| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
PHYSIOLOGICAL REVIEWS Vol. 79 No. 1 January 1999, pp. 215-262
Copyright ©1999 by the American Physiological Society
Institut National de la Santé et de la Recherche Médicale U. 127, Hôpital Lariboisière, Paris, France
I. DEFINITION: LIMITATIONS
II. MORPHOLOGICAL AND CLINICAL BASIS
A. Anatomic Basis
B. Cellular Alterations
C. Clinical Setting and Treatment
III. PHENOTYPIC CHANGES IN MYOCYTES
A. Methodological Problems
B. General Process of Biological Adaptation
C. Cardiac Hypertrophy
D. Energy Metabolism
E. Membrane Proteins
F. Contractile Proteins
G. Contractile Cycle and Excitation-Contraction Coupling
H. Cardiac Autocrine Functions
I. Cytoskeleton
IV. PHENOTYPIC CHANGES IN THE EXTRACELLULAR MATRIX
A. Normal Cardiac Extracellular Matrix
B. Myocardial Fibrosis of Known Origin
C. Cardiac Overload Without Fibrosis
D. Pressure Overload Hypertrophy
E. Other Components of Fibrosis
F. Other Components of the Extracellular Matrix
V. CELL DEATH
A. Ischemia
B. Cardiac Hypertrophy and Failure
C. Vasoactive Peptides and Catecholamines
VI. CHANGES IN GENE EXPRESSION IN RELATION TO MYOCARDIAL FUNCTION
A. Systolic Ejection and Active Relaxation
B. Diastolic Dysfunction
C. Arrhythmias and Changes in Heart Rate
D. Transition to Cardiac Failure
VII. CARDIAC REMODELING DUE TO SENESCENCE
A. The Senescent Heart: an Overloaded Heart
B. Senescence of the Myocardium: a Specific Biological Process
VIII. CONCLUSION
REFERENCES
 |
ABSTRACT |
|---|
 Top References
|
|---|
Swynghedauw, Bernard. Molecular Mechanisms of Myocardial Remodeling. Physiol. Rev. 79: 215-262, 1999. 
"Remodeling" implies changes that result in rearrangement of normally existing structures. This review focuses only on permanent modifications in relation to clinical dysfunction in cardiac remodeling (CR) secondary to myocardial infarction (MI) and/or arterial hypertension and includes a special section on the senescent heart, since CR is mainly a disease of the elderly. From a biological point of view, CR is determined by 1 ) the general process of adaptation which allows both the myocyte and the collagen network to adapt to new working conditions; 2) ventricular fibrosis, i.e., increased collagen concentration, which is multifactorial and caused by senescence, ischemia, various hormones, and/or inflammatory processes; 3) cell death, a parameter linked to fibrosis, which is usually due to necrosis and apoptosis and occurs in nearly all models of CR. The process of adaptation is associated with various changes in genetic expression, including a general activation that causes hypertrophy, isogenic shifts which result in the appearance of a slow isomyosin, and a new Na+-K+-ATPase with a low affinity for sodium, reactivation of genes encoding for atrial natriuretic fator and the renin-angiotensin system, and a diminished concentration of sarcoplasmic reticulum Ca2+-ATPase, 
-adrenergic receptors, and the potassium channel responsible for transient outward current. From a clinical point of view, fibrosis is for the moment a major marker for cardiac failure and a crucial determinant of myocardial heterogeneity, increasing diastolic stiffness, and the propensity for reentry arrhythmias. In addition, systolic dysfunction is facilitated by slowing of the calcium transient and the downregulation of the entire adrenergic system. Modifications of intracellular calcium movements are the main determinants of the triggered activity and automaticity that cause arrhythmias and alterations in relaxation.
"Remodeling" qualifies changes that result in the rearrangement of normally existing structures. Although remodeling does not necessarily define a pathological condition, myocardial remodeling is usually restricted to diseased conditions. The above definition eliminates gestational and developmental aspects and also the so-called physiological cardiac hypertrophy that follows intensive exercising. Remodeling concerns the two components of the cardiovascular system. The structure of both the myocardium and the vessels, including the coronary vessels, is indeed able to change under the influence of external factors, such as ischemia and mechanical overload. This review is focused on acquired cardiac remodeling (CR) of the left myocardium. Remodeling of coronary arteries, right ventricle, and atria [with the exception of the left atrial changes that compensate for the deficit in left ventricular (LV) filling during CR], CR of genetic origin, the mechanisms of coronary restenosis, and remodeling of the large vessels under the influence of the atherosclerotic process and arterial hypertension have been excluded. Although "myocardial remodeling" is now a widely used term, from a historical point of view it was initially used to describe the remodeling that occurs following myocardial infarction (MI). The meaning of the word was subsequently extended and used to qualify a variety of conditions including pure mechanical overload as well as hypertensive, valvular cardiopathy, familial hypertrophic, and dilated cardiomyopathy (reviewed in Ref. 410; see Table 1). Transgenic manipulations (reviewed in Ref. 446) as well as experimental or clinical hormonal intoxications (due to thyroxine, angiotensin II, aldosterone) are also able to remodel the myocardium. In general, myocardial remodeling is a reversible process provided the cause of the remodeling has been either suppressed or attenuated. This has considerable links to clinical and basic pharmacology and has been extensively reviewed recently (432).
The literature on the subject is extensive, and the present review is focused on permanent modifications in the left ventricle in relation to clinical dysfunction, in the most commonly observed conditions of CR, i.e., myocardial ischemia and/or arterial hypertension. Hibernating myocardium is a reversible state of persistently impaired ventricular function at rest which copes with a reduced coronary blood flow. The hibernating response may be acute or chronic and could be considered to be a process of adaptation. The biological mechanisms of hibernation need a specific approach and are excluded from the present review. From an epidemiological point of view, MI and arterial hypertension are both more frequent in aged persons. In addition, the structural modifications that are observed during senescence in healthy individuals has several points in common with the mechanically overloaded heart (32). Cardiac remodeling is triggered by mechanical stretch; nevertheless, there are also several different factors including ischemia, hormones, and vasoactive peptides, which can modify the effects of the mechanical factor. The most common clinical situation during which remodeling is known to occur is a rather complex mixture of ischemia, stretch due to pressure overload, stress due a myocardial scar, and increased plasma levels of hormones or vasoactive peptides. An additional factor has to be listed, namely, the unknown signal that provides to the heart the information concerning the amount of substance that has been lost after MI and that is exactly recovered by the compensatory hypertrophy; such a signal should be similar to that occurring after uninephrectomy, unipneumonectomy, or partial hepatectomy. The mechanisms by which cardiac growth occurs have already been reviewed (85, 108, 234, 346, 515), and because they could themselves constitute a full review, they are not developed in this review. The permanent changes in molecular structure and their consequences in terms of cell physiology can be divided into three principal mechanisms: the deleterious consequences of the general process of adaptation, including cardiac hypertrophy, cell death, and fibrosis. The goal of this review is not only to review the extensive literature concerning CR, but also to try to simplify a complex problem and to identify pathways for future research. A review on the same topic was published in 1982 (449); only a selection of the articles published before this date have been quoted. A. Anatomic Basis
Cardiac remodeling is accompanied by an increase in LV mass and volume and a change in the shape of the ventricle (89, 90). If the process is triggered by MI, the remodeling is asymmetric and is associated with infarct expansion. Early ventricular remodeling following coronary occlusion is dominated by infarct expansion, which is an acute dilation of the infarction area that cannot be explained by additional myocardial necrosis. Infarct expansion is caused by death and slippage of the myocytes; it usually predominates in the apical region of the left ventricle and dramatically alters ventricular volume and geometry. In rats, two days after a MI causing a loss of 63% of the myocytes, the transverse mid chamber diameter increases by 20% and the wall thickness decreases by 33%. Simultaneously, both the total number of myocytes and capillary profile in the spared region of the LV free wall are reduced by 40%. The combination of these abnormalities indicates side-to-side myocyte slippage and accounts for the seven- to eightfold increase in diastolic wall stress (338). Ventricular remodeling is the result of infarct expansion of akinetic-dyskinetic segments and volume-overload hypertrophy of noninfarcted segments. Catheterization and ventriculography have demonstrated a volume enlargement, a lengthening of the ventricular perimeters, and an increased sphericity index both in systole and diastole that is accompanied by a blunting of the normal curvature of the apex. Later, what is usually called remodeling is characterized by additional enlargement and sphericity of the ventricle, a decrease in stroke volume, and impaired diastolic filling (88, 148, 208, 288, 303, 352). Based on these observations, the following model of ventricular remodeling after MI has been proposed: systolic impairment secondary to the loss of contractile material results in an increased end-systolic volume, an increased cardiac size, and a secondary augmentation of the diastolic filling pressure and distensibility. As the fibrosis increases, the distensibility decreases, resulting in an increase in diastolic pressure and volume. Peripheral mechanisms including vasoconstriction subsequently increase both preload and afterload, which results in an increased wall stress and a progressive thinning of the area. Simultaneously, in noninfarcted segments, elevation of the end-diastolic stress causes volume-overload hypertrophy that tends to normalize the wall stress according to Laplace's law. The extent and location of the infarction, therapy, or associated diseases may considerably modify remodeling (288). When CR originates from a more general process such as arterial hypertension or valvular disease, the ventricular hypertrophy remains symmetric. Compensated cardiac hypertrophy (CCH) is concentric and is initially characterized by a thick ventricular wall and septum, a normal internal volume and wall stress, and a high mass-to-volume ratio. Cardiac failure (CF) is progressive and is accompanied by a progressive enlargement of the ventricular cavity, and the mass-to-volume ratio returns to normal values (444). B. Cellular Alterations
The first demonstration that, in adults, cardiac myocytes hypertrophy and cardiac nonmuscular cells both hypertrophy and divide by mitosis was made using thymidine labeling in the laboratory of Rabinowitz and co-workers (163, 319). This finding provided the basis for the general belief that postnatal growth of cardiocytes occurs through myocyte hypertrophy and that cardiac myocytes are terminally differentiated cells that are unable to reenter the cell cycle. In contrast, in the nonmuscle cells, which include both fibroblasts and endothelial cells, a hyperplastic component persists. Therefore, CR is a result of myocyte hypertrophy, hyperplasia, and hypertrophy of the nonmuscular cells and interstitial cell growth. Such a paradigm has however to be reevaluated, since CR is also known to be associated with polyploidy (this has been well documented in the hypertrophied human heart; Refs. 2, 487), myocyte loss (see sect. V ), myocardial damage, and probably DNA repair. In addition, there is evidence based on quantitative morphometry that, at least in human hearts (but it is nearly impossible to experimentally obtained very bulky hearts, discussed in Ref. 187), severe degrees of cardiac hypertrophy are accompanied by cardiocyte proliferation (21, 258). There are also convincing data showing mitotic images in atrial and ventricular cardiocytes during CF in humans and in the region adjacent to the necrotic area in both human hearts and in experimental models of MI (363, 384). With the use of a morphometric approach, the mitotic index for cardiocyte nuclei was rather high in the end-stage failing hearts compared with fetal hearts and was increased in the surviving tissue bordering the acute infarction. However, mitotic images observed during CF may not necessarily be indicative of cell division but could be a prerequisite for polyploidy, multinucleation, DNA repair, and even DNA fragmentation and cell death (363). Compensated cardiac hypertrophy is essentially caused by myocyte hypertrophy and hyperplasia of the nonmuscle cells; nevertheless, adult cardiocytes may not all be terminally differentiated cells, and mitotic divisions of the myocytes may constitute a growth reserve for severely damaged myocardium. Cardiocyte hypertrophy is the result of a sarcomeric reorganization. Dilation of the heart is associated with myocyte lengthening, which is mediated by the generation of new sarcomeres in series resulting in a pronounced enhancement of the length-to-width ratio of the myocytes. The same phenomenon is observed in the rat 2 or 6 days after birth. This would suggest that during CF a fetal gene program involving activation of the cytoskeletal proteins is initiated that regulates the overall shape of the myocyte by preferentially directing lengthening of the cell. In contrast, hypertrophy of cardiomyocytes is the result of the addition of new sarcomeres in parallel (152). Such a hypothesis is favored by the observation made by Samuel and co-workers (393, 394) showing a reversible rearrangement of the microtubular network during the early stage of cardiac overload. C. Clinical Setting and Treatment
Cardiac remodeling after MI is accompanied by a progressive decline in ejection fraction over time, which suggests that CR, in this condition, would be better quantitated by measuring the LV ejection fraction (89, 148). In hypertensive cardiopathy, there is a consistent and graded relation between blood pressure and the degree of concentric ventricular hypertrophy. Systolic and diastolic dysfunction is progressive and is correlated with ventricular dilation. Several groups of clinical trials, using vasoactive drugs (89), Recent clinical trials have demonstrated the existence of two groups of vasoactive drugs: The effects of Because of the development in molecular and cellular biology, we have learned that, in response to mechanical stimuli, both the myocardium and the vascular walls adapt to increased work loads by changes in gene expression. However, nature has not provided infinite possibilities of new genetic programs. If one looks back during development, such programs are usually available. In other words, to know, in a given animal species, what the genetic programs could be reexpressed in a given condition, the best strategy consists of exploring simultaneously the pattern of gene expression during development and in the pathological state. Such a strategy has been successfully applied to mechanical overload and has led to the discovery that, during mechanical overload, only fetal isoforms were reexpressed in this condition (Table 2). In skeletal muscle, hypertrophy due to intensive training or during muscle regeneration is accompanied by the reexpression of several programs, including a fetal and an embryonic program because there are two programs that are expressed during development (442).
Nevertheless, in terms of genetic expression, there are differences between an embryo and patient suffering from arterial hypertension. In the first case, the changes in genetic expression are part of a sequential program. In the second case, the endogenous program is associated with several other programs resulting from complex and variable interactions between hormones, neurotransmitters, and plasma peptides. It is one of the goals of this review to try to separate these two components from each other. A. Methodological Problems
Molecular or cellular studies on CR have raised several specific problems. 1 ) There are several experimental models of cardiac hypertrophy available, such as Goldblatt, aortic insufficiency, or abdominal aortic stenosis that need experienced hands to provide reproducible results and a reasonable (>35%) degree of cardiac hypertrophy (187). Several outstanding biochemical or biological studies have indeed been based on experimental models that gave rise to a cardiac hypertrophy of 20% or less (421). Experimental models of CF are rare, expensive [i.e., aging spontaneously hypertensive rats (SHR); Ref. 47], and sometimes far away from the clinical reality, for example, the tachycardia-induced model which is not accompanied by hypertrophy (523). 2) The world-wide initiation of active cardiac transplantation programs has provided the possibility to obtain human tissue samples from both control and end-stage failing hearts [New York Heart Association (NYHA) stage IV], including ischemic cardiopathy and idiopathic cardiomyopathy, for physiological, molecular biological, and enzymatic use. Control samples can be obtained from donor hearts which proved to be unsuitable for transplantation or from tissue discarded during surgical procedures. End-stage failing hearts always come from patients with end-stage disease, and it is therefore frequently difficult to decide whether the differences reported between experimental animal models and human material are due to species differences or reflect real different stages of the disease. 3) Technical and methodological problems are particularly frequent in this domain; extensive purification procedures may lead to selective extraction (see below a good example concerning the calcium-regulating proteins). For unknown reasons, highly reputed journals accept results based on small series of three to four patients, and even worse accept the utilization of a Student's t-test in such condition (some good examples are given in Tables 5 and 7). 4 ) In any model, including the human heart, the biological results cannot be interpreted without knowing two parameters: the stage of the disease, which is usually rather well documented, and the concentration of the major plasma hormones and peptides. Despite their major importance, the latter have unfortunately been rarely documented (142).
Species specificity is another crucial methodological problem. The molecular determinants of ventricular adaptation and dysfunction are indeed species specific (451). In contrast, an important conclusion of the previous thermodynamic study is that the fundamental physiological process of adaptation is the same, both in rabbits and humans (7). The causes of cardiac hypertrophy may differ from one animal species to another (for example, rats are resistant to atherosclerosis, dogs have a very particular valvular pathology, and CF in humans in the affluent western countries is mostly secondary to hypertension and/or coronary insufficiency). Nevertheless, the most important cause of such a species specificity is the fact that ventricular contraction is regulated by different processes that depend on both the type and the size of the animal. There are species such as rats or rabbits in which the adaptational process is caused by modifications of both the contractile machinery and the calcium-regulating systems (129). In contrast, in humans (at least human ventricle) and guinea pigs, the contractile proteins are unlikely to be modified. The hypothesis is that in the latter group membrane proteins responsible for the calcium movements play a major role in the adaptational process (306, 481). By comparing the results of mechanical studies performed on isolated fresh papillary muscles with those made on isolated skinned muscle fibers (fibers without any membrane structure) from pressure overloaded rats or guinea pigs (481), one can identify two different processes. When maximum shortening velocity (Vmax ) is determined on fresh muscles, it is altered in both animals. In contrast, when muscular contraction is performed on skinned fibers, the shortening velocity is reduced only in pressure-overloaded rat hearts, but not in guinea pigs. This indicates that in the absence of membrane structures, the fundamental process of adaptation is still present in the rat, whereas it has disappeared in guinea pigs. Calculations based on heat production have indicated that in the rabbit ventricle the two mechanisms, namely, the sliding process and calcium movements, play roughly equivalent roles in the improvement of the economy (Fig. 1) (6). Molecular investigations have confirmed such a view, and with the comparison of the ventricles and atria, it has been shown that there is, in addition, a tissue specificity.
B. General Process of Biological Adaptation
Biological adaptation is a general process by which an organ or organism expresses a new genetic program in response to environmental changes that unbalance its thermodynamic status. The expression of such a program has to result in an improved thermodynamic state, which is referred to as adaptive. Mechanical overload of a striated muscle immediately reduces the instantaneous shortening velocity by simply bringing into play the mechanical properties of the myofiber. When the load is greater, the fiber contracts more slowly, and this results in an immediate decrease in the economy (i.e., the number of ATP molecules burned per gram of tension) of the system, since the muscle fiber has adapted to have an optimal economy for a given velocity (153, 158, 435). The adaptational process is dominated by several changes in genetic expression that allow the heart to recover a normal economy. Quantitative modifications lead to cardiac hypertrophy and, according to Laplace's law, will normalize the wall stress. Simultaneously qualitative changes will allow the myocardial fiber to contract more slowly and to recover a normal economy by having a lower Vmax and by using a different force-velocity curve (7, 246, 247). This phenomenon proceeds by trial and error, and the permanent changes in genetic expression simultaneously have, even at the beginning, both beneficial effects in terms of muscular economy and detrimental, or even useless, effects (383). Cardiac failure, when it occurs, indicates the limits of the process of biological adaptation (448). The phenotypic changes (also called phenoconversion) are very similar in a variety of conditions, including fetal development, volume and pressure overload, senescence, or hypothyroidism. This coincidence is not fortuitous, since in each of these conditions the change in genetic expression is due to an external event, and the same program is always used simply because this is the only developmental program available (266). C. Cardiac Hypertrophy
The increase in cardiac mass is the most easily detectable adaptational factor triggered by mechanical overload. Cardiac hypertrophy adapts the heart to the new working conditions by both multiplying the contractile units and reducing the wall stress according to Laplace's law. Following the pioneer studies by Schreiber et al. (406) performed on isolated working guinea pig hearts, a considerable amount of work using radioactive amino acid labeling was performed, and extensive reviews have been published (311, 313, 443, 449). In brief, the activation of protein synthesis is a rapid and rather homogeneous phenomenon. Schreiber et al. (406) showed that total protein synthesis, including myosin, but not myoglobin, and collagen, were activated after 3 h of increased aortic pressure. Hatt et al. (189) were able to demonstrate polysomes in the myocardium 30 min after an acute overload. Following the imposition of a pressure overload, such as aortic stenosis, radioactive labeling of total proteins or myosin remained unchanged for 1 or 2 days, peaked at 5-6 days, and returned to control values within 2-3 wk (304, 370, 527). The accumulation of total proteins or myosin is due to an increased rate of synthesis; nevertheless, the rate of degradation is augmented in parallel, such a paradoxical wasting effect is a general feature in protein metabolism (304, 313). Molecular biological techniques using global approaches such as hybridization curves to evaluate mRNA complexity were unable to detect major differences between normal and hypertrophied hearts (26, 450). Different results were subsequently obtained using more sophisticated techniques such as differential display and specific hybridization procedures and are described in section VIII. D. Energy Metabolism
1. Energetics
I. DEFINITION: LIMITATIONS
Top
Next
References
View this table:
TABLE 1.
Main sources of cardiac remodeling
II. MORPHOLOGICAL AND CLINICAL BASIS
Top
Previous
Next
References
-adrenergic blocking agents (123), converting enzyme inhibitors (CEI) (93, 430), or spironolactone (522) have provided evidence that structural remodeling can be not only attenuated but even reversed and that the treatment of CR and its adverse effects can be targeted solely to the biological function of the myocardium. 
1 -blockers, such as prazosin, which have no effect either on mortality rate or ejection fraction, and drugs such as hydralazine, isorbide dinitrate, and CEI that improve both the mortality rate and the ejection fraction. The first category of drugs has no effect on the progressive development of cardiac hypertrophy and dilation. In contrast, CEI significantly reduces LV mass and volume (90). Therefore, it is now possible to treat CR. In addition, such trials show that vasoactive peptides, like angiotensin II, have both a hemodynamic and a trophic effect. -adrenergic blockade with bucindolol or carvedilol have also been examined in several placebo-controlled studies involving patients with CF and systolic dysfunction due to profound CR (123). In all studies, LV ejection fraction was consistently improved and ventricular volume reduced. These effects appeared after 3-6 mo of continuous therapy. Such a long-term treatment also reverses CR by decreasing ventricular mass and volume. There was also an increase in the sphericity index, which indicates that the ventricle recovers its ellipticity.
III. PHENOTYPIC CHANGES IN MYOCYTES
Top
Previous
Next
References
View this table:
TABLE 2.
Fetal program reexpression during cardiac remodeling
View this table:
TABLE 5.
Calcium-regulating proteins in the failing human heart
View this table:
TABLE 7.
Cardiac receptor densities
View larger version (15K):
[in a new window]
FIG. 1.
Energy transduction in cardiac remodeling. Data are expressed as percent controls. CCH, compensated cardiac hypertrophy due to pulmonary artery stenosis; HF, heart failure due to dilated cardiomyopathy in humans. Force-time integral is measured in isometric conditions. Shortening velocity is calculated in isotonic conditions. ** Significantly different from controls. [Data recalculated from Alpert et al. (7) and Hasenfuss et al. (183).]
2. Energy metabolism
Whether CF is caused by a defect in energy production or a deficit in energy utilization is a continuing debate (290, 340, 449). Of course, it is easy to oppose two extreme conditions, i.e., CF due to acute anoxia and CF occurring a few weeks after massive MI and involving an important loss of contractile material. Nevertheless, these are rather rare situations that are unrelated to CR as it is usually met in clinical practice. The real problem is to know whether, during the compensatory stage, modifications in the cellular apparatus that are responsible for energy production (mitochondria or anaerobic energy pathways) could account for further impairment of myocardial function or, conversely, could CF be more easily explained by a deficit in energy utilization at the level of the contractile machinery. Myoglobin facilitates the transport of molecular oxygen from erythrocytes to mitochondria and plays an important role in myocardial oxygenation. A 40-50% reduction in myocardial myoglobin protein and mRNA has been well documented in various models of congestive CF, including CF in humans (335, 336). Such a reduction would affect the energy flux in the myocardium and could be considered to participate in the energy deficit that is associated with CF. In CCH, mitochondria adapt to the new situation by increasing their number and decreasing their size. The mitochondrial mass increases (188) by activation of mitochondrial DNA replication and by removal of the block in the conversion of DNA D-loops to other intermediates (364). The mitochondria-to-myofibril ratio remains unchanged, but even with a decreased ratio, the efficiency of the organelles would be normal or even improved because fragmentation enlarges the average surface area for oxygen exchange. This kind of morphological adaptation occurs in every model of cardiac hypertrophy, including aorta-caval fistula, which is a model of pure mechanical overload without any neurohormonal modification (188, 189). In vitro studies showed normal or even improved mitochondrial oxygen function and coupling (reviewed in Ref. 449) and, as discussed in section IIID1, heat recovery (which depends of mitochondria activity) is not modified in CCH (Fig. 1). In addition, more recent studies based on 31P-NMR spectroscopy showed no significant differences in either the content or turnover rates of the phosphoryl group in CCH (211, 212). Oxygen and substrate uptake measured in vivo by estimating arteriovenous differences with a catheter in the coronary sinus were unchanged in cardiac hypertrophy (40). Rather ancient studies reported a lactate dehydrogenase isoenzyme shift toward a more skeletal muscle type in humans, with an increase in the amount of M subunits at the expense of the H subunits (374). Several detailed studies on cardiac metabolism in the fully compensated mechanical overloaded ventricle demonstrated an increased glucose utilization and a pronounced enhancement in the activity of several enzymes that control glycolysis (including hexokinase, glycogen phosphorylase, and lactate dehydrogenase) together with a diminished activity of enzymes responsible for ketone body metabolism (including acetoacetyl-CoA synthase). This suggests that there is a preferential utilization of glucose when compared with ketone bodies as competing substrates for the fuel of cardiac respiration (225, 454). There is no doubt that the rate of glycolysis from exogenous glucose is accelerated in cardiac hypertrophy. Recent studies have strongly suggested that this process is regulated at the level of membrane enzymes, since myocardial glycogen metabolism does not change (4). A gene regulatory mechanism involved in the reexpression of the gene encoding medium-chain acyl-CoA dehydrogenase (which catalyzes a rate-limiting step in the fatty acid oxidation) has been identified (388). This mechanism involves reactivation of fetal transcriptional control via members of the Sp and chicken ovalbumin upstream promoter transcription factor (COUP-TF)/erbA-related protein families of transcription factors and is likely to be also involved in the reexpression of the skeletal actin isoform. In CF, several rather ancient studies had initially suggested that myocardial ATP content was normal (357, reviewed in Ref. 449), even in humans (84), and that energy depletion was a consequence and not the cause of CF (357). More recent studies have contradicted these findings, and CF has been reported to result in the depression of mitochondrial function and a modification in several of the mitochondrial enzymes including malate and glutamate dehydrogenase (349). Other studies have also revealed a decreased oxidation of fatty acids in the failing heart (512). These alterations are probably a consequence of the neurohormonal imbalance and ischemia that are associated with CF (Table 3). It has also been suggested that the failing heart is "energy starved" (229) and that its capacity to synthesize ATP via the creatine kinase system is impaired. The pioneer studies of Ingwall and co-workers (211, 212, 321) performed in vivo using 31P-NMR spectroscopy have shown a decrease in the creatine kinase reaction and of its products, namely, ATP and phosphocreatine, and an isoenzymic shift resulting in an increased amount of the B subunit (the fetal isoform, see Table 2). The myocardial creatine phosphate-to-ATP ratio can be more directly assessed in situ on beating hearts using NMR spectroscopy. This ratio is reduced in the intact residual myocardium after MI both in rats (8 wk after ligation) and in dogs with documented ventricular dysfunction (11 mo after infarction) (286, 323). Calculations of the rates of synthesis of ATP showed that phosphoryl transfer via creatine kinase does not limit contractile performances during baseline conditions, but performance will be impaired in extreme conditions such as acute stress (323).
|
It is evident that there is an energy deficit in the failing heart; however, similar isoenzymic shifts in creatine kinase (in rat) and a creatine deficit and changes in creatine kinase activity (in cardiomyopathic hamsters) have also been described during CCH (323, 465, 517). Hence, it is impossible to conclude whether such a deficit originates early during the development of the adaptational process or is only created or aggravated by other external factors during CF.
E. Membrane Proteins
1. Ion channels and currents
The increased Q-T (and Q-Tc, which is the Q-T interval corrected for heart rate) interval duration on electrocardiogram (ECG) and the enhanced action potential duration on isolated cardiomyocytes are well-documented characteristics of the hypertrophied heart and cardiocyte (Table 4). The duration of the action potential duration depends on the activity of several ion channels and can increase either when an outward current is depressed or when an inward current is enhanced. The modifications of these current activities could result from either functional or structural changes. At present, there are enough data to support the second hypothesis. Structural changes are mainly due to modifications in the expression of genes encoding ion channels. Acquired modifications of the repolarization time probably reflect the adaptational response compensating for the mechanical overload such as the isomyosin shift. Recent studies on an inherited monogenic multiallelic disease, the long Q-T syndrome, have demonstrated that a prolonged action potential can indeed be caused by several mutations located on ion channel genes including a subunit of the sodium channel, and HERG, which encodes the-subunits of a potassium channel. The electrophysiological pattern observed during CCH and CF has recently been reviewed (18, 181, 447, 448).
|
1 ) The slow inward current ICaL contributes to the plateau phase and could be involved in lengthening the action potential provided that the corresponding current density is increased. The changes in inward current are strongly correlated to the density of the dihydropyridine binding sites, and, in CCH, both the current density and the concentration of dihydropyridine binding sites remain unchanged (233, 307, 400). The same results have been observed in the senescent rat heart (491). These results suggest that the expression of the corresponding genes is sensitive to mechanical stress and is activated in proportion to the degree of hypertrophy which finally will result in an unchanged current density. Nevertheless, such an unchanged density is also accompanied by isoform modifications. An alternative splicing in the third membrane-spanning region of motif IVS3 of the
1-subunit (which forms the channel) occurs in the L-type calcium channel. In the rat, this region undergoes developmentally regulated mutually exclusive splicing of two adjacent exons (called IVS3A and IVS3B). IVS3B is predominant in normal adult rat heart, whereas both isoforms are equally expressed in both fetal tissue and noninfarcted hypertrophied ventricle 21 days after MI (154).
During CF, the situation could be different. Electrophysiological studies in humans have shown the same unchanged density as in experimental models of compensatory hypertrophy (34). Nevertheless, there are data suggesting that in some models of CF the density of the calcium channels is modified (discussed in Refs. 181, 457, 489, 490) (Table 4).
2) Sodium/calcium exchange creates an inward depolarizing current. Its activation could create a long-lasting slow inward current and contribute to the lengthening of the action potential. As explained above, for the moment, there are arguments suggesting that such a current is prolonged as a consequence of the prolongation of the calcium transient. Whether or not the rather well-established increased sodium/calcium exchanger density could contribute to the electrophysiological abnormality is purely speculative.
3) Potassium currents are outward currents that accelerate the repolarization and include a transient early current (Ito ), a delayed current (IK ), a background current (IK1 ), and several currents that are activated by ATP or muscarinic effects. The major defect in CCH in rats is a pronounced depression of the peak Ito after normalization to cell surface area. Such a decrease is proportional to the degree of cardiac hypertrophy and has been confirmed by several laboratories both in humans and in different experimental models (29, 35, 98, 233, 362) (Table 4). The peak Ito density is also reduced during aging (491). Such a reduction is likely to participate in the reexpression of the fetal program, since, at least in the human atria, Ito is developmentally regulated and its density is twice that found in the atria of young subjects (100).
The relative contribution of the cloned potassium channels to cardiac function has only been partly elucidated (106). A major advance concerning CR was the discovery, by El-Sherif and co-workers (155), that the electrophysiological modifications of the two components of Ito, Ito-s and Ito-f, reflect changes in the genetic expression of the corresponding potassium channels at the level of both the protein and the mRNA. The expression of Kv1.4 and Kv2.1, which, for these authors, are genes encoding Ito-s, and that of Kv4.2, which is the gene encoding Ito-f, are decreased by 60 and 54%, respectively, suggesting that the diminution of Ito and the corresponding lengthening of both action potential and Q-T interval are in fact transcriptionally regulated and participate in the adaptational process.
In feline models of cardiac hypertrophy, the delayed rectifier outward current IK is also reduced, whereas the background inward rectifier current IK1 is increased. During CF in humans, convincing reports have shown that the prolongation of the duration of the action potential is, at least in part, caused by Ito, since 3 mM 4-aminopyridine, a specific inhibitor of Ito, does not entirely restore a normal duration (29, 98). Moreover, there is a coordinated decrease in Ito and IK1. The delayed rectifier is hardly detectable and plays only a minor role in the human cardiocyte (35).
A major advance in our understanding of the modifications in ion channels was made by Nuss et al. (334) using gene transfer technology and the canine tachycardia-induced model of CF. In this model, isolated cardiocytes have prolonged action potentials that are due to a 66% reduction in Ito. A genetic construct was made using an adenoviral shuttle vector and the Drosophila Shaker B (ShK ) gene as a prototype of the voltage-dependent class of potassium channels. Failing cardiocytes were infected with the genetic construct. Two to three days after infection by ShK, the action potentials were dramatically shortened in a dose-dependent manner. In other words, it is possible to correct the increased repolarization time by gene transfection with potassium channel genes, and in so doing to increase the contraction velocity which then returns to control values. This is the first demonstration that one of the primary events in the adaptational process is located at the level of ion channels.
4 ) At least three currents or proteins specific for the sinus node have been found in overloaded ventricles, suggesting that the ventricles are able to acquire some degree of automaticity. 1 ) Cerbai and co-workers (72, 73) have recently shown the spontaneous occurrence of If (the main current responsible for the spontaneous depolarization of the pacemaker) in isolated ventricular myocytes from senescent SHR. A good correlation exists between the duration of the pressure overload and the number of cells in which this current is found. 2) Reexpression of ICaT, a calcium channel insensitive to calcium blockers and specific for the sinus node, has been demonstrated in the hypertrophied ventricle (333), although there is indirect evidence that this channel is not functional (15). 3) More recently, we have found that the 3-isoform of the sodium pump, which is probably a marker of the conduction system (520), is reexpressed in the overloaded rat ventricle (77). 4 ) In addition, an increased susceptibility to pacemaker-like activity has been demonstrated in human trabeculae from failing hearts during superfusion with a modified Tyrode solution (482).
The reappearance of If, ICaT, and the 3-isoform of the Na+-K+-ATPase is suggestive of the reinduction of a fetal program (such a program is also expressed in the adult conduction system). Attempts to trigger automaticity in pure models of CCH by increasing the external calcium concentration have, however, been unsuccessful (15, 72, 73).
5) Connexins are components of the gap junctions; they are confined to intercalated disks and constitute the channels that establish cell-to-cell electrical coupling. There is evidence of a rather complex rearrangement of connexin distribution in CR both in human and experimental models with a shift in genetic expression from connexin43 (the most abundant ventricular connexin isoform) to connexin40 (which is normally only expressed in atria and in the conduction system and has a greater unitary conductance than connexin43 and -45) (24, 348). Connexin40 is increased in CR, and the accumulation of this molecule may explain the increased susceptibility of the hypertrophied heart to arrhythmias.
2. Control of intracellular pH
There is little information available concerning the control of intracellular pH (pHi ). The Na+/H+ antiporter is a polymorphic transmembrane protein that participates in the regulation of pHi. Pressure overload in the rabbit results in a twofold increase in the Na+/H+ antiporter isoform NHE-1 mRNA 3 days after the beginning of the load, and this level remains constant for 2 wk. It was proposed that this activation is involved both in the regulation of pHi and as an intracellular signaling system (459). The same type of activation has been observed following ischemia (139). In contrast, the activity of the exchanger is reduced in the diabetic heart in response to alterations in intracellular calcium (250). As yet, there is no additional information concerning the other main determinants of pHi, namely, the intrinsic buffering power, the Na+-HCO3 cotransporter, and the Cl
/HCO3 exchangers in CR.
3. Calcium-regulating proteins
Intracellular calcium participates both in the control of the contractile process and in the coupling between mechanical activity, energy metabolism, and most likely protein synthesis. The free cytosolic intracellular calcium in cardiac tissue, as in every tissue, has two origins: the extracellular space and a closed internal store, called the sarcoplasmic reticulum (SR). From a purely molecular point of view, the intracellular calcium homeostasis and calcium transient depend on various proteins that are responsible for both the input and output of calcium into these two compartments. Energy-dependent systems are required to release calcium from the cytosolic space, namely, the Ca2+-ATPase of SR; phospholamban and the couple formed by the Na+/Ca2+ exchanger and the Na+-K+-ATPase, the latter, also called the sodium pump, provides energy for calcium transfer at the external membrane level. In contrast, input of calcium into the cytosol is controlled by gated systems, namely, the ryanodine receptor of the SR, which is responsible for the calcium-induced calcium release phenomena, and the calcium channels for the external membrane (reviewed in Refs. 14, 494) (Fig. 2).
|
A) SARCOPLASMIC RETICULUM. An altered calcium uptake by the SR of failing myocardium has been reported several times over the past 25 yr (216, 254, 436). However, convincing evidence for this has only recently been obtained by routine use of molecular biological techniques. A decreased concentration of the SR Ca2+-ATPase in CR, regardless of the etiology or model, has now been well documented. To date, the diminished expression of SERCA2a, the gene encoding the cardiac isoform of the enzyme, is considered to be one of the best markers of chronic cardiac overload. The reduction in the concentration of the enzyme is progressive and is linked to the progression of cardiac hypertrophy, which in turn is correlated with the appearance of symptoms of CF, at least in the experimental models, which by definition are untreated animals. A reduced level of ATPase has therefore been proposed as a marker of CF (132, 232). Several different investigations have demonstrated (262) that 1 ) the cardiac isoform is unique (239, 404) and does not shift to another isoform during cardiac overload. 2) Both mRNA and protein concentrations decrease in proportion to the degree of cardiac hypertrophy. This suggests that the corresponding gene is not activated by mechanical overload that results in a dilution of both the corresponding mRNA and protein. 3) The oxalate-stimulated calcium uptake, an essential property of the SR, is impaired. This alteration is due to a parallel reduction in the functionally active SR Ca2+-ATPase molecules, as determined by quantitation of the calcium-dependent phosphorylated intermediate (25). 4 ) The diminution can be correlated with the force-frequency curve (184), the cardiac index (294), and the levels of brain and atrial natriuretic factors (13, 456) and Na+/Ca2+ exchanger (433). In a comparative study, Schwinger et al. (415) have shown that the purification procedure may normalize the observed modifications and has suggested that additional factors could regulate the activity of the SR Ca2+-ATPase, a suggestion which has also been made for the ryanodine receptor (389) (see below). Such modifications have also been reported in experimental models of CCH (25, 283, 318) and in end-stage CF in humans (Table 5). Because SR Ca2+-ATPase activity is almost absent in 14-day fetal rat hearts and increases substantially at the end of the fetal life and in the early postnatal period to reach a maximum 4 days after birth (264), it is generally accepted that the diminished concentration observed during cardiac overload participates in the reexpression of the fetal program (Table 2). Adenovirus-mediated expression of a SERCA2 transgene can reconstitute endogenous SERCA levels and activity obtained in cultured neonatal cardiocytes treated with phorbol 12-myristate 13-acetate. It is conceivably possible that in the future it will be possible to correct the above modification using gene transfer technology (156). Phospholamban and calsequestrin have also been investigated both in experimental models and in humans. A drastic diminution of phospholamban has been reported in the human heart with end-stage failure (13, 14, 297) and in experimental cardiac hypertrophy (232, 283, 318). In both cases, the fall more or less parallels that of the SR Ca2+-ATPase and might play a role in the attenuated response of the contractile apparatus to catecholamine. The targeted cardiac overexpression or knockout of phospholamban has confirmed the important role of such a protein in regulating both relaxation and contraction (224, 270). In contrast, calsequestrin remained unchanged, except in the rabbit model (283) (Fig. 2). Two forms of intracellular calcium-release channels are expressed in the heart: the ryanodine receptor and the inositol 1,4,5-trisphosphate (IP3 ) receptor (66, 157). The situation for the ryanodine receptors is rather complicated. In experimental CCH, in rat, rabbit, guinea pig, and ferret, the ryanodine receptor binding site density is reduced and parallels the Ca2+-ATPase concentration, with a more severe alteration being present in ferret, rabbit, and guinea pig than in rat (283, 322, 365). In end-stage failure in humans, several investigators, including those from our laboratory, dealing with a sufficient number of experiments, found a decreased mRNA content and an unchanged protein level (Table 5). In addition, binding studies using radioactive ryanodine revealed a twofold increase in the number of high-affinity sites (231, 389) with an unchanged dissociation constant, suggesting that during CF additional regulatory factors affect the ryanodine binding properties. Supporting the latter hypothesis are the differences in the gating properties of the caffeine-induced calcium release observed by D'Agnolo et al. (103) and by Kim et al. (231), the decreased ryanodine calcium accumulation into SR in cardiomyopathic human hearts (329), and the fact that ryanodine binding is affected by the purification procedure (329, 389). Another possibility is that the downregulation of the ryanodine receptors would be compensated, at least in humans, by an upregulation of the IP3 receptors (157). To summarize, in CCH, the proteins responsible for calcium movements in the SR are downregulated in parallel, suggesting that at this stage the activity of SR is reduced for adaptational purposes. Nevertheless, at this particular level, calcium homeostasis is maintained. In contrast, during end-stage CF, it is proposed that additional factors, such as hormones, modify the landscape, which results in a normal ryanodine receptor density but, still, a downregulation of the SR Ca2+-ATPase (Fig. 2). B) CALCIUM TRANSPORT THROUGH THE SARCOLEMMA. During CCH, the concentrations of both total (i.e., the dihydropyridine sites) and active calcium channels (representing ~5-8% of the total channels; Ref. 413), i.e., the calcium inward current density ICaL, remain unchanged and parallel the degree of cardiac hypertrophy. This suggests that calcium entry from the extracellular space is normal (285, 400). This process is not species specific, and the total number of calcium channels parallels the degree of cardiac hypertrophy in at least two different animal species (rat and guinea pig) that have different mechanisms of calcium metabolism (361). As explained above, additional isoform modifications occur which, for the moment, have no electrophysiological significance (154). Different results have been observed in end-stage CF in humans. Beuckelmann et al. (34) found in CF the same type of results as those obtained in CCH, namely, an unchanged density of ICa. In contrast, others have reported a pronounced and parallel reduction in both dihydropyridine receptors and in the level of mRNA encoding the
1-subunit of the calcium channel (Table 5). Interestingly, a reduction in the calcium channel concentration has also been observed in intact myocytes from chick embryo ventricles after a 4-h exposure to 1 mM isoproterenol, in parallel to a downregulation of the -adrenergic receptors (281), thus suggesting that the decrease in the number of the calcium channels observed during CF is in fact a consequence of the neurohormonal disorders which accompany the hemodynamic failure. Both the receptor level and related voltage-sensitive calcium channels increase in the Syrian cardiomyopathic hamster; nevertheless, such an alteration is likely to have a genetic origin and has simultaneously been observed in heart, brain, and skeletal and smooth muscles (489).
Calcium output is mainly regulated by a functional duo composed of the Na+/Ca2+ exchanger and the Na+-K+-ATPase. The first is responsible for the calcium release, whereas the second provides energy to the first. Both are modified in CCH, suggesting that calcium homeostasis may be unbalanced at this level.
Several papers have shown alterations in the activity of the Na+-K+-ATPase, the so-called sodium pump (360, 516). More recent studies using molecular biological techniques have shown modifications in the sodium pump that are species specific. The Na+-K+-ATPase is composed of two subunits, and . The -subunit is polymorphic, and the normal adult rat heart contains two different isoforms, 1 (1) (75%) and 2 (2) (25%). A third isoform, 3, is mainly embryonic but is also present in the conductive tissue of normal adult hearts (520). This enzyme has two important properties, namely, its affinity for sodium and for ouabain. In rats, 3 has a low affinity for sodium and a high affinity for glycosides. In contrast, 1, which has a higher affinity for sodium, has a low affinity for ouabain and is likely to be responsible for digitalis toxicity (76, 79). Studies during CCH in the rat using a highly purified membrane preparation (276) have shown a complex pattern (76, 77, 249) that combines 1 ) an unchanged specific activity, 2) an increased density in the high-affinity sites for ouabain, and 3) a slowing of the dissociation rate constant (k1 ) for ouabain (Table 6). Such modifications reflect an isoenzymic shift of the -subunit from the adult form, 2, to the fetal form, 3. The 3 form has a much lower affinity for sodium than 2, and this shift finally results in a lower affinity of the overall enzyme for sodium. This would mean that, in the rat, the sodium pump should be less active in cardiac hypertrophy.
|
The situation is clearly species specific and is different in humans. In the failing human heart, the activity of the Na+-K+-ATPase is reduced, and it is less sensitive to sodium (417). Molecular biological studies on the
-subunit of the enzyme, made on a limited number of cases, have found a shift from the 1- to the 3-isoform. Because the human heart possesses an 1-subunit with a low affinity for sodium, one can conclude that in humans, as in rats, the final result is a diminution of the affinity of sodium for the enzyme and then a slight accumulation of sodium below the membrane surface (Table 6).
The activity of the Na+/Ca2+ exchanger has been a controversial issue. Pioneer studies in rats using isolated membrane vesicles during both cardiac overload and in the senescent heart have concluded that the activity of the exchanger was depressed (113, 114, 176, 198). Such results have been contradicted by investigations performed using new molecular tools. Such contradictions may be due to the fact that the membrane vesicles are heterogeneous in size and that the procedure used for their isolation does in fact select different vesicles in the different experimental groups. It has now been generally accepted that the molecular density of the Na+/Ca2+ exchanger is increased during both compensatory hypertrophy and in end-stage CF in humans (433, unpublished data) (Fig. 2). The increased expression of the exchanger can be correlated with the reduction in the Ca2+-ATPase of the SR (433). The Na+/Ca2+ exchanger is an electrogenic transporter that generates the current INa-Ca. The corresponding potential ENa-Ca becomes more and more positive as the intracellular calcium concentration increases. For this reason, the exchanger generates an inward depolarizing current during almost all of the duration of the action potential. The prolongation of the calcium transient will lengthen the INa-Ca, and at present, it is thought that the activity of the exchanger is both lengthened and attenuated by the slightest increase in intracellular sodium.
Very few studies have been carried out on the Ca2+-ATPase of the external membrane, which is thought to remain unchanged during cardiac hypertrophy (114).
C) CALCIUM TRANSPORT WITHIN THE CELL. Calmodulin is a four-calcium site binding protein that acts as a cofactor in numerous reactions. A reduction in the amount of calmodulin mRNA has been reported in failing human hearts (222). Calmodulin also has a role in myocardial protein synthesis and cell proliferation as shown using transgenic technology (164), and the decreased calmodulin expression may also have a significance in terms of growth signaling.
4. Cardiac receptors
A rather large number of receptors have been identified in the adult human heart, including seven domain membrane receptors, such as the adrenergic, angiotensin II, and muscarinic receptors, as well as nuclear DNA-binding receptors such as aldosterone, thyroxine, and glucocorticoid receptors. Numerous changes in the expression of most of the known myocardial receptors as well as modifications in the functioning of the corresponding signal pathways have been reported to occur during CR both during compensation and failure (Table 7). Such modifications most likely have a dual origin and reflect both an adaptation to the mechanical stress, i.e., a heterologous regulation, and additional homologous regulations in response to changes in the plasma levels of the corresponding agonists. The concentrations of both1-adrenergic and adenosine A1 receptors are unchanged in CF (1, 45, 58); nevertheless, it has been proposed by Simpson, who discovered the direct hypertrophic effect of 1-adrenergic agonist (424), that the 1C-adrenergic receptor subtype can be induced by a variety of hypertrophic agonists, including catecholamines and endothelin-1 (378).
A) SS-ADRENERGIC AND MUSCARINIC RECEPTORS AND THEIR TRANSDUCTION SYSTEM. The positive inotropic and cAMP-elevating effects of -adrenoceptor agonists and phosphodiesterase inhibitors are attenuated in the hypertrophied nonfailing and failing heart and also during senescence. The inotropic response to calcium, dibutyryl cAMP, or ouabain is unmodified (44, 56, 63, 81, 133, 400). During CF, isoproterenol is unable to compensate for the excitation-contraction uncoupling (160). The isoproterenol-induced increase in the calcium transient is maintained (37), suggesting that there is a defect located "down the road" (166). The regulation of receptor density is a species-specific phenomenon (99).
Following the pioneer studies of Bristow and co-workers (56, 57), the cardiac -adrenergic receptors have been extensively investigated (reviewed in 59) (Table 7), and two situations have been identified. 1 ) Studies on CCH have shown that, in spite of normal plasma levels and depressed myocardial catecholamine content (147), the -adrenergic receptor density is decreased by 29%, whereas the 1/2-adrenergic receptor density remains unchanged. In addition, competition curves with isoproterenol have revealed two sites, one with a high affinity and the same inhibition constant (Ki ) (2-8 nM) as in normal hearts, and the other with a low affinity for the agonist (Ki 6-13 µM), which represents an unusual subpopulation of uncoupled receptors (81, 275, 305). The corresponding mRNA are decreased in parallel (308), whereas the total number of receptors per heart remains unchanged, suggesting that the corresponding gene is not activated by mechanical stress and that the changes in receptor density are regulated. 2) In failing hearts, even in humans, the elevated circulating catecholamine levels induce a homologous downregulation of the 1-adrenergic receptor, and this creates an additional decrease in the receptor density that is superimposed onto the heterologous downregulation occurring during the compensation. The downregulation that occurs during CF is specific for the 1-adrenoceptor subtype and, at least in dilated cardiomyopathy, does not affect the 2-subtypes. Such a selective subtype regulation is controversial as far as ischemic cardiopathy is concerned and does not exist in mitral valve disease (59). The downregulation of the -adrenergic receptor is a complicated dose- and time-dependent phenomena and requires a preliminary phosphorylation of the receptor through a specific -adrenergic receptor kinase whose expression is reduced in CR (475). Targeted overexpression of this enzyme using transgenic technology attenuates the isoproterenol-induced increase in cardiac contractility (236). 3-Adrenoceptors are present in the human myocardium, but for the moment, their pathophysiological role has not been documented (149).
Reflex bradycardia, in response to baroreflex hypertension, is attenuated in CF, which indicates an additional defective control at the parasympathetic level in this disease. Muscarinic receptors and mRNA are also downregulated in CCH (275, 308), and, at least in the rat, the muscarinic-to-1-adrenergic receptor ratio remains unchanged, suggesting that they reflect a compensatory mechanism more than a homologous downregulation. Pressure overload accompanied by CF in dogs lowers the muscarinic receptor density by 36%. This loss is rather specific for the high-affinity population and is associated with a depressed functional efficiency of the muscarinic system (478). However, in this article, receptor density was measured on a highly purified membrane preparation. At present, the situation is far from being clear, and the previous results have been contradicted by Fu et al. (144) and by Böhm et al. (45), who found normal muscarinic receptor content in the failing heart of rats after MI.
One of the most prominent features of CF in humans, which has been confirmed by three different laboratories, is the existence of a 35-40% increase in the level of the Gi subunit, whereas Gs remains unchanged. Molecular biological studies have revealed that this augmentation is due to a 75% increase in Gi-2 (in the human heart Gi-1 is almost absent and Gi-3 is unchanged) (45, 125, 132, 324). Such an alteration could account for the discrepancies between changes in adrenoceptor density and the adenylate cyclase-stimulating effects of agonists. Prolonged infusion of isoproterenol (2.4 mg·kg1·day) not only reduces the -adrenoceptor density and the adenylate cyclase activity, but it also increases the cardiac level of Gi-2. On the basis of the studies of isolated cardiomyocytes, it is now generally accepted that there is a transcriptional cross-regulation of G protein pathways. The increased level of this particular G subunit in CF is most likely to be caused by elevated levels of plasma catecholamines (125, 126).
Muscarinic regulation may have important effects on cardiac function in CF. Such a question has recently been addressed by Newton et al. (325), who found in the failing human hearts, and not in controls, a negative effect of muscarinic stimulation on myocardial relaxation, and by Eschenhagen et al. (125), who showed in rats that chronic treatment with carbachol favors the occurrence of isoprenaline-induced arrhythmias.
The adenylate cyclase activation by isoproterenol, 5'-guanylylimidodiphosphate, and forskolin is reduced, and this reduction is accompanied by a comparable reduction of the isoproterenol and forskolin activation of myocardial contraction (44, 58, 81, 126). Recent investigations using molecular probes specific for adenylate cyclase isoforms showed in CF a decreased type VI adenylate cyclase isoform, which is the minor isoform in the rat heart; such a diminution may explain in part the depressed activity (127).
The decreased 1-adrenergic receptor density protects the heart against acute stress and, in CCH, participates in the overall process of cardiac adaptation by attenuating the inotropic effects of catecholamines. As such, noninduction of the genes encoding this system has the same significance as noninduction of the SR Ca2+-ATPase or the isomyosin shift to the slow isoform V3.
B) ANGIOTENSIN II RECEPTORS. Angiotensin II receptors (ATR) are expressed at a rather low density in the heart, including the human heart, as compared with the -adrenergic receptors (Table 7). In rats, ATR predominate in the atria and in the conduction system (391). At present, three different receptor subtypes have been identified: ATR1a, ATR1b, and ATR2. The ATR1 are responsible for both the vasoconstrictive and inotropic effects of the peptide. Trophic effects differ according to the target cell and are mediated by the three subtypes through multiple pathways, and the exact role of ATR2 has not been defined (reviewed in Refs. 372 and 387). The normal rat and human heart contains 30-60% ATR1 and 70-40% ATR2 (330, 371). In humans, ATR1 are located both on myocytes and on nonmuscular cells (17, 379).
Cardiac remodeling following experimental MI induces changes in ATR as initially shown on isolated cardiocytes by Meggs and co-workers (291, 525) (Table 7). Reactive hypertrophy was accompanied by almost a twofold increase in ATR1 density, which parallels the depression in the velocity of myocyte shortening and relengthening. Using competitive PCR and binding assays, Nio et al. (330) analyzed the receptor subtypes and demonstrated that MI in rats causes a two- to fourfold increase in ATR1a and ATR2, with no changes in ATR1b, both in the infarcted and noninfarcted areas. Nuclear run-off showed that these changes are transcriptionally regulated. Such an augmentation occurs soon after acute infarction in the zone of acute ischemia and then remains located in the scar tissue and noninfarcted fibrotic area (439). Therapy with an ATR1, but not with an ATR2, antagonist reduces this increased expression.
Results concerning CCH are more controversial as shown Table 7, mainly because these studies have been performed on very limited number of experiments. In several reports, except two (267, 513), models of CCH, including SHR, renal hypertension (441), and abdominal aortic stenosis (128), result in substantial increases in ventricular ATR with an unchanged ATR1-to-ATR2 ratio (Table 7). Cardiac hypertrophy is also partially prevented by type 1 receptor blockade (128).
The first demonstration that a pronounced loss of ATR1 occurs in end-stage CF but not in CCH was made by Regitz-Zagrosek et al. (371). It was subsequently shown that the decreased receptor density was accompanied by a parallel diminution in the mRNA levels. In addition, ATR1, but not ATR2, was significantly altered (17, 192), as in experimental MI. The drop in ATR1 density is correlated with a decrease in 1-adrenergic density, suggesting that the two phenomena have a common origin and may be related to the increased plasma levels of the corresponding agonists (17).
F. Contractile Proteins
1. Isomyosin shift to V3
The myosin molecule is composed of a pair of myosin heavy chains (MHC) and two different pairs of myosin light chains (MLC), MLC1 and MPLC2. In the rat heart, there are three isomyosins: V1, V2, and V3, all of which possess the same pairs of MLC, MLC1v, and MPLC2v, but which differ by their MHC composition, in V1, in V2, and 
, increased activity (as compared with controls); 
, decreased activity; =, activity unchanged; SR, sarcoplasmic reticulum; SL, sarcolemma; Gs