Austin Flint, 1870 [1]

Heart failure is traditionally viewed as a clinical syndrome in which the impaired ability of the heart to pump blood decreases cardiac output and increases venous pressures. This focus on the paradigm of organ physiology, however, fails to emphasize what I believe is the most important clinical problem in patients with heart failure—that of decreased life expectancy, which averages less than 5 years even in patients with moderate heart failure [2]. Advances in a second paradigm, that of cell biochemistry, have provided a detailed understanding of the calcium cycles responsible for excitation-contraction coupling and relaxation and of how abnormalities in these cycles can impair the performance of the failing heart [3]. Recent evidence that growth abnormalities, which accompany hypertrophy of the overloaded myocardium, may play an important role in the deterioration of patients with heart failure [4, 5] has shifted our focus to a third paradigm, that of gene expression (molecular biology) [6].

One way to view the overloaded, failing heart is by analogy to an automobile that has begun to break down while pulling a heavy trailer. It is becoming apparent that the problem in patients with heart failure is not simply impaired performance of the engine (abnormal organ physiology) or even malfunction of the engine's components (abnormal cell biochemistry). As described below, recent clinical studies have provided surprising and, in some cases, counterintuitive findings that are not readily explained by the organ or cell paradigms. Instead, these results suggest that heart failure is accompanied by abnormalities in gene expression analogous to the replacement of normal engine parts with abnormally brittle components. Accordingly, efforts to understand the progressive deterioration and long-term therapeutic response of the failing heart are shifting from the more obvious organ and cellular abnormalities to as-yet poorly understood disorders of cell growth and proliferation.

Response of the Heart to Long-Term Overload

The importance of cardiac hypertrophy in the response to overload was noted in the middle of the last century by Austin Flint [1] who postulated that, like skeletal muscle enlargement in athletes, hypertrophy compensates for hemodynamic overloading of the heart. Whereas Flint attributed the eventual failure of this adaptive response to worsening of the primary cause of the overload, such as valvular insufficiency, Osler [7] suggested that weakening and degeneration of the hypertrophied heart muscle caused the progressive failure of the chronically overloaded heart.

Direct evidence for Osler's hypothesis—that changes in the hypertrophied myocardium contribute to deterioration of the overloaded heart—was provided by the pioneering studies of Meerson [8], who examined the natural history in experimental animals after surgical coarctation of the aorta. Meerson found that although the acute heart failure that immediately followed aortic constriction was alleviated by left ventricular hypertrophy, premature cell death in the hypertrophied, overloaded heart caused the animals to die of progressive heart failure. This progression is caused partly by a vicious cycle in which cell death in the overloaded heart adds further to the overload on surviving cardiac myocytes. Evidence that this progression from adaptive to maladaptive hypertrophy is not caused simply by the increased load was provided by Grossman and colleagues [9], who found that the hypertrophic response to aortic stenosis and insufficiency initially normalizes left ventricular wall stress. These findings indicate that the maladaptive response that follows a long-standing overload involves more than the progressive increase in the "causes," as proposed by Flint [1]. Instead, the ability of long-term overloading to cause myocardial cell death and cardiac fibrosis indicates that long-standing cardiac hypertrophy is accompanied by a progressive, eventually lethal growth abnormality that I have called a cardiomyopathy of overload [4].

Treatment of Heart Failure

As recently as a decade ago, when heart failure was viewed largely as a disorder of organ function, the major goal of therapy was to reverse the adverse effects of the body's attempts to compensate for the decreased cardiac output caused by the failing cardiac pump (Table 1). Thus, the mainstays of therapy were diuretics, which act on the kidney to counteract salt and water retention, and vasodilators, which alleviate excessive afterload by relaxing arteriolar resistance vessels. Except for the cardiac glycosides, treatment was not directed to the heart itself.

The significance of prognosis as a major problem in patients with heart failure led, in the 1980s, to several long-term clinical trials, many of which have generated counterintuitive findings that are leading us to reevaluate the pathophysiologic mechanism of heart failure. As noted below, powerful inotropic agents (which intuitively seemed to be logical therapy for patients who have decreased cardiac pumping) actually increased mortality, whereas some negative inotropic drugs (when administered for several months) were found to improve pump function and symptoms. Equally striking—and equally counterintuitive—is recent clinical experience with vasodilators. Because of evidence that the failing heart is in an energy-starved state [10], it was expected that vasodilator drugs, which not only increase cardiac output but also decrease cardiac energy expenditure, would improve survival in patients with heart failure. Yet, only two classes of vasodilators—the converting enzyme inhibitors, and the combination of nitrates and a direct-acting arteriolar dilator—have been found to prolong life in these patients. Unexpectedly, several recent clinical trials have shown that other vasodilators, although they "unload" the failing heart, accelerate deterioration and worsen prognosis in patients with chronic heart failure.

Positive and Negative Inotropic Agents

Although drugs that increase myocardial contractility are clearly of value in the short-term treatment of acute cardiac decompensation, long-term inotropic stimulation of the failing heart can be harmful. Adverse effects, notably arrhythmias, have been especially prominent when agents that increase cellular cyclic adenosine monophosphate (AMP) levels, either by inhibiting its breakdown (phosphodiesterase inhibitors) or increasing its production (ß-adrenergic agonists), were given in inotropic doses on a long-term basis to patients with heart failure [11-18]. Even the intermittent administration of dobutamine, which during the short term improves symptoms, has been reported [19] to decrease long-term survival in these patients. Conversely, ß-adrenergic blockers, which have negative inotropic effects that can worsen heart failure symptoms acutely, have been found to improve symptoms and left ventricular function after long-term administration to patients with heart failure [20-23].

The ability of digoxin to improve symptoms in patients with heart failure has recently been confirmed in two trials [24, 25] where this drug was withdrawn, but evaluation of the effects of this drug on survival must await the results of a large clinical trial now under way and sponsored by the National Heart, Lung and Blood Institute and the Veteran Affairs Cooperative Studies Center. The cardiac glycosides, which are often viewed simply as positive inotropic agents, have important central actions that increase parasympathetic tone and inhibit sympathetic outflow in patients with heart failure [26, 27]. A role for these central effects is suggested by evidence that the most prominent hemodynamic effect of digoxin withdrawal is an approximately 15% increase in heart rate [24, 25].

These seemingly paradoxical long-term consequences of the administration of positive and negative inotropic agents were not entirely unexpected because of their effects on the energetics of the failing heart [28, 29] and because positive inotropic agents are arrhythmogenic [30]. More surprising, however, is evidence that many vasodilators, despite the fact that when given in the short term they increase cardiac output and decrease cardiac energy expenditure, reduce long-term survival in patients with heart failure.

Vasodilators

A solid rationale exists for using vasodilators to "unload" the failing heart. Indeed, it was for this reason that nitrates were introduced more than a century ago to relieve angina [31], and the ability of vasodilators to alleviate acute pulmonary edema has been known for more than 40 years [32]. Although vasodilators were introduced to relieve symptoms of chronic heart failure in the 1970s [33], it was not until 1986 that the combination of hydralazine and isosorbide dinitrate became the first therapy shown to prolong survival in these patients [34]. The following year, the converting enzyme inhibitors were found to have an even more dramatic effect to improve prognosis in patients with severe heart failure [35]. Subsequent clinical trials have established that the converting enzyme inhibitors prolong survival in less sick patients, largely by slowing progressive left ventricular deterioration [36-38], and that this class of drugs even delays the appearance of clinical heart failure in patients with left ventricular dysfunction who have not yet developed symptoms of their cardiac condition [39]. This benefit was partly caused by an unexpected effect of the converting enzyme inhibitors to prevent adverse ischemic events in patients with ischemic heart disease [38, 40]. However, similar effects on survival in animal models of heart failure that do not involve progressive coronary occlusive disease [41, 42] indicate that the improved prognosis in clicaused by actions of the converting enzyme inhibitors on the failing myocardium itself.

The recent surprise has been that not all vasodilators share the same ability to improve prognosis in patients with chronic heart failure (Table 2). For example, long-term administration of the combination of hydralazine and isosorbide dinitrate is less effective than enalapril, a converting enzyme inhibitor [43]. Currently available data cannot tell us whether administration of both hydralazine and isosorbide dinitrate is necessary to achieve long-term benefit from this combination. However, several direct-acting vasodilators have adverse effects on prognosis, whereas nitrates may have beneficial growth-inhibitory effects (see below). Because the benefit of the combination of hydralazine and isosorbide dinitrate may be caused largely by isosorbide dinitrate, until new data become available to show that hydralazine has beneficial long-term effects, many physicians (including the author) no longer recommend hydralazine for the treatment of heart failure.

Not only are some vasodilators more effective in improving prognosis than others, but it is now apparent that many vasodilators do not have a beneficial effect on prognosis and that several are harmful. Prazosin, an -adrenergic blocker, did not improve survival after 42 months [34], and after 66 months, it slightly worsened prognosis [44]. Other classes of vasodilators have clearer adverse effects on prognosis. These include the phosphodiesterase inhibitors (see above), whose cardiovascular effects include a positive inotropic response as well as vasodilatation. The calcium channel blockers, which act directly to relax arteriolar resistance vessels, also appear to hasten deterioration in patients with heart failure [45-50]; this finding is often attributed to their negative inotropic effects. As already noted, however, long-term administration of ß-blockers, which also decrease contractility, can improve symptoms and left ventricular function in these patients. Instead, as discussed below, activation of neurohumoral responses—especially by the short-acting "first generation" calcium channel blockers—may contribute to these adverse effects. Another class of vasodilators, the prostaglandins, was recently found to have effects so adverse that a large clinical trial of prostacyclin in severely ill patients with heart failure was terminated prematurely (Letter to physicians from Burroughs Wellcome Co., 18 June 1993). High doses of flosequinan, a direct-acting vasodilator with positive inotropic actions [51] that relaxes smooth muscle by inhibiting phosphodiesterase activity [52] and formation of intracellular second messengers (inositol 1,4,5-trisphosphate and diacylglycerol [53]), were also found to have adverse effects on long-term survival (Letter to physicians from Boots Pharmaceuticals, 24 May 1993). These findings make it clear that many vasodilators fail to improve survival in patients with chronic heart failure, although all unload the heart and improve symptoms when administered for short periods.

The mechanisms responsible for the differences among various classes of vasodilators are not clear. Abrupt lowering of blood pressure by some vasodilators, especially those that are short acting, may activate harmful neurohumoral reflexes [54]. If this explanation is correct, vasodilators with prolonged pharmacokinetic responses (which minimize their hypotensive effects) may be able to unload the failing heart without provoking maladaptive baroreceptor responses [55]. As noted below, however, several vasodilators have effects on cell growth that may also contribute to their long-term beneficial response in patients with heart failure.

The Converting Enzyme Inhibitors

These agents inhibit the catalytic actions of a family of proteases commonly referred to as angiotensin converting enzymes. Although the most important target of these drugs was initially believed to be the serum angiotensin converting enzyme, it is now clear that the converting enzyme inhibitors also inhibit tissue converting enzymes [56, 57], which are sometimes more sensitive to these drugs than are the serum enzymes. Converting enzymes are dipeptidyl carboxypeptidases [58] that not only catalyze the formation of the vasoconstrictor angiotensin II from its inactive precursor angiotensin I but also the breakdown of the vasodilator bradykinin [59]. Thus, converting enzyme inhibitors can modify the concentrations of at least two extracellular messengers, thus inhibiting the production of a vasoconstrictor and the destruction of a vasodilator (Figure 1).

View larger version (22K): [in this window] [in a new window]   Figure 1. Converting enzyme inhibitors decrease the formation of angiotensin II and the breakdown of bradykinin. Decreased levels of angiotensin II (AII), by blunting the cascade shown at the left of the figure, inhibit the production of two intracellular messengers: inositol 1,4,5-trisphosphate (InsP3) and diacylglycerol (DAG). Because the former releases calcium (Ca2+) from internal stores, decreased InsP3 production in smooth muscle cells has a vasodilator effect. Inhibition of the cascade initiated by angiotensin II also decreases the production of the mitogen DAG, so that converting enzyme inhibitors can inhibit cell growth. Inhibition of bradykinin breakdown has even more complex effects, shown at the right of the figure. Increased bradykinin levels lead to the production of InsP3 and DAG. However, because the net effect of increased bradykinin is vasodilation, the calcium released by the bradykinin cascade appears to favor the production of several vasodilator molecules; these include nitric oxide (produced when an increase in cytosolic calcium activates nitric oxide synthase) and the eicosanoids, prostaglandin E2 and prostacyclin (released when calcium activates the lipolytic enzyme phospholipase A2 [PLA]). Bradykinin also activates adenylyl cyclase, which increases the production of yet another vasodilator molecule, cyclic adenosine monophosphate (cAMP). The nitric oxide formed by the bradykinin cascade appears also to have important growth-inhibitory effects. Together, these effects account for the ability of the converting enzyme inhibitors to unload the heart by decreasing peripheral resistance (afterload), and to exert growth-inhibitory effects. The latter, by slowing the progression of maladaptive hypertrophy, may play a role in the ability of the converting enzyme inhibitors to prolong survival in patients with heart failure. G protein = guanine nucleotide-binding coupling proteins; PKC = protein kinase C; and PLC = phospholipase C.

An additional consequence of the catalytic actions of converting enzymes is that angiotensin II and bradykinin have important effects other than to regulate arteriolar tone: Angiotensin II is not simply a vasoconstrictor but is also a mitogen, and bradykinin is not only a vasodilator but is also an inhibitor of cell growth and proliferation (Figure 1). This means that, in addition to unloading the failing heart, the converting enzyme inhibitors inhibit the formation of a growth promoter (angiotensin II) and the breakdown of a growth inhibitor (bradykinin). The resulting effects on the growth response of the overloaded myocardium may, as suggested below, contribute to the beneficial effects of the converting enzyme inhibitors in patients with heart failure.

Cellular Actions of Angiotensin II

The actions of angiotensin II begin when this extracellular messenger binds to specific angiotensin II receptors on the external surface of the plasma membrane (Figure 1). The bound receptor then activates a G-protein that increases the catalytic activity of phospholipase C, a lipolytic enzyme that hydrolyzes phosphatidylinositol 4,5-bisphosphate, a membrane lipid, to form two intracellular second messengers—inositol 1,4,5-trisphosphate and diacylglycerol. The former mobilizes Ca²+ from intracellular stores causing smooth muscle contraction and a weak inotropic response by the heart. Diacylglycerol also plays an important role in mediating the cellular response to angiotensin II by activating a large and diverse family of phospholipid-dependent enzymes, called protein kinase C, that regulate several cellular responses by catalyzing protein phosphorylation. The complex signaling cascade initiated by angiotensin II therefore has important effects on cell growth [60-64], as well as on muscle contraction. Some of these growth effects are mediated by an immediate-early gene response [65], which is an important element of the complex signaling cascade that initiates cardiac hypertrophy. As suggested below, reduction of angiotensin II levels by the converting enzyme inhibitors may prolong survival in patients with heart failure by blunting adverse effects of this growth response.

Cellular Actions of Bradykinin

The actions of bradykinin, like those initiated by angiotensin II, begin when this extracellular messenger binds to a membrane receptor, called the B₂ kinin receptor (Figure 1). This receptor, whose effects are also mediated by G-proteins, initiates cellular responses that are complex and differ among different tissues; these include activation of phospholipase C, which, as in the case of the angiotensin II receptor, generates inositol 1,4,5-trisphosphate and diacylglycerol [66, 67]. The ability of inositol 1,4,5-trisphosphate to release calcium from intracellular stores might, as in the case of angiotensin II, be expected to exert vasoconstrictor effects; however, other effects of the signaling cascade initiated when bradykinin binds to its receptor allow bradykinin to relax, rather than contract, resistance vessels. These vasodilator effects appear to be initiated when an increase in intracellular calcium concentrations activates nitric oxide synthetase, which catalyzes the production of nitric oxide, a potent vasodilator believed to be the physiologic endothelium-derived relaxing factor [59, 68]. The increase in cytosolic calcium initiated by bradykinin also activates phospholipase A₂, a lipolytic enzyme that catalyzes the production of a number of eicosanoids, including the vasodilators prostacyclin and prostaglandin E₂ [69]. The G-protein that is activated when bradykinin binds to its B₂ kinin receptor also activates adenylyl cyclase, an enzyme that produces cyclic AMP, which relaxes smooth muscle. Because the predominant effect of bradykinin is vasodilatation, inhibition of bradykinin breakdown, like that of angiotensin II formation, allows the converting enzyme inhibitors to unload the failing heart.

Evidence that bradykinin is a growth inhibitor as well as a vasodilator [66-68, 70] provides further support for the view that inhibition of a maladaptive growth response contributes to the beneficial actions of the converting enzyme inhibitors in patients with heart failure. It is now apparent, for example, that the signal transduction systems activated by bradykinin not only decrease vascular tone but also inhibit DNA synthesis [66, 67]. The resulting inhibition of cell growth can explain why the antiproliferative response of blood vessels to converting enzyme inhibitors is partially suppressed when the effects of the increased bradykinin levels are blocked [70]. Especially provocative is evidence that nitric oxide mediates the antiproliferative effect of bradykinin [68], which is consistent with the view that growth-inhibitory effects also contribute to the ability of a drug combination (hydralazine and isosorbide dinitrate) that includes nitrates [34], like the converting enzyme inhibitors, to prolong survival in patients with heart failure.

Evidence is increasing that the responses initiated when the converting enzyme inhibitors block bradykinin breakdown can differ from those responses caused by inhibition of angiotensin II formation. For example, the ability of the converting enzyme inhibitors to decrease infarction size after 30 minutes of ischemia in the rabbit heart appears to be caused mainly by increased bradykinin levels rather than by inhibition of angiotensin II effects [71, 72]. That bradykinin may also play an important role in mediating the beneficial effect of the converting enzyme inhibitors in patients with heart failure is suggested by evidence that angiotensin II receptor blockers, which inhibit binding of angiotensin II to its receptor but do not increase bradykinin levels, lack the protective effects of the converting enzyme inhibitors in some experimental models of heart failure [73]. Reports [74] that the ability of low doses of converting enzyme inhibitors to inhibit left ventricular hypertrophy after aortic banding can be abolished by administration of a B₂ kinin receptor blocker also suggest that inhibition of bradykinin breakdown may play an important role in the protective effects of converting enzyme inhibitors in patients with heart failure.

Importance of the Growth Response in Cardiac Hypertrophy

The possibility that the beneficial effects of the converting enzyme inhibitors—and perhaps also the nitrates—are caused in part by inhibition of maladaptive hypertrophy returns our discussion to the paradigm shifts described at the beginning of this article: that the poor prognosis in patients with heart failure is not caused simply by abnormalities in organ physiology (decreased cardiac output) and in cell biochemistry (depressed contractility and energy starvation) but also by abnormalities in the third paradigm of gene expression. To understand how accelerated gene expression induced by long-term overload might exacerbate heart failure, it is useful to examine the mechanisms that control myogenesis, the process by which the proliferating cells of the embryonic heart withdraw from the cell cycle to become differentiated adult cardiac myocytes.

Adult cardiac myocytes are terminally differentiated, which means they have ceased to divide but instead have settled into a quiescent phase of the cell cycle termed G_o. At the same time, protein synthesis, which had been rapid in the proliferating embryonic heart, slows to rates appropriate for cell maintenance and repair, and the pattern of protein synthesis changes with the expression of adult, rather than fetal, isoforms of many important proteins. Because terminally differentiated adult cardiac myocytes are unable to divide, our myocardial cells are the same age as we are. This means that the cardiac myocytes of an elderly person have suffered the genetic "wear and tear" associated with the normal lifetime of exposure to such cytotoxic agents as ionizing radiation and free radicals. Loss of myocardial cells in the hearts of aged rats has been reported [75], and mitochondrial gene mutations have been reported [76] to increase with normal aging. Fortunately, for most of us the cardiac myocytes, which differentiate in neonatal life, survive for 90 to 100 years before substantial numbers wear out and die; however, myocytes do appear to wear out, as suggested by the increasing prevalence of heart failure in elderly persons, which increases from 0.8% in the 6th decade of life to about 8% in the 9th decade [77].

One way to view the cardiomyopathy of overload is that chronic hypertrophy decreases the life span of cardiac myocytes to 4 to 5 years, the average survival of the patient with heart failure. This highlights the need to understand the transition from the initial adaptive hypertrophic response, where increased myocyte size distributes the overload among an increased number of sarcomeres, to the maladaptive hypertrophic response that develops after long-standing overload (see Table 1). As suggested below, new understanding of the factors that regulate muscle growth and differentiation has provided clues that may help explain, at least in broad outline, how this transition is brought about.

Myogenesis

Maturation of cardiac myocytes is accompanied by changes in the overall rate of protein synthesis and the nature of the proteins that are synthesized. In the proliferating myocytes of the embryonic heart (Figure 2, left panel), which are stimulated by a complex array of growth factors, rapid cell division is accompanied by active protein synthesis where fetal muscle-specific genes are expressed preferentially. The pattern changes drastically when, shortly after we are born, cardiac myocytes withdraw from the cell cycle and lose their ability to divide (Figure 2, center panel). At the same time, protein synthesis slows to rates appropriate for cell maintenance and repair, and the overall pattern of gene expression shifts to the production of adult muscle-specific gene products. As noted by Nadal-Ginard [78], proliferation and differentiation appear to be mutually exclusive in cardiac myocytes.

View larger version (19K): [in this window] [in a new window]   Figure 2. Overall growth patterns in the proliferating myocytes of the embryonic heart, the terminally differentiated myocytes of the normal adult heart, and the maladaptively hypertrophied myocytes of the failing heart. Left. In the embryonic heart, growth factors stimulate synthesis of fetal-specific gene products that, because protein synthesis is matched to active cell cycling, lead to normal cell division. Center. Withdrawal of growth factors and binding of myogenic factors to the E box in the terminally differentiated cells of the normal adult heart slows protein synthesis, inhibits the cell cycle, and favors the synthesis of adult muscle-specific gene products. Right. Overloading of the adult heart initiates an immediate-early gene response that reactivates growth factor stimulation; this, in turn, accelerates protein synthesis and favors the expression of fetal muscle-specific gene products. However, because the cell cycle remains blocked, the overloaded heart undergoes an unnatural growth response. One consequence of this response is accelerated myocardial cell death, which may be partly caused by the ability of some of the activated growth factors not only to stimulate protein synthesis but also to cause programmed cell death (apoptosis).

The Myogenic Determination Genes

These genes encode a family of helix-loop-helix proteins (for example, MyoD and myogenin) that regulate proliferation and differentiation when they bind to a DNA consensus sequence, called the E box, found in the control regions of a number of muscle-specific genes. Binding of the myogenic factors to E box sites favors differentiation by promoting the synthesis of adult muscle protein isoforms and by inhibiting the cell cycle [79-81]. In the heart, where regulation of these processes appears to be more complex than in skeletal muscle, E box-independent pathways, as well as E box sites, participate in myogenesis [82].

The Tumor Suppressor Genes

These genes, which also regulate myogenesis, encode the retinoblastoma gene product and p107, a related protein found in the heart [83, 84]. These regulators, sometimes called "pocket proteins" because their structure includes a pocket that binds the helix-loop-helix myogenic factors, block cell division and inhibit malignant transformation. The retinoblastoma gene product and p107 exist in phosphorylated and dephosphorylated states, which allows myogenesis to be controlled by protein kinases, enzymes that catalyze protein phosphorylation [85]. In terminally differentiated myocytes, the pocket proteins are in the dephosphorylated form that, along with the helix-loop-helix myogenic determinants, act on the E box to block the cell cycle. Phosphorylation reactions, which may include those initiated by angiotensin II and inhibited by bradykinin (see above), reverse this inhibition so as to stimulate protein synthesis.

Growth factors normally present in the fetus stimulate protein synthesis and proliferation of cardiac myocytes. In the terminally differentiated myocytes of the adult heart, however, the cell cycle remains blocked; as a result, the response to a renewed growth stimulus cannot end with normal cell division. Instead, activation of growth factors appears to induce an abnormal response in the terminally differentiated cardiac myocytes, which begin to express fetal protein isoforms, but the myocytes divide rarely, if at all (see below).

Maladaptive Hypertrophy and the Cardiomyopathy of Overload

A possible mechanism for the maladaptive hypertrophy that leads to the cardiomyopathy of overload is outlined in Figure 2. Hemodynamic overloading, through the actions of such factors as stretch and an energy deficit [86, 87], activates an immediate-early gene response [88-92]. In nonmyocytes, notably cardiac fibroblasts that are able to divide in the adult heart, this growth response stimulates proliferation. Protein synthesis is also accelerated in terminally differentiated myocytes, which allows the immediate-early gene response triggered by overload to increase the number of sarcomeres in the adult heart. As noted earlier, this provides a short-term adaptive response to the overload [9]. However, the new sarcomeres formed in response to the immediate-early gene response are not normal, as shown, for example, by the reappearance of fetal muscle-specific gene products [93]. Further, these terminally differentiated cells have little [94] or no [95] ability to re-enter the cell cycle. For this reason, hemodynamic overloading of the adult heart initiates an unnatural growth response where, as in the "natural" growth of the proliferating embryonic myocyte, protein synthesis is accelerated and fetal genes are preferentially expressed. In the overloaded adult heart, however, the growth response that in the fetal heart would normally end with cell division cannot be completed because the cell cycle remains blocked. Thus, as is true for many other aspects of the heart's response to overload (see Table 1), hypertrophy appears to be of short-term benefit, but when the overload is sustained, this growth response appears to cause long-term harm.

Cell Death in the Failing Heart

The mechanism by which the unnatural growth response triggered by overload leads to accelerated myocardial cell death is not clear. Loss of myocytes in the failing heart might represent a form of apoptosis or programmed cell death that, unlike necrosis (accidental cell death), is associated with proto-oncogene activation, protein synthesis, and energy expenditure [96, 97]. Although direct evidence to support this hypothesis is lacking, several of the growth factors that accelerate protein synthesis in the overloaded heart have recently been found also to stimulate programmed cell death. Notable is evidence that the proto-oncogenes c-myc and c-fos and the peptide transforming growth factor-ß, all of which participate in the immediate-early gene response [86-92, 98-100], also promote apoptosis [101-103]. Most striking is evidence that the same regions of the c-myc protein that activate the immediate-early gene response in the overloaded heart can induce programmed cell death [101]. Like Aesop's satyr, who with the same breath blew cold to cool his soup and hot to warm his frozen hands, the immediate-early gene response may stimulate adaptive and maladaptive hypertrophy.

Mitochondrial gene damage, which as noted earlier increases during normal aging [76, 104], has also been reported [105-107] to be more frequent in hearts from patients with idiopathic dilated cardiomyopathy and heart failure. Such abnormalities, by decreasing the capacity for oxidative phosphorylation, could contribute to the state of energy-starvation in the failing heart (see above). However, these findings have recently been challenged [108], so that the importance of mitochondrial gene abnormalities in the pathogenesis of the cardiomyopathy of overload needs further clarification.

Conclusions

When considered in terms of the evolution of our species, it is logical for the short-term response to overload to benefit the individual patient, but when it is sustained, to harm society [109]. An obvious advantage exists for nature to provide powerful means for the injured brave hunter, the wounded defender of hearth and home, and the fecund woman to survive the short-term decrease in cardiac output caused by hemorrhage. Yet, the ability of the same mechanisms to hasten death—when called on to compensate for a long-term decrease in cardiac output as occurs in patients with heart failure—would offer some advantage to a primitive social group by eliminating disabled persons unable to make a quick recovery.

Returning to our earlier analogy of the failing heart as an overloaded automobile with a brittle engine, the growth response of an overloaded adult heart that is failing may be similar to running this vehicle with one foot on a "growth accelerator" (activation of growth factors and the immediate-early gene response) and the other foot on a "growth brake" (blockage of cell division by the myogenic determinants and pocket proteins). Of course, in the car this would cause the engine and drive train to wear out quickly, much as long-term overload accelerates deterioration of the heart.

The view of maladaptive hypertrophy as an unnatural growth response by overloaded cardiac myocytes, in which protein synthesis is stimulated but cell division is blocked, may also explain why the converting enzyme inhibitors, and possibly nitrates, improve prognosis in patients with heart failure. The possibility that these drugs inhibit abnormal growth stimuli—and possibly apoptosis—in the overloaded myocardium could, like easing up on the "growth accelerator" of our overloaded automobile, explain how they slow clinical deterioration and prolong survival in these patients.