In five viable lines with levels from 60- to 350-fold normal in a model of ß-2 AR overexpression, at lower levels of ßAR overexpression the hearts were virtually normal in pathology, normal size, wall thickness and function; similar to initial reports from other investigators of ß-2AR overexpression enhancing cardiac function. Toxic levels, 350-fold baseline, resulted in a phenotype with left and right ventricular dilation, wall thinning, massive atrial enlargement with formation of thrombus in the atria-indicative of a murine DCM.

In vivo hemodynamic analysis measuring left ventricular (LV) dv/dt as a function of isoproterenol infused showed a parallel increase in isoproterenol dose and LV contractile function (Fig. 2). In the lower level overexpressers, basal LV function was maximal. In the higher level overexpressers, cardiac function was measured at two time points. At 12 weeks, basal function was enhanced over non-transgenic, with no response to isoproterenol, and at high levels of isoproterenol function was diminished. At 20 weeks, the functional decline was even more enhanced. Basal function was essentially normal, with no response to isoproterenol.

A time-dependent decline in LV function in the ß-2AR mice at high levels of overexpression was suggested by these data, prompting a longitudinal analysis. The survival curve showed that at about 11 weeks, the 350-fold overexpressers began to die, and all were dead by 35 weeks (Fig. 3,top panel). This suggested the "classic" progression towards DCM, if indeed a hypertrophy phenotype exists. Molecular analysis using RNA dot blot expression of hypertrophy-associated genes in a time course of 7, 11 and 17 week-old ß-2AR-350 animals revealed no molecular progression (Fig. 3, bottom panel). The same molecular abnormalities of increased beta myosin heavy chain and atrial natriuretic factor (ANF) expression were present in 7-week and 17-week old animals, the age at which a selection bias began to appear in the cardiac phenotype due to death.

A rather pronounced cellular progression was present. Full thickness LV histologic samples showed: at 7 weeks a normal histology, at 11 weeks areas of cardiomyocte dropout, at 17 weeks more pronounced areas of cardiomyocte dropout with indication of some fibrotic replacement, and at 25 weeks nearly complete replacement of the ventricular myocardium by fibrosis. This is progressive cardiomyocyte necrosis and replacement fibrosis resulting in a DCM.

A model of peripartum DCM in a transgenic mouse overexpressing the Gaq signaling protein was studied. Gaq is the first common signaling molecule for a variety of hypertrophy-stimulating agonists, including angiotensin I, endothelin and ß-adrenergic agonists. Gaq overexpression at 4- to 5-fold endogenous levels resulted in a hypertrophy phenotype with increased LV wall thickness and size, a 20% increase in cardiac mass, extremely increased levels of beta myosin heavy chain and ANF expression, increased ß-skeletal actin expression, and decreased sarcoplasmic reticulum calcium ATP expression.

The model very closely reproduced pressure overload hypertrophy. The animals had about a 20% decrease in contractile function and enlarged cardiac myocyte cross-sectional areas.

Animals with about 8-fold overexpression of the Gaq signaling protein resulted from breeding animals with about 4- to 5-fold overexpression. A different phenotype resulted: chronic progressive cardiomyopathy, quite large heart, left and right ventricle dilation, quite large atria, and cardiomyocyte dropout with a small amount of fibrosis. In the peripartum period, a more severe form of the Gaq CM is seen. An impregnated transgenic mouse developed a severe rapidly progressing CM, generally within one week of giving birth. The mortality rate was about 50% per pregnancy event. The hearts are generally about 300-fold larger than a normal Gaq heart, organized and acute thrombus in the left and right atria can be seen, indicating a chronic low cardiac output state.

The involvement of an apoptotic process was suggested by the nuclear morphology and in vitro studies. On histology, a cellular transparency and an associated nuclear abnormality in which the chromatin began to clump in the periphery or the center, much unlike the H&E stain of the normal heart was seen. TUNEL assay, DNA laddering, and ultrastructural examination revealed a relatively massive apoptotic process, with apoptotic indices approaching 20%. The mitochondrial abnormality in the Gq peripartum cardiomyopathy models was striking ultrastructurally, with a great deal of disarray and the internal tubular structures falling apart, seemingly out of proportion to the sarcomeric degeneration.

Transverse aortic coarctation

This represents another method to progress the Gq heart to DCM. The animals were aortic-banded, and followed for 3 weeks. In the non-transgenic animals pre- and post-aortic coarctation there was the expected development of concentric hypertrophy, with an increase in the relative ratio of wall thickness to ventricular radius. However, in the Gq animals, an eccentric form of hypertrophy progressed to a DCM. A small amount of hypertrophy was seen, but was counterbalanced by an increase in ventricular dimension.

The banded Gq model is associated with cardiomyocyte apoptosis, as shown by recent studies performed in collaboration with Ross. This apoptotic process differs somewhat from the peripartum model. The TUNEL staining occurs in clusters rather than diffusely throughout the myocardium, and the area of staining corresponds to areas of cardiomyocyte dropout and the beginning of a fibrotic process.

Interestingly, the levels of cardiomyocyte apoptosis seen in the Gq mice inversely correspond with the depression of cardiac function after banding, as assessed by echocardiographic LV shortening. There is a shortening between apoptosis and diminished LV function in these animals.

In the Gq mice, there is a progression of phenotype from normal in non-transgenic, to stable hypertrophy in the Gq overexpressers, to apoptotic CM in either high level Gq overexpression, peripartum CM, or after aortic banding (Fig 4). Koch has previously shown that dominant negative inhibition of Gq causes failure to hypertrophy in the presence of transverse aortic banding. Thus, it is believed that the complete spectrum of Gq signaling and its association with hypertrophy has been identified.

Progression of hypertrophy to apoptotic CM

An hypothesis that a genetic program for Gq apoptosis exists, recapitulating the genetic program for Gq hypertrophy, was then formed. This is rather novel, as no genetic program for apoptosis is known. A small cluster of apoptosis genes that is upregulated was found by comparative analysis of the Gq and non-transgenic gene expression animals using Incyte DNA microarrays.

The Gq animal is not undergoing apoptosis, but is pre-disposed to apoptosis. This finding is supported by DNA microarray analysis and conventional assay of ANF. For the apoptosis gene caspase-1, the relative differential expression on the chip was 4.0 and Northern blot analysis showed substantial upregulation. Increased levels of pro-caspase-1 can be measured in the Gq animals, relative to the non-transgenic animals, but is not associated with caspase activation. Upregulation of the Fas-associated factor was found. A non-identified express sequence tag, with substantial upregulation and some homology to a Bcl-interacting protein, was studied. Multiple transcripts were noted on RNA dot blot analysis. Using that cDNA fragment, they cloned from a mouse cardiac library mouse Nix cDNA. Nix is a recently described pro-apoptotic mitochondrial protein, that forms pores, interacts with Bcl-2 and Fas, and is probably the mitochondrial effector for the pro-apoptotic proteins. Mitochondrial degeneration is a hallmark feature of the Gq apoptosis. Alternately spliced variants of Nix that are differentially regulated in the Gq mouse have been found in their lab, and they are now looking at this gene in human hypertrophy and heart failure. They believe they are now beginning to understand the mechanism whereby the Gq signaling pre-disposes to cardiomyocyte apoptosis.

Through genetic manipulation they recently have shown it is possible to modulate phenotypes by increasing the size or number of cardiomyocytes or by preventing their loss. Unpublished data on epsilon PKC translocation inhibition and activation, which can result in a hypoplastic phenotype, was reviewed.

The Gq pathway downstream of phospholipase C becomes quite complex due to a number of endogenous protein kinase Cs. These may have different effects on the cardiomyocyte, and determining which may be effectors of the Gq or other phenotypes and which may result in hypertrophy, heart failure or apoptosis or a combination is difficult. Rosen has demonstrated that the mechanism of activation of PKCs is directly related to their location in the cell and to the steric form of the molecule. Inactive PKC is folded over onto itself and is associated with receptors for inactive C kinases. When the PKCs need to be activated by calcium or diacylglycerol (DAG), they unfold and thereby expose their substrate binding site and a binding site for a receptor for activated C kinase (RACK), which can be conceived of as a factory. The unfolded PKC goes to work in the factory where the substrates are, allowing substrate-specific activity of different PKC isoforms, as each PKC isoform has a different RACK.

Rosen conceived of the notion of using competitive peptides based on the RACK binding sequences in active PKC to prevent their interaction with RACK. Alternately, other peptides can be used even in the absence of calcium and DAG to force the PKC to unfold and expose its RACK binding site, thereby translocating and activating it. These are catalytic-inactive peptides that can modulate the translocation and hence activation of PKC.

Studies with epsilon PKC

This approach has been used in Dorn's lab to express PKC-specific peptides for epsilon, delta, alpha, beta, and for the atypical PKCs. The biochemical data for ePKC activating peptide, called pseudo epsilon RACK, and an inhibiting peptide, shows there is no effect on the overall amount of ePKC. However, the activating peptide does increase the amount of PKC in the particulate fraction, about 20% at baseline. The inhibitory peptide decreases the amount of ePKC in the particulate fraction, about 20% at baseline. The lack of an effect on alpha PKC shows this to be an isoform-specific effect.

Epsilon PKC activation and inhibition can increase or decrease the number of cardiac myocytes in the heart. Activation of PKC increased the number of cardiac myocytes that grew during the early post-embryonic period in their studies. This activation led to a hypertrophic phenotype with small hearts with thick walls, functionally normal by echo, biochemical and invasive measures. The molecular phenotype showed an increase in beta myosin heavy chain alone; the ANF transcript was not altered. The heart weight to body weight was increased at 15 weeks but not at 8 weeks in the transgenic animals. This was not cellular hypertrophy, because the myocardial cells were smaller than those in the non-transgenic heart.

A model of inhibition of cardiomyocyte proliferation during post-natal development resulting in DCM was developed through ePKC inhibition with epsilon B-1 peptide. Lower or modest levels of expression of epsilon B-1 peptide resulted in normal looking hearts, whereas very high levels of expression resulted in DCM with lethality at about 30 weeks. Histologic examination showed cardiomyocytes to be low in number but very large with no replacement fibrosis, as seen in the ßAR model.

The addition of the pseudo-epsilon RACK transgene to the Gaq transgene resulted in a smaller heart with more cardiac myocytes, and normal liver and lungs. Enhanced cardiac function and increased fractional shortening in a smaller heart with normal wall thickness resulted.