Having a Change of Heart: Reversing the Suicidal Proclivities of Cardiac Myocytes

Abstract

The molecular and cellular basis for cardiac remodeling have been difficult to establish. Transcriptional analysis and genetic manipulation of the mouse heart have revealed expression of a molecular program for cardiac myocyte suicide under conditions of myocardial injury or hemodynamic stress. Interrupting the cardiomyocyte suicide program by selective ablation of inducible apoptosis genes has proven to be remarkably effective in preventing remodeling and heart failure following myocardial infarction. Since these apoptotic genes are similarly dysregulated in human heart disease, the stage is set for a new era of therapeutics targeting cardiac suicide genes and their products.

Most of What You Need to Know as a Cardiologist in Two Minutes

It is axiomatic amongst lay persons that having an “enlarged heart” is bad news. It is somewhat paradoxical therefore that pathological cardiac enlargement, as distinguished from the trained athlete's heart, is almost always associated with a decrease in the number of cardiac myocytes. Indeed, heart failure (defined as any condition wherein the rate of blood delivery to systemic tissues is insufficient to meet their metabolic needs) tends to be irreversible because it results from the physical loss of functioning cardiac myocytes, so-called “cardiomyocyte drop-out” (¹). In coronary atheroslerosis and ischemic heart disease, cardiomyocyte drop-out tends to be focal, whereas it is more diffuse in non-ischemic failure (^{2, 3}). Importantly, the loss of functional contractile units of the heart and their replacement by fibrous scar or diffuse interstitial fibrosis produces secondary structural changes in the ventricles that increase the work required for ventricular contraction. Surviving myocytes slip past one-another and elongate as the extracellular matrix remodels (^{4, 5}). Cardiomyocyte drop-out and slippage produce “ventricular remodeling”, resulting in an enlarged, thin-walled ventricle.

Ventricular wall thinning and dilation oppose ventricular contraction and blood ejection into the aorta due to adverse physical properties of this ventricular geometry. A mathematical description of this physical relationship was developed by the 18^th century French physicist, Pierre Laplace, and indicates that wall stress (in the heart, afterload) is directly proportional to the radius of the chamber (r) and its intracavitary pressure (p), and is inversely proportional to the chamber wall thickness (h); stress = pr/2h (⁶). Thus, any primary injury to the heart that causes cardiomyocyte dropout and ventricular remodeling produces a secondary stress on the heart in the form of increased afterload that is chronic and unremitting. Studies performed over the past decade have shown that this hemodynamic stress is a powerful stimulus for programmed cardiomyocyte death, initiating a vicious cycle of unfavorable geometrical remodeling that stimulates yet more cell death (^{7, 8}). Since the myocardium has little ability to regenerate, loss of cardiac myocytes can have profound effects and initiate a downward functional spiral progressing to dilated cardiomyopathy and end-stage heart failure.

The Heart-Hormone Relationship

Pathological cardiac hypertrophy (i.e. reactive hypertrophy in response to myocardial injury or stress) is generally regarded to be a neurohormonal response. The seminal observation that demonstrated this relationship was norepinephrine stimulation of hypertrophy in cultured neonatal rat cardiac myocytes (⁹). In this and subsequent studies it became apparent that norepinephrine-, angiotensin II-, and endothelin-stimulated cultured neonatal rat cardiac myocytes develop the cellular and genetic features of pressure overload, including characteristic calcium cycling abnormalities (¹⁰). However, the important issue of whether contractile dysfunction was inherent in hypertrophied cardiac myocytes could not be addressed in the neonatal rat cardiomyocyte system because their contractile elements are not aligned in parallel, and fractional shortening can not therefore be measured. To address this question requires an in vivo system of pathological cardiac hypertrophy that does not alter cardiac loading. This precluded the approach of infusing norepinephrine or other pro-hypertrophic hormones into animal models, as they each are potent hypertensive agents. Therefore, we bypassed the hormones and their individual receptors, and transgenically overexpressed their common signal transducer, the alpha subunit of the heterotrimeric Gq G-protein, in cardiac myocytes. The resulting mice (Gq mice), exhibit cardiac hypertrophy with many of the structural and functional characteristics of pressure overload hypertrophy, but without pressure overload (¹¹). One of the most intriguing features of this model is that, with increased Gq expression (¹¹) or in the context of a superimposed physiological stress such as pregnancy (¹²), the hearts progressed from compensated hypertrophy to dysfunctional or lethal dilated cardiomyopathy, thus recapitulating the natural history of conventional pressure overload. Histologically, a hallmark of the Gq dilated cardiomyopathy was cardiac myocyte drop-out and replacement fibrosis.

To define the cellular mechanism for Gq-mediated cardiomyoycte dropout, we returned to the cultured neonatal rat cardiac myocyte system, in which cardiac load is not a variable. Infecting the cells with adenovirus expressing wild-type Gq resulted in stable hypertrophy, whereas infecting them with adenovirus expressing a constitutively active Gq mutant caused initial hypertrophy that rapidly progressed to apoptotic cardiomyocyte death. These findings were echoed in peripartum cardiomyopathy hearts from Gq transgenic mice, in which rapid cardiac dilation, functional decompensation, and death was associated with very high rates of cardiomyocyte apoptosis (¹²). Together, the Gq transgenic mouse studies established a biochemical basis for neurohormonal stimulation of cardiac hypertrophy, and demonstrated that Gq-dependent hypertrophies are both intrinsically dysfunctional from the perspective of contraction, and are abnormally sensitive to cardiomyocyte death from apoptosis.

Suicide Genes and the Transition from Hypertrophy to Failure

Myocardium is intrinsically resistant to apoptosis due in part to high levels of endogenous caspase inhibitors, such as X-linked inhibitor of apoptosis protein (XIAP) and apoptosis repressor with caspase recruitment domain (ARC) (^13–15). Accordingly, apoptotic cardiomyocytes are found in normal human hearts at a rate of ∼1/10,000 cardioyocytes, but the frequency increases orders of magnitude in ischemic and dilated cardiomyopathy and reactive cardiac hypertrophy (^{3, 16, 17}). Because adult cardiac myocytes are incapable of meaningful cellular regeneration (¹⁸), chronic persistent apoptosis at these relatively low rates (∼0.1-1% of total cardiomyocytes) is regarded as sufficient, over time, to produce cardiac dilation and heart failure (¹⁹).

To investigate possible molecular mediators of Gq-stimulated cardiomyocyte apoptosis, we performed an unbiased analysis of regulated mRNA transcripts in non-failing Gq transgenic hearts using an early Incyte DNA microarray. Of 153 genes that were significantly upregulated in Gq myocardium, four were annotated as being from apoptosis pathways. Of these, one unidentified expressed sequence tag showed striking upregulation in northern blots of pressure overloaded mouse hearts and human hypertensive hypertrophy. We cloned the full-length cDNA and identified the hypertrophy-regulated gene as Nix/Bnip3L, a pro-apoptotic member of the Bcl2 family of mitochondrial-targeted death factors (^{20, 21}). Recombinant Nix expression in HEK293 cells showed that it localized to mitochondria, caused release of cytochrome C into the cytoplasm, activated caspases, and produced apoptosis (i.e. that it activated the intrinsic apoptosis pathway).

Cardiac myocytes express a number of pro- and anti-apoptotic Bcl2 family proteins that are transcriptionally regulated in heart disease (^{22, 23}). Pro-apoptotic Bcl2 proteins are classified according to structural features that reflect their different functions (²⁴). The “multidomain” proteins, Bax and Bak, are the essential pore-forming proteins that lead to mitochondrial outer membrane permeabilization, cytochrome c release, and induction of the intrinsic apoptosis signaling cascade (see Figure 1). The activity of these pore-forming proteins is enhanced by pro-apoptotic “BH3 domain-only” proteins, like Nix, and opposed by anti-apoptotic factors like Bcl2 and Bcl-xl that bind to and inhibit the BH3-only proteins. In the heart and elsewhere, the principal role of BH3-only proteins appears to be to sense stress and initiate mitochondrial translocation of Bax or activation of mitochondrial-localized Bak. Cardiomyocyte fate is therefore determined in part by regulated expression of these factors (^{24, 25}). It seems surprising that cardiac myocytes, which are essentially incapable of replacing or regenerating their own losses, would dynamically express programmed cell death genes under conditions of stress. Yet, Bax expression increases and Bcl2 expression decreases during chronic experimental cardiac pressure overloading (²⁶). Nix is upregulated in human and experimental cardiac hypertrophy (^{20, 21}) and Bnip3 expression increases in ischemic cardiac models (^{21, 27}).

Fig. 1.

Multiple programmed cardiomyocyte death pathways influenced by Nix. Nix stimulates caspase-dependent and independent apoptosis by interacting with Bax or Bak to permeabilize mitochondrial outer membranes. Smaller membrane pores permit release of cytochrome C (cyt C) that activates caspases, which are tonically inhibited in cardiac myocytes by the inhibitor of apoptosis proteins (XIAP) and apoptosis repressor with caspase recruitment domain (ARC). As pores in the mitochondrial outer membrane enlarge, apoptosis inducing factor (AIF) and endonuclease-G (endo G) escape and translocate to the nucleus to stimulate DNA degradation in a caspase-independent manner. Sarcoplasmic reticular (SR) Nix can increase calcium transport to mitochondria, resulting in opening of the mitochondrial permeability transition pore (MPTP) that targets mitochondria for autophagic degradation and can lead to necrotic cell death.

Since cardioyocytes are intrinsically resistant to apoptosis, we determined if increased Nix expression alone was sufficient to cause functionally significant programmed cardiomyocyte death using transgenic overexpression driven by the α-myosin heavy chain (αMHC) promoter (²⁰). Two important characteristics of the αMHC expression system are that it is cardiomyocyte-specific, and that measurable overexpression begins shortly after the birth of the mouse. As a consequence, intrauterine cardiac development is not affected by αMHC-driven transgenes. αMHC-Nix mice were born at expected Mendelian ratios, but their neonatal growth was stunted, and they all succumbed after 1 week of life. Echocardiography performed on day 6 revealed left ventricular dilation and poor contractility, and histological studies of αMHC-Nix hearts on day 6–7 revealed striking increases in the rates of cardiomyoyte apoptosis (15–20%). These finding were similar in two independent transgenic lines, demonstrating that increased Nix expression is sufficient to cause a lethal apoptotic cardiomyopathy in neonatal mice. In a subsequent study using a conditional (tetracycline suppressible) αMHC promoter to drive cardiomyocyte Nix expression beginning after birth or in young adulthood, cardiomyocyte apoptosis was much more prevalent in neonates (∼15%) than adults (∼3%), and Nix overexpression at even high levels in adults produced little effect on cardiac size or contractile function. However, when adult Nix overexpression was superimposed on microsurgical creation of an aortic pressure gradient that in normal mice produces functionally-compensated pressure overload hypertrophy, the result was left ventricular remodeling and progression to heart failure (²⁸). These studies suggest that Nix is not only an initiator and effector of cardiomyocyte apoptosis, but is also a sensor of cardiac stress that helps to coordinate transcriptional and physiological cues for programmed cardiomyocyte death (²⁹).

We hypothesized that interrupting Nix might be able to prevent the adverse remodeling and heart failure that results from chronic cardiomyocye drop-out in ischemically-damaged or hypertrophied hearts. We have examined this notion by engineering the cardiac-specific Nix knockout mouse and subjecting it to pressure overloading by microsurgical creation of a transverse aortic coarctation (TAC), which normally upregulates Nix. We then compared apoptosis, ventricular remodeling, and cardiac performance over time between identically-treated wild-type and knockout mice (³⁰). We observed that TAC in Nix knockout mice produced only half the rate of early cardiomyocyte apoptosis and late replacement fibrosis as was seen in control hearts, and that ventricular dilatation and wall thinning was largely prevented over the subsequent weeks, resulting in preserved systolic function. Importantly, both at the cellular level and in terms of whole organ weights, Nix ablation did not change reactive left ventricular hypertrophy after pressure overloading. These data provided the first evidence that apoptosis is truly causal in decompensation of pressure overload, and establish Nix as a critical inducible factor that mediates this response.

In a parallel series of studies that are not described in detail, we have examined the role of related Bnip3 in infarction-associated myocardial apoptosis and ventricular remodeling. Whereas Nix is specifically upregulated in hypetrophy, Bnip3 is upregulated by ischemia. Accordingly, we examined the consequences of Bnip3 ablation on post-infarction remodeling and heart failure (³¹). Bnip3 gene ablation did not affect infarct size. However, when we examined post-ischemic apoptotic cardiomyocyte drop-out in the days after the acute ischemic insult, cardiomyocyte apoptosis was reduced by ∼50% in both the peri-infarct region and the myocardium remote from the ischemic zone of Bnip3 null mice. This decreased apoptosis in infarcted Bnip3 knockout hearts was associated with striking declines in left ventricular remodeling and improved cardiac performance measured by magnetic resonance imaging. Thus, prevention of Bnip3-mediated cardiomyocyte death during the days and weeks after a transient ischemic insult decreased apoptosis and prevented ventricular dilation, which preserved left ventricular ejection performance and prevented heart failure. Together, the Nix and Bnip3 knockout remodeling rescue studies also prove that necrotic cardiomyocyte death is not the inevitable consequence of preventing programmed death in ischemic or hypertrophied cardiomyocytes [as has occasionally been suggested (⁷)], which supports the development of therapies aimed at preventing ventricular remodeling by targeting transcriptional pro-apoptotic events specific to different forms of cardiac injury or stress.

DISCUSSION

Runge, Chapel Hill: Gerald that was a beautiful talk! When you think about NIX or BNIP as potential therapeutic targets, what other roles do they have either in development or outside the heart?

Dorn, St. Louis: That is an excellent question. When we did the germline knockout of NIX, we actually got polycythemia vera. NIX is necessary for mitochondrial clearance or mitochondrial pruning, and this is especially important in erythrocytes. If you have erythrocytes that are full of mitochondria, they are subjected to oxidative stress, and you get a hemolytic anemia. Between the time of reticulocyte formation and erythrocyte release, NIX is transcriptionally regulated in the bone marrow and the spleen (in the case of the mouse), which is important for mitochondrial clearance to prevent hemolysis. In the heart, NIX and BNIP3 actually don’t seem to have a significant effect on development.

Abboud, Iowa City: So the suggestion is hypertrophy as you presented is maladaptive. Any thoughts on an adaptive hypertrophy that one might get with exercise or NIX not generated in that setting?

Dorn, St. Louis: Sure. There is very little in common between reactive hypertrophy as I have described and normal physiologic hypertrophy. That is not the same genetic program. Yes, cardiac myocytes get larger, but they are very, very different conditions. As you are aware, some people have suggested that exercise or training-induced hypertrophy is mediated through the IGF growth factor response and not through these types of neurohormones. The different hypertrophies may also be due to the nature of the stimulus and the fact that the stimulus is not chronic and persistent as it would be with aortic valvular disease or with hypertension or with myocardial injury. So when we refer to “hypertrophy” or “pathological hypertrophy”, we are not talking about exercise-induced hypertrophy that is not only generally beneficial, but can be an excellent treatment for heart failure.

Chapman, Jackson: This may be a naïve question coming from an infectious disease person, but what happened to the role of spironolactone and antifibrosis and how this interacts with some of the genes that you are talking about?

Dorn, St. Louis: Aldactone is being used for the treatment of heart failure, or aldosterone antagonists are, and I think that there is a compelling reason to not only block angiotensin pathways and beta-adrenergic pathways, but also aldosterone pathways. These are all effective in experimental models certainly. Sometimes neurohormonal antagonists in human studies regress hypertrophy and prevent the fibrosis—both interstitial fibrosis that seems to be more of a generalized matrix problem and also cardiomyocyte dropout in replacement fibrosis. So that is what we are doing right now. We are using pharmacological neurohormonal antagonism to effectively treat heart failure and potentially reverse hypertrophy.

Chapman, Jackson: Is it working through any of these genes?

Dorn, St. Louis: I believe that these genes are further downstream, and when you interrupt the cycle pharmacologically you also diminish the genetic program, particularly in experimental models.

Alexander, Atlanta: You mentioned regeneration or lack of it, creation of new myocytes. What about stem cells?

Dorn, St. Louis: There is a tremendous amount of interest. I think the things the things that we are trying to do, which is to prevent myocytes from dying needless deaths, are complimentary to those folks who are trying to use various stem cell approaches to regenerate myocardium. There is, I think, little doubt that cardiac myocytes can undergo very, very infrequent cell divisions. The problem is that it is not sufficient to meaningfully affect pretty much anything. So do we use endogenous stem cells, or bone marrow mesenchymal stem cells, to try to repopulate the heart? One of the interesting facets that we are working on is when you are doing these stem cell transplantations in the heart, you get maybe one in 10,000 of your transplanted stem cells that will die. Many of the rest of them undergo apoptotic cell deaths. So there may be complimentary means of targeting apoptosis not just in the heart but in those cells that might be successfully engineered in order to regenerate myocardium. We are looking at that.