonly induced in response to several different growth factors (139). Retroviral cell lineage studies do not however support a common origin from the cardiogenic mesoderm, but show rather that the cardiogenic mesoderm consists of two distinct subpopulations (88). Similarly, studies in zebrafish (309) identified that the two lineages have already separated from a common progenitor cell at the blastula stage, before formation of mesoderm. Some of these studies (139) have indicated that appropriate levels of retinoic acid, the major active metabolite of vitamin A, are required for the initial formation of myocardial and endocardial lineages and of the primary heart tube. Quail embryos cultured in the absence of retinoic acid have an underdeveloped endocardium (223) and fail to form the primary heart tube (285).
Endocardial endothelial cells subsequently invade the cardiac jelly and migrate toward the myocardial tube.1 Soon, the endocardial endothelium comes to line most of the cardiomyocytes as the sole cardiac endothelial cell monolayer. Endocardial endothelium and cardiomyocytes together constitute the primitive spongy heart tube. The first rhythmic cardiac contractions are initiated at this double-walled stage of heart development (352). Some beating myocytes seem to migrate toward the endocardium, generating a number of protrusions or trabeculae (352). The formation of this spongy and trabeculated pattern substantially increases the endocardial endothelial surface area. At a later stage, cardiac valves and septum develop through endocardial cushion formation. Cushion formation is followed by the development of more compact myocardium in the periphery of the developing cardiac walls. The formation of compact myocardium, or myocardial "compaction," is accompanied by the penetration of small buds of primitive coronary vessels invading it from the epicardial surface.
Accordingly, maturation of the heart into a four-chambered muscular pump organ requires at least three essential, consecutive developmental steps: 1) formation of a trabecular myocardial layer, 2) endocardial cushion formation with valve development and septation, and 3) growth and thickening of an outer compact ventricular chamber wall with formation of a coronary circulation (514). Unique and reciprocal signaling pathways between the primitive endocardial endothelial cells and the cardiomyocytes seem to be involved in all these processes.
A) MYOCARDIAL TRABECULATION (FIG. 6, TOP). The spongy, trabecular myocardium is largely responsible for the maintenance of blood flow during early stages of cardiac morphogenesis before expansion of the compact zone. Trabeculation of the spongy heart constitutes the first step in cardiac maturation for which endocardial-myocardial signaling is prerequisite. This signaling is essential for normal maturation and functioning of the heart and for survival.
|
/ mutant mouse, the collagen2a1-Cre VEGF Flox/+, the endocardium
appeared detached from the underlying myocardium, which was much
thinner with less-developed trabeculae and embryonic lethality
(214).
These observations emphasized that, although cardiomyocytes in the
early embryo may differentiate normally in the absence of an
endocardial tube, endocardial endothelial cells are from the onset of
cardiac development prerequisite for myocardial maturation, normal
function, and survival. In addition, threshold levels of VEGF-A
expression are required for proper endocardial-myocardial interactions.
III) Neuregulin. Further evidence for the obligatory role of
the endocardial endothelium in orchestrating myocardial cell maturation
and function has been confirmed after the discovery of the neuregulin
growth factor signaling pathway in the endocardial endothelium
(189, 287, 308,
356, 380,
651).4 At an
early embryological stage (though later still than cloche and Flk-1 expression), neuregulin-1 expression is confined to the
endocardial endothelium from where it is released as a paracrine growth
signal to the tyrosine kinase ErbB2 (HER-2, or Neu) and ErbB4 (HER-4) receptors, expressed on nearby
cardiomyocytes. Activation of the ErbB2/ErbB4
receptor complex by neuregulin is required for trabeculation of the
primitive spongy heart. Mutant embryos lacking either one of these
three genes exhibit no trabeculation and die in utero at an early stage
(Fig. 7).
|
/ mutant embryos lacked myocardial trabeculation as did mutant ErbB2 / embryos. Consistent with previous
observations that selective 5-HT2B receptor antagonists led
to defects of trabeculation (411), the findings with
ErbB2 / animals suggest interaction between these two
signaling pathways already in early heart development (411). 5-HT may indeed act as a cardiac mitogenic factor
via a 5HT2B receptor-mediated pathway, involving the
neuregulin-ErbB2 signaling pathway. The surviving
5HT2B-deficient mice exhibit severe cardiomyopathy with a
loss of ventricular mass due to a reduction in number and size of
cardiomyocytes (412).
V) Angiopoietin. An interesting parallel development was the
discovery of another family of receptor tyrosine kinases termed TIE.5 These receptors are
rather specifically expressed in endothelial cells, including embryonic
endocardial endothelial cells. Their ligand, angiopoietin-1, which is
the first member of a new growth factor-like family specific for
the TIE-2 (previously called TEK) receptor, is highly expressed in
myocardium surrounding the TIE2-expressing endocardium.6 In mutant mice
lacking either myocardial angiopoietin-1 or its receptor TIE-2 in the
endocardium, a less complex and less well differentiated endocardial
endothelial lining was observed; even though the endothelial cells were
present in normal number, the endocardial monolayer appeared somewhat
collapsed and retracted from the myocardium (461,
500, 562). This abnormality too resulted in
markedly diminished myocardial trabeculation. Angiopoietin-1 produced
by the myocardium thus appears to influence the development of the
adjacent endocardium which, in turn, provides key signals required
for normal myocardial trabeculation. Angiopoietin-2, in contrast to
the myocardium-derived angiopoietin-1, is highly expressed in
microvascular endothelial cells (353, 422).
It is a naturally occurring antagonist of angiopoietin-1 in that it
competes for binding to TIE-2 and blocks angiopoietin-1-induced TIE-2
autophosphorylation. Overexpression of angiopoietin-2 in microvascular
endothelial cells in transgenic mice resulted in embryonic defects
reminiscent of those observed in angiopoietin-1 and TIE-2 knock-out
mice (347). The implications with respect to
endothelial-myocardial signaling in the developing heart need further investigation.
An essential, direct or indirect transcriptional regulator of
angiopoietin-1 and VEGF in the developing heart is the MEF2C gene that
is expressed in cardiac muscle cells as well as in endothelial cells.
Targeted deletion of MEF2C in mice resulted in dramatically reduced
cardiac expression of angiopoietin-1 and VEGF with malformed and
irregularly oriented endocardial endothelial cells and severely constricted endocardial lumen (40), loss of the right
ventricle, and failure of the remaining ventricle to loop
(328). Some of these malformations resemble endocardial
alterations produced by null mutations in VEFG-A (209)
and angiopoietin-1 or its receptor TIE-2 (500,
562).
Accordingly, just before myocardial trabeculation there seems to exist
a critical interdependence between various signaling pathways,
including VEGF, neuregulin, serotonin, and angiopoietin-1 growth factor
pathways, with reciprocal signaling between the endocardial endothelial
cells and the developing myocardium (19, 102,
574). These endocardial endothelial-myocardial
interactions are highly specific and essential steps in normal
cardiogenesis. If one of these signaling pathways is disrupted, the
ventricle remains untrabeculated and most of the animals soon die in utero.
B) ENDOCARDIAL CUSHION DEVELOPMENT: VALVE FORMATION AND SEPTATION
(FIG. 6, BOTTOM).
I) Endocardial endothelial lineage. In
the region of future cardiac valve formation, endocardial endothelial
cells are transformed into mesenchymal cells which migrate into the cardiac jelly. Transformation into mesenchymal cells is accompanied by
downregulation of the endothelium-specific PECAM-1
(299) and QH1 (119, 407)
expression. It occurs during a narrow spatiotemporal window of intense
programmed cell death, or apoptosis (264,
456, 597, 653). The resulting
local mass of mesenchymal cells pushes the endocardial layer into the
cardiac lumen and increasingly expresses smooth muscle -actin,
necessary for cell motility and migration (119,
407), as well as urokinase, necessary for remodeling of
the matrix and cell migration (348, 369,
545). These protrusions are known as the endocardial
cushions from which the cardiac valves and septa are formed
(302, 616). Predictably, the
cloche mutant heart in zebrafish, which lacks endocardium,
lacks also cardiac cushions and valvuloseptal formation
(549).
. An understanding of the
mechanisms by which cushion endothelial cells respond to this
myocardial induction is fundamental to the understanding of growth
responses during cardiac development (339,
409). Candidate molecules that have been implicated in the
cushion transformation process are transforming growth factor-
(TGF-) and bone morphogenetic protein (BMP), among others. Virtually
every cell in the body produces TGF- and has receptors for it. Of
the three isoforms, TGF-1 and TGF-3 are
expressed in the heart early during embryonic development, in
endothelial, and mesenchymal cells, respectively. TGF- is increasingly expressed in cushion endothelial cells just before their
transformation into mesenchyme, but also in the immediate subjacent
myocardium (50). TGF- expression by the endocardium is
under regulation of the ES proteins (54, 120,
406).TGF-2 was only transiently expressed
at the time of induction of the cushion in the cardiomyocytes
underlying the regions in which the endocardial-mesenchymal
transformation will take place (30). Absence of
TGF-2 in TGF-2-null mice resulted in
abnormal valve leaflets and septa that derive from endocardial cushion
tissue. These TGF-2 knock-out mice die around birth.
TGF- binds to its receptors TR-I, TR-II, and TR-III
(54) and interacts with other molecules such as betaglycan
and endoglin. The latter two binding molecules are components of the
TGF- receptor complex, which may modulate its biological action.
Endoglin was found to be strongly but transiently expressed in
endocardial endothelial cells, but not in the myocardium, during
endocardial cushion formation (464). The receptors for
TGF-, TR-I, and TR-II, as well as betaglycan, were found
mostly on cardiac myocytes and were detectable only at low levels on
endocardium. Transient upregulation of the various constituents of the
TGF- receptor complex, and in particular of endoglin, is also an
essential step in valve formation and ventricular septation.
Yet, there is no apparent cardiac malformation in single-knock-out
mutant mice lacking either TGF-1 or
TGF-3, which suggests pharmacological redundancy of
TGF- in cushion formation even though TGF- is essential for
cardiogenesis (409). BMP, as a member of the TGF-
growth factor superfamily, is also expressed in the subjacent
myocardium and acts synergistically with TGF- to initiate
endothelial-to-mesenchymal transformation (50,
409, 634, 635).
Finally, there exists some interaction between the TGF- pathway with
retinoic acid, the major active metabolite of vitamin A. The retinoic
acid signaling pathway, noted above, plays an essential role at an
earlier stage in the establishment of the primary heart tube and in the
expression of induction protein complexes in the cardiomyocytes before
cushion formation. Retinoic acid deficiency in mice resulted also in
downregulation of TGF-1 (33), whereas
retinoic acid treatment resulted in an increased expression of
TGF-1 in endocardial and mesenchymal cells
(346). Mice deficient in RXR-, one of the retinoic acid
receptors, display hypoplastic cushion formation (208).
TGF- would thus seem to act downstream of the retinoic acid
signaling in cushion formation.
IV) Neuregulin. The neuregulin growth factor
(NRG-1) pathway (discussed above) was shown to also play a specific and
essential step in endocardial cushion formation through the
ErbB2/ErbB3 receptor complex, which is
expressed by the prevalvular mesenchymal cells adjacent to the
neuregulin-expressing endocardium of the endocardial cushions
(144). Null mutation in the ErbB3 receptor resulted in abnormal cardiac valve formation with subsequent congestive heart failure and death in utero.
V) Insulin-like growth factor I. Endocardial
endothelium also expresses insulin-like growth factor I (IGF-I).
IGF-I expression is highest in the endocardial endothelium, appears
to decrease in the mesenchymal cells, and is absent in the myocardium
(388). While having no apparent effect by itself, in
contrast to NRG-1, IGF-I interacts with NRG-1 synergistically to
promote expansion of the atrioventricular cushion tissue
(226).
VI) Myocardial-derived hepatocyte growth factor.
Myocardial-derived hepatocyte growth factor (HGF) may also
participate in inducing and maintaining endothelial-to-mesenchymal
transformation, at least partly by upregulating the expression of
urokinase in the endocardium-derived mesenchymal cells
(545).
VII) Nuclear factor of activated T cells. The
transcription factor NF-ATc (or NF-AT2)
(nuclear factor of activated T cells), known to participate in the
immune system, was selectively expressed by endocardial endothelial
cells near the endocardial cushion (111, 416,
472). Mutant NF-ATc / embryos did not
undergo outflow tract valve formation and underwent only partial
ventricular septation with subsequent congestive heart failure and
premature death of the embryo. The atrioventricular valves were either
less severely affected (472) or normal (111)
in these NF-ATc / mutant embryos. NF-ATc
would appear to be involved in events that are essential for valve
formation, but after, and independently of endothelial-to-mesenchymal
transformation, when TGF-, which is expressed at normal levels in
NF-ATc / heart, is no longer involved
(472).VEGF was recently shown to play an important role also in valve formation by inducing NF-ATc activation
specifically in these valvular endothelial cells (495).
VIII) Neurofibromin. Another interesting
development was the observation that neurofibromin (Nf1), which is
transiently overexpressed in cushion mesenchymal cells, downregulates
endothelial-to-mesenchymal transformation (301). Mutant
mice deficient of Nf1 (Nf1 /) display a dramatic overabundance of
spongy endocardial cushion tissue, with failure to condense and thin to
form mature valve leaflets due to a lack of appropriate apoptosis in
the Nf1 / endocardial cushions in addition to increased
proliferation (301). Subsequent compaction of the
ventricular myocardium is also hampered, and often accompanied by
ventricular septum defect and thinning of the compact myocardial layer,
with death in midgestation.
IX) Platelet-derived growth factor. Another
candidate receptor tyrosine kinase expressed in the transforming
cushion tissue is the platelet-derived growth factor receptor
-subunit (PDGF-R) (301). PDGF-R can bind the
secreted ligand PDGF-A which is expressed at high levels in the
cardiomyocytes. This was initially believed to be indispensable
(502), but PDGF-R knock-out mice did not show
endocardial cushion defects (547).
X) Pre-B-cell growth-stimulating factor. A
variety of other molecules including the chemokine Pre-B-cell
growth-stimulating factor/stromal cell-derived factor-1
(PBSF/SDF-1) (405) are also expressed in the endocardial
endothelium and seem to be involved in endocardial-myocardial
signaling. Mutant PBSF/SDF-1-deficient embryos develop cardiac
ventricular septal defects. As cushion and valve formation are slightly
retarded but normal in these mutant embryos, these signaling pathways
appear to have an essential but rather late developmental function. In
situ hybridization revealed that PBSF/SDF-1 mRNA was expressed in the
endocardium of the muscular septum, when the membraneous portion of the
ventricular septum is to be formed, suggesting a key role of this
chemokine in ventricular septum formation (405). Most mice
lacking PBSF/SDF-1 died perinatally.
XI) Endothelin. A putative role of the
endothelium-derived endothelin (ET) pathway has also been
demonstrated. In the embryo, ET-1 mRNA is expressed in the entire
endocardial endothelium around the time of cushion formation
(296, 297). Moreover, although ET-1 mRNA and
protein were also expressed in cardiomyocytes of rat embryonic hearts,
immunostaining with anti-ET antibody was strongest near the
endocardium (415). The mRNA for ET-converting enzyme
(ECE-1) and for the ETA receptor were expressed both in the
endocardial endothelium and subjacent myocardium (641).
Yet, in neither ET-1- or ETA-deficient mutant mice embryos
nor after pharmacological suppression of
ET-1/ETA/ETB signaling was the induction of
endothelial-to-mesenchymal transformation inhibited (87,
296, 408); in contrast, a large proportion of
the ET-1-deficient animals developed ventricular septal defects at a
later stage (296). Hence, ET-1/ETA signaling
appears not to be essential for initiating cushion formation but may
play a role at later stages near the completion of the transformation
process as well as during early compaction of the septum and
ventricular wall. This is supported by the more severe cardiac defects
observed after disruption of ECE-1 (440). ECE-2 mRNA is
largely absent in the heart before these stages and seems to be
expressed exclusively in the mesenchyme of the endocardial cushions
(640). Most intriguingly, although neither single ECE-1
/ nor ECE-2 / mice display any detectable developmental defect
in cushion formation, combined ECE-1 / and ECE-2 / double null
mice embryos exhibited abnormal atrioventricular valve formation,
despite significant residual tissue levels of ET-1/ET-2. This suggests
that ECE-2 may contribute to the production of mature ET-1 in situ in
the endocardial cushion mesenchyme or that it would function in
endocardial cushion by cleaving non-ET peptides essential for valve
formation (640).
XII) NO. A similar rather late contribution of
endothelium-derived NO in cushion formation has been suggested from
experiments in mice lacking endothelial constitutive NO synthase
(ecNOS). These ecNOS / animals survived into mature animals but
demonstrated a high incidence of bicuspid aortic valves; in the
corresponding wild-type embryonic mice, ecNOS was localized to the
cardiomyocytes and endocardial endothelial cells lining the developing
valve leaflets (310). In another study on ecNOS /
mice, a high incidence of atrial and ventricular septal defects, again
suggesting cushion malformation, was observed, with death soon after
birth (155). It may also be worth mentioning that ecNOS
and inducible NO synthase (iNOS) were prominently but transiently
expressed at a late phase in the myocardium of mice and rat embryos and
in a model of embryonic stem cell-derived cardiomyocytes during
maturation into adult cells, becoming undetectable in these
cardiomyocytes when terminally differentiated (43). As
pharmacological inhibition of NOS resulted in a pronounced
differentiation arrest of the cardiomyocytes, these observations would
suggest that at some stage, corresponding probably to the late phase of
cushion formation and early myocardial compaction,
myocardium-derived NO could in addition to the
endothelium-derived NO also play a role in cardiomyogenesis.
That NO is probably essential for cardiomyocyte survival at these early
stages of development is further supported by the observation that
ecNOS-deficient mutant mice show increased cardiomyocyte apoptosis
(154, 155). We (Andries, Sys, and Brutsaert,
unpublished data) and others (594) observed that ecNOS
expression in the EE in the developing rat heart occurred at a much
earlier stage than ecNOS expression in MyoCapE, even when myocardial
capillaries in the compact zone of the ventricular wall were already
detectable by rat endothelial cell antibody (RECA) staining. During
later compaction, ecNOS was present not only in EE but also in the
developing MyoCapE and in coronary vessels within the compact
myocardium, although absent from the cardiomyocytes (594).
XIII) Angiotensin. Mutant mice deficient of
angiotensin type 1A and 1B receptor genes similarly survive normally in
utero but may develop ventricular septal defects suggesting an
important but late ontogenic role of angiotensin II (ANG II) and its
AT1 receptor in cardiac development (590).
C) FORMATION OF COMPACT MYOCARDIUM (MYOCARDIAL COMPACTION). At a later stage during cardiac development, the outer myocardial layers of the embryonic heart become compact. Initially, the compact zone is only about two myocardial cells thick and is avascular. As compaction expands through cardiomyocyte proliferation to a thickness of three to four cells, angioblasts begin forming vascular tubes from the epicardium (581).
The compact zone comprises the outermost layers of the ventricular cavities and a major portion of the muscular septum. In birds, the septum arises from coalescence of trabecular sheets. In mammals in contrast, the septum has a compact arrangement from the onset (514). The ventricular septum is thus formed by two distinct processes: 1) through cushion formation with septation of the conotruncus and outflow tract and coalescence of trabeculae at the interventricular groove, and 2) through inward growth, or myocardialization (596), and compaction of the fused medial walls of the expanding ventricles forming the major muscular portion of the septum. The molecular and functional phenotype of the ventricular compact myocardium differs distinctly from the ventricular trabecular component (169-171, 398). The maturing and differentiating trabeculated myocardium expresses specifically the cyclin-dependent kinase inhibitor p57Kip2, which is absent in the proliferating compact myocardium (278). Trabeculations, moreover, retain the characteristics of the primary myocardium. They express predominantly 1) atrial-specific isoforms of the contractile protein, compared with the ventricle-specific isoforms of the contractile proteins in compact myocardium; 2) atrial-like gap junctional Cx40 mRNA and protein rather than the ventricle-like Cx43 mRNA and protein in compact myocardium; and 3) a higher sarcoplasmic reticulum Ca2+-ATPase (SERCA2) and lower phospholamban (PLB) mRNA than in compact myocardium (170). As a result, trabecular myocardium has a quicker twitch contraction and faster impulse conduction (85) but shows a longer lasting relaxation. This performance pattern suggests that trabeculations may play an important transient role by driving contraction and impulse conduction during maturation when specialized conduction tissue and mature myocardial cells have not yet been fully formed. Gene inactivation studies in mice have indicated that myocardial compaction is under the control of a great variety of growth signaling pathways and their myocardial-specific receptor genes, which apparently share a common molecular pathway necessary for compaction, such as retinoic acid (RA) (the major active metabolite of vitamin A) and the myocardial RA receptors (RAR, RXR) (46, 47, 258, 259, 514, 559), basic FGF (bFGF) (581) and its receptor tyrosine kinase (FGFR1) (381), TGF-2
(497), the nuclear protein jumonji (311), the
FOG-2 gene as an essential developmental cofactor for
GATA-4/5/6 (575), the -dystrobrevin gene
(233), and probably many more pathways still to be
discovered.8 Many of the
known pathways have been shown to initiate or to enhance compaction,
but some may counteract this process and serve as a suppressor of
cardiomyocyte proliferation. Mutant TGF- / mice, for example,
develop a significant thickening of the ventricular wall and a loss of
ventricular chamber volume, both of which result from cardiomyocyte
hyperplasia (315).
I) Endothelin. Many genes that encode the
transcription factors for initiating the thickening of the outer
compact ventricular myocardium are expressed in the cardiomyocytes,
but it is at present unclear whether additional signals from the
endocardial endothelium, from the developing epicardial coronary
vascular endothelium, or from the MyoCapE are also required
(581, 582). At the onset of compaction, for
example, mRNA for ECE-1 and ET-1 are expressed both in the endocardial
cushion tissue and in the endocardial endothelium of the ventricular
cavities (640), and they have been shown to contribute to
the compaction process of the ventricular wall and the large portion of
the muscular septum. Kurihara et al. (296) observed
hypoplasia of the compact zone of the ventricular wall, including
severe ventricular septal defects, in the mutant ET-1-deficient embryos
when they were treated also with the ET antagonist BQ-123 to eliminate
the effects of circulating maternal ET. Intriguingly, an
interventricular septum defect was observed in all the ECE-1 /
mutant embryos, despite significant remaining ET-1 tissue levels
(641).
II) Erythropoietin. Erythropoietin (EPO) is also
an essential growth factor during compaction (628). Hearts
from mutant EPO / or EPO receptor (EPO-R) / embryos suffered
from ventricular hypoplasia, coupled to defects in the interventricular
septum, due to a reduction in the number of proliferating
cardiomyocytes in the septum and ventricular compact layer. EPO-Rs
are highly expressed in endothelial cells (5) including in
the endocardium, but not in the cardiomyocytes (628).
Moreover, EPO has been shown to promote endothelial cell proliferation
in vitro (121). It would, therefore, be feasible that EPO
would trigger cardiomyocyte proliferation indirectly through its action
on the endothelial cells, both in the endocardial endothelium of the
trabeculated layer and in the cardiac endothelial cells of the
proliferating myocardial capillaries in the expanding compact
myocardial layer. An obvious potential candidate for this
EPO-triggered endothelial-myocardial interaction at these two sites
could be the EPO-induced release of ET-1. ET-1 is highly expressed
in these two types of endothelial cells as is the ETA
receptor on the cardiomyocytes at this later stage of development.
Isolated cardiomyocyte cell proliferation and
[3H]thymidine incorporation studies have indeed revealed
a mitogenic action of EPO in cells isolated from EPO / mice, but
not in cells from EPO-R / animals. The involvement of MyoCapE
is supported by the observation of epicardial detachment and lack of
defined myocardial capillary structures in EPO / and EPO-R
/ hearts (628).
III) NRG-1 and IGF. Interaction of the cardiac
endothelium-derived growth factors NRG-1 and IGF-I
synergistically induced substantial DNA synthesis of cardiomyocytes
with significant growth and thickening of the ventricular compact zone
(226). As pointed out by the authors, this is a most
intriguing observation since NRG-1, when given alone to a whole mouse
culture system, induced trabeculation of the ventricular wall but
without a significant increase in the proliferation of cardiomyocytes,
whereas IGF-I, while essential for cardiac growth in neonatal
heart, had no apparent effect by itself on cardiac development.
D) CORONARY VASCULARIZATION. The embryonic heart of higher vertebrates is avascular until the myocardium becomes thickened through compaction. In human embryos, the first sign of definitive blood vessels was found to be localized in the subepicardial space of the apical interventricular incisura (228). These primitive coronary vessels are lined by a monolayer of (coronary) vascular endothelial cells which soon become physically continuous and contiguous with the endocardial endothelium. Hence, the development of a mature coronary circulation and the establishment of a myocardial capillary vascular plexus are relatively late events, the phylogenetic and embryological importance of which depends entirely on the ratio of compact-to-spongious myocardium.9 Recent studies support the hypothesis that coronary vascular growth during compaction is genetically induced by the myocardium (575) and somehow regulated, at least in part, by the rate and magnitude of myocardial growth (580). This is supported further by the observation that threshold levels of VEGF-A expression in the myocardium are required, not only for proper formation of endocardial endothelium as we have seen above, but also for the establishment of a myocardial vascular network (72, 214, 586). Currently, however, the underlying regulation of embryonic coronary vasculogenesis (through in situ local formation) or angiogenesis (through outgrowth or branching of preformed vessels) are still not well understood (581, 582). Experiments with replication-defective retrovirus-mediated genetic tags (386) have, however, revealed vasculogenesis, rather than angiogenesis, as the mechanism of coronary vessel formation (383, 385).
Are endocardial and vascular endothelia derived from distinct cell lineages? From traditional embryological observations, the answer to this question has been unclear. Recent biochemical evidence, however, indicates that endocardial endothelial cells are distinct from coronary vascular endothelial cells, including MyoCapE (339, 383). For example, as noted above, expression of NF-ATc is restricted to specified EE cells (111, 472). Other evidence that EE cells and MyoCapE are distinct is provided by the zebrafish mutant cloche, which lacks an endocardial tube but not blood vessel endothelium (549). The fact that cloche acts upstream of VEGF and its receptor Flk-1 (327) also indicates that the formation of EE and MyoCapE cells may be regulated through different signaling pathways. In further support of these functional differences is the observation that adult EE cells in culture tended to proliferate and to reach confluence significantly faster than cultured MyoCapE or coronary vascular endothelial cells (9, 371, 373, 417, 418). Accordingly, the emerging field of targeted gene manipulation has over the past 10 years strongly supplemented the contention that cardiac endothelial-myocardial interaction is a prerequisite for normal cardiac development, structure, and function. It has, moreover, resulted in a better understanding of some of the molecular mechanisms and cellular signals governing this interaction. This interaction involves a finely tuned series of molecular, morphological, and functional events that, if perturbed even slightly, can have dramatic consequences. Information from null phenotype experiments grows and as more molecular pathways are discovered and characterized, this overview will inevitably need updating. The extent to which these molecular pathways govern physiological or pathophysiogical cardiac adaptations in the adult heart remains to be elucidated (79).2. Mature and adult heart
Most studies on the cardiac endothelial-myocardial signaling pathways discussed above have focused on their role in cardiac development rather than their regulatory role postnatally.
It is still generally believed that cardiomyocytes in the postnatal and adult heart are terminally differentiated, although recent reports suggest the contrary (12, 37, 250).
Cardiomyocytes may, however, respond by hypertrophic growth to a wide array of stimuli, including mechanical and oxidative stress, as well as metabolic (hypoxia), neurohormonal, and growth factors. Many factors with myocardial growth-modulating properties have been shown to be released by adult cardiac endothelial cells. The possibility that auto- or paracrine pathways including cardiac endothelial-myocardial interactions might modulate cardiac growth or gene expression during adaptive hypertrophy is intriguing.
That cardiac endothelial cells might be indispensable for myocardial
growth in the adult was suggested by experiments with cardiomyocytes
cocultured with endothelial cells. Only in the presence of cardiac
endothelial cells could the cardiomyocytes maintain their adult
phenotype; when vascular endothelial cells from aorta or fibroblasts
were added, cultured cardiomyocytes continued to undergo
dedifferentiation and reexpression of fetal proteins observed in
long-term monoculture of adult cardiomyocytes (138).
Endothelial cells stimulated also the secretion of atrial natriuretic
peptide from atrial cardiomyocytes in coculture (318, 319). Reciprocal signaling from cardiomyocytes to MyoCapE
has also been observed: MyoCapE cell growth was indeed significantly enhanced when cocultured with cardiomyocytes (414).
Moreover, only in coculture with cardiomyocytes were MyoCapE able to
express mRNA for both TGF- precursor and pre-pro-endothelin (ppET).
In these coculture experiments, TGF- acted as an autocrine cytokine, increasing ppET mRNA and inhibiting the rate of MyoCapE cell
proliferation (414). In other studies, expression of von
Willebrand factor in MyoCapE was shown to be induced through a
signaling pathway mediated by cardiomyocyte-dependent,
platelet-derived growth factor AB heterodimer (PDGF-AB); only MyoCapE
with PDGF- receptors could transduce the signal for the expression
of the gene for von Willebrand factor, whereas neighboring MyoCapE
devoid of these receptors lacked this ability (2,
136, 484). Finally, endothelial cell progenitors in mice embryo as well as differentiated endothelial cells
from mice umbilical vein can differentiate into beating cardiomyocytes
when cocultured with neonatal rat cardiomyocytes or when injected near
the damaged area of the heart after occlusion of a coronary vessel, the
latter transdifferentiation being independent from signaling molecules
active in embryogenesis (91a).
A) ROLE OF PARACRINE ENDOTHELIAL-MYOCARDIAL SIGNALING. Endothelial-derived molecules such as NO, ET, PGI2, and ANG II may influence myocardial growth, as they are known to influence vascular smooth muscle growth. NOS and ET are expressed, albeit at negligibly low levels, in quiescent cardiomyocytes as well as constitutively expressed in cardiac endothelial cells of the normal adult heart. ecNOS expression in cardiac endothelium is highest in EE and only scant in MyoCapE (8). Whether endothelial or myocardial in origin, NO and ET may theoretically participate in growth responses in the adult heart. Cardiac endothelial dysfunction has for example been invoked to explain maladaptive growth responses during the progression of heart failure.
I) NO. Many experimental and clinical observations support an antigrowth effect of NO in the adult heart, which may, at least in part, explain the beneficial effects of angiotensin converting enzyme (ACE) inhibitors on ventricular "remodeling" (68, 216, 220, 238, 247, 332, 345, 363, 367, 368, 524). Their inhibition of bradykinin breakdown with enhanced bradykinin-induced NO release will contribute to their antigrowth, representing another example thereby of cardiac endothelial-myocardial signaling pathways. Bradykinin has a direct growth-stimulating effect on cardiomyocytes in monoculture, but an antigrowth effect which is critically dependent on the presence of endothelial cells in coculture, and in particular on the release of endothelium-derived NO and PGI2 (479, 480). Disruption of the bradykinin B2 receptor in mice leads to inappropriate cardiac hypertrophy with marked chamber dilatation and reparative fibrosis (142), an effect which was prevented by an ANG I receptor antagonist (142). Targeted disruption of the tissue kallikrein gene triggered dilated cardiomyopathy in mice (375), further supporting the view that kinins normally stimulate the release of NO and PGI2 through activation of bradykinin B2 receptors. Blunted basal activation of the NO (and PGI2?) pathway in B2/ knock-out mice would leave unbalanced ANG
II-induced myocardial growth. The kallikrein kinin system may,
therefore, be cardioprotective by counterac
