ε Protein Kinase C in Pathological Myocardial Hypertrophy

The ε isoform of protein kinase C (PKC) has a critical cardiotrophic function in normal postnatal developing heart as demonstrated by cardiac-specific transgenic expression of εPKC-selective translocation inhibitor (εV1) and activator (ψεRACK) peptides (Mochly-Rosen, D., Wu, G., Hahn, H., Osinska, H., Liron, T., Lorenz, J. N., Robbins, J., and Dorn, G. W., II (2000) Circ. Res. 86, 1173–1179). To define the role of εPKC signaling in pathological myocardial hypertrophy, εV1 or ψεRACK were co-expressed in mouse hearts with Gαq, a PKC-linked hypertrophy signal transducer. Compared with Gαq overexpression alone, co-expression of ψεRACK with Gαq increased εPKC particulate partitioning by 30 ± 2%, whereas co-expression of εV1 with Gαqreduced particulate-associated εPKC by 22 ± 1%. Facilitation of εPKC translocation by ψεRACK in Gαq mice improved cardiac contractile function measured as left ventricular fractional shortening (30 ± 3% Gαq versus 43 ± 2% ψεRACK/Gαq,p < 0.05). Conversely, inhibition of εPKC by εV1 modified the Gαq nonfailing hypertrophy phenotype to that of a lethal dilated cardiomyopathy. These opposing effects of εPKC translocation activation and inhibition in Gαqhypertrophy indicate that εPKC signaling is a compensatory event in myocardial hypertrophy, rather than a pathological event, and support the possible therapeutic efficacy of selective εPKC translocation enhancement in cardiac insufficiency.

The protein kinase Cs (PKCs) 1 constitute a large family of ubiquitous phospholipid-dependent serine-threonine kinases postulated to have diverse effects in the heart, including mediating cardiac hypertrophy/failure and myocardial protection (1)(2)(3). It has been difficult to establish particular roles for individual PKC isoforms due to the absence of isoform-specific agonists and antagonists. Targeted cardiac overexpression of PKC has been helpful in establishing possible effects of myocardial PKC ␤ signaling (4,5), but the potential for nonspecific or nonphysiological signaling of overexpressed enzymes warrants a cautious interpretation of resulting phenotypes. Based on recent insights into the mechanism for differential PKC isoform subcellular translocation upon activation (6), a strategy was developed whereby the effects of endogenous myocardial PKC isoforms such as ⑀PKC could be elucidated by cardiacspecific expression of peptides that specifically modulate PKC isoform translocation/activation (7,8). These ⑀PKC-derived peptides, ⑀RACK (HDAPIGYD, rat ⑀PKC amino acids 85-92) and ⑀V1 (the first variable region of rat ⑀PKC, amino acids 2-144) have previously been shown to specifically modulate ⑀PKC translocation (7)(8)(9). Compared with transgenic overexpression of PKC, a critical advantage of translocation modification is that PKC isoforms are specifically regulated without changing the natural stoichiometric relationships of PKC to its upstream activators and downstream substrates. Cardiac transgenic models expressing either the ⑀PKC translocation inhibitor (⑀V1) or activator (⑀RACK) were previously created with opposing cardiac phenotypes of dilated cardiomyopathy and physiological hypertrophy, respectively (8), demonstrating that ⑀PKC signaling is both necessary and sufficient for normal myocardial growth in the developing young mouse. This apparently beneficial role for ⑀PKC in normal physiologic cardiac growth is, however, at odds with the widely held notion that pathological cardiac hypertrophy is the consequence of ⑀PKC signaling (1, 10 -13). In addressing this apparent contradiction, we hypothesized that the effects of ⑀PKC on a diseased heart might be different than during normal heart development. We further considered that these putative pathologic effects could be identified using ⑀PKC-specific translocation modification in the context of a form of cardiac hypertrophy where ⑀PKC is thought to be a pathological mediator, like the cardiac-specific G␣ q -overexpressing mouse, which develops myocardial hypertrophy and contractile depression associated with ⑀PKC translocation (11,12). ⑀V1 or ⑀RACK were therefore co-expressed with G␣ q by cross-breeding the respective individual transgenic mice to generate compound transgenic mice.

EXPERIMENTAL PROCEDURES
Transgenic Models-Details of transgenic mice (FVB/N background) overexpressing G␣ q , ⑀V1, or ⑀RACK under control of the full-length mouse ␣Ϫmyosin heavy chain (MHC) promoter have previously been described (7,8,11). G␣ q mice express G␣ q at ϳ5 times nontransgenic levels (the line previously described as G␣ q -40) and develop nonfailing ventricular hypertrophy with modest contractile depression, which is unresponsive to ␤-adrenergic agonists (11). ⑀RACK transgenic mice develop mild hypertrophy with normal ventricular function and normal response to ␤-adrenergic agonists (7,8). The ⑀V1 low (lowest expressing) mice used herein have normal cardiac mass and function, whereas higher expressing ⑀V1 lines develop ventricular dilation with wall thinning, diminished systolic function, and heart failure (8). Compound transgenic mice overexpressing ⑀V1 or ⑀RACK and G␣ q were obtained by breeding heterozygous ⑀V1 low or ⑀RACK with heterozygous G␣ q -40 mice. Transgenes, alone and in combination, were identified by genomic Southern analysis of tail clip DNA. All experiments were performed with comparison of littermates.
Western Blot Analysis-Samples for measurement of transgene products and PKC expression were prepared as described previously (12). Briefly, mouse ventricles frozen at Ϫ80°C were homogenized in buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 10 mM EGTA, 5 mM dithiothre-itol, 10 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 10 g/ml pepstatin, 10 g/ml aprotinin, and 10 g/ml leupeptin) followed by centrifugation at 100,000 ϫ g for 1 h. The pellet was further extracted with buffer containing 1% Triton X-100 for 30 min on ice. Proteins were separated on SDS-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membrane. ⑀V1 and G␣ q transgene expression were evaluated by Western blot analysis using anti-FLAG M2 (Sigma) and anti-G␣ q/11 (Santa Cruz) monoclonal antibodies, respectively. PKC isoform expression was measured by quantitative immunoblot analysis with anti-␣PKC (Santa Cruz) and anti-⑀PKC antibodies (Transduction Laboratories) using recombinant human ␣PKC and ⑀PKC (Calbiochem) as quantitative standards (8,12). Protein loading and transfer efficiency were evaluated by amido black staining of the membrane after immunoblotting.
Assessment of Cardiac Hypertrophy and Function-Morphometric analysis and histological examination of Masson's trichrome-stained ventricles used standard techniques. Echocardiography was performed to determine cardiac contractile function in a noninvasive manner; left ventricular fractional shortening (the proportion of blood ejected during systole compared with maximal ventricular capacity in diastole) and left ventricular mass were calculated (14). Cardiac gene expression was assayed by RNA dot blot analysis using total RNA (3 g/dot) extracted from ventricles of nontransgenic and transgenic mice and 32 P-labeled oligonucleotides as probes (11). Hybridization was quantified using a Storm PhosphorImager (Molecular Dynamics) and expression of each cardiac gene was normalized to glyceraldehyde-3-phosphate dehydrogenase.
Statistical Analysis-All data are expressed as mean Ϯ S.E. Comparisons between G␣ q and dual transgenic mice were evaluated using Student's t test, and the p value of Ͻ0.05 was considered as statistically significant. Values for nontransgenic (NTG) siblings are also presented for comparison, but were not included in the statistical analysis.

RESULTS AND DISCUSSION
Cardiac-targeted transgenic expression of ⑀V1 and resulting inhibition of myocardial ⑀PKC translocation has proven that ⑀PKC activation is necessary for normal physiological myocardial growth in the postnatal period (8). Complementing this finding is the observation that enhancing ⑀PKC activation by expression of ⑀RACK stimulates growth of normally functioning myocardium (8). Herein, we utilized in vivo translocation modulation to characterize ⑀PKC effects in G␣ q -stimulated cardiac hypertrophy, a form of hypertrophy in which ⑀PKC translocation is increased and in which ⑀PKC has therefore been considered to be a possible pathological mediator (11,12). Compound transgenic mice were generated and ventricular expression of ⑀V1 and G␣ q was compared between parent mice and dual transgenic progeny by Western blot analysis (Fig. 1A). Compound transgenic mice expressed both ⑀V1 and G␣ q at the same levels as parent lines, indicating that no cross-talk occurred at the level of transgene expression. As reported previously, it is not possible to detect ⑀RACK expression by Western blot analysis due to the small molecular weight of the peptide (8). However, Northern blot analysis showed no change in ⑀RACK and G␣ q transgene expression in compound transgenic mice, compared with parent lines (data not shown).
⑀RACK/G␣ q mice were healthy when followed for 5 months. Effects of enhanced ⑀PKC translocation/activation on G␣ q -mediated cardiac hypertrophy and contractile dysfunction were assessed in 12-week-old ⑀RACK/G␣ q mice. Compared with G␣ q transgenic siblings, ventricular and atrial weights in ⑀RACK/G␣ q mice were reduced (Table I). Left ventricular mass of ⑀RACK/G␣ q mice calculated from noninvasive echocardiographic analysis was correspondingly decreased (Table  II). Echocardiography further revealed that ⑀RACK/G␣ q hearts had reduced end diastolic and systolic dimensions, i.e. that the cardiac ventricles were smaller before or after contraction, with no change in left ventricular wall thickness (Table II). Thus, the ratio of wall thickness to ventricular radius (h/r) was increased, which suggests concentric ventricular remodeling. Perhaps as a consequence of this more favorable ventricular geometry, ⑀RACK expression enhanced left ventricular contractile function of G␣ q mice, measured echocardiographically as the proportion of blood volume ejected per cardiac cycle, or fractional shortening (NTG 51 Ϯ 3%, ⑀RACK 47 Ϯ 1%, G␣ q 30 Ϯ 3% and ⑀RACK/G␣ q 43 Ϯ 2%).
A molecular marker of cardiac hypertrophy is increased expression of the embryonic cardiac genes ␣-skeletal actin, ␤-MHC, and atrial netriuretic peptide (ANF) in ventricle (16). These genes are highly expressed in G␣ q overexpressing hearts (11,14), and Northern analysis revealed that ␣-skeletal actin gene expression was selectively reduced in ventricles of ⑀RACK/G␣ q mice, whereas ANF and ␤-MHC gene expression was not changed from their characteristically elevated levels (11) (Fig. 2B). Since inhibition of ⑀PKC with ⑀V1 in normal hearts was previously shown to selectively increase expression of ␣-skeletal actin gene (8), the isolated reduction of this mRNA by ⑀RACK in the G␣ q mice suggests a role for ⑀PKC in transcriptionally regulating actin, but not ANF or ␤-MHC expression, in the heart. Taken together, the results with ⑀RACK/ G␣ q mice indicate that enhanced ⑀PKC translocation improves ventricular function and normalizes ventricular geometry, but surprisingly, diminishes the extent of cardiac hypertrophy in G␣ q overexpressors.
Heart failure is not typically observed in G␣ q -overexpressing mice, although the characteristic eccentric hypertrophy is associated with measurable contractile depression (11). Heart failure also is not a feature of ⑀V1 low mice in which hearts are functionally and histologically normal (8). Thus, breeding G␣ q with ⑀V1 low is the equivalent of crossing two nonfailing cardiac models. Strikingly, combined expression of ⑀V1 low with G␣ q caused development of lethal heart failure at 14 Ϯ 2 weeks of age (n ϭ 9). At 12 weeks of age, ⑀V1/G␣ q mice could not tolerate echocardiographic analysis under even light anesthesia, so a confident assessment of in vivo ventricular function at an age matching that of the ⑀RACK/G␣ q mice was not obtained. However, gross pathological examination of 12-week-old hearts revealed ventricular dilatation, focal ventricular wall thinning, and massively enlarged atria containing intracavitary blood clots, which are typically observed in chronic low cardiac output states, i.e. heart failure ( Fig. 2A). These gross pathological findings were much more striking than the histological appearance which revealed only areas of mild myocardial fibrosis, as detected by blue staining with Masson's trichrome (Fig. 2C), Interestingly, ⑀V1 overexpression did not alter the program of ANF, ␤-MHC, or ␣-skeletal actin gene expression in G␣ q mice (Fig. 2B). Since a clear mechanism for development of heart failure was not readily apparent from these postmortem studies, and as 12-week-old mice were too sick to undergo in vivo   physiological assessment, we performed echocardiographic examination of apparently healthy 6-week-old ⑀V1/G␣ q mice to help define the relevant pathophysiology. At 6 weeks, ⑀V1/G␣ q hearts exhibited ventricular enlargement and wall thinning with depressed contractile function (fractional shortening) compared with G␣ q , which at this age have not yet developed contractile dysfunction (Table II). These findings suggested that ⑀V1 low inhibits myocardial growth in the young developing G␣ q mice, as ⑀V1 high did in mice with a normal genetic background (8). This notion was confirmed by a 27% reduction in dry ventricular weight of ⑀V1/G␣ q mice, compared with G␣ q (Table I). In contrast, atrial weight was increased, which, as noted above, probably reflects contractile depression. Finally, the ratio of ventricular wall thickness to ventricular radius (h/r) was significantly diminished, indicating that dilatory remodeling, i.e. ventricular wall thinning with chamber enlargement, was already occurring in 6-week-old ⑀V1/G␣ q mice. Thus, inhibition of ⑀PKC translocation, which is already "enhanced" in the G␣q mouse, impairs ventricular function, causes cardiac enlargement with ventricular dilatation, and reduces ventricular mass.
The reciprocal cardiac phenotypes of ⑀RACK/G␣ q and ⑀V1/ G␣ q mice described above unequivocally refute our hypothesis that ⑀PKC signaling contributes to the pathology associated with G␣ q -mediated hypertrophy. Although ⑀PKC activation occurs in G␣ q hearts, "superactivation" with ⑀RACK proved to be therapeutic, whereas even the modest impairment of ⑀PKC translocation caused by ⑀V1 low , essentially "normalization," greatly exaggerated the underlying G␣ q pathology. Most surprising is that the opposing effects on cardiac function were not due to opposite effects on cardiac "hypertrophy" per se, in that ⑀RACK and ⑀V1 both decreased left ventricular mass of G␣ q mice. Rather, it is apparently an ⑀PKC effect on ventricular geometry that is the critical determinant of viability, with benefit accruing from ⑀RACK-induced concentric remodeling (of the eccentrically hypertrophied G␣ q heart (11)) and ⑀V1induced ventricular dilatation proving fatal. These changes in geometry were surprisingly well defined in that they were not associated with significant changes in ANF or ␤-MHC gene expression, which are widely considered to be indices of cardiac pathology (16). In the context of our prior study (8), the current results show that the molecular and morphometric effects of ⑀PKC are similar in nontransgenic and G␣ q overexpressing hearts, but the effects of ⑀PKC activation and inhibition are amplified in the G␣ q model.
An intriguing difference between the current studies of ⑀PKC in the G␣ q background and our previous characterization of ⑀PKC translocation modulation in the normal background is the absence of a "hypertrophic" effect of ⑀RACK plus G␣ q . We previously found a modest increase in ventricular mass of 12-week-old ⑀RACK mice, with concentric remodeling. Compared with G␣ q , however, ⑀RACK/G␣ q ventricles were concentrically remodeled, but smaller. This suggests that the intrinsic contractile depression and eccentric remodeling of G␣ q mice may itself contribute to the hypertrophic response. ⑀RACK improved ventricular geometry and function and as a conse-quence diminished the external stimulus for hypertrophy.
A strength of the current studies is that ⑀PKC activity/ translocation was modulated by expressing catalytically inactive peptides rather than by overexpressing ⑀PKC itself, essentially using transgenesis as a means of organ-specific drug delivery. However, the effects of ⑀PKC modulation were measured in the G␣ q transgenic mouse, which achieves its phenotype by 5-fold increased expression of G␣ q (11). Thus, concerns regarding stoichiometric excess and nonspecific activity of overexpressed signaling proteins (which drove our unique approach of PKC isoform translocation modification) apply to the G␣ q model itself. While the rationale for choosing the G␣ q model is strong, i.e. changes in ⑀PKC translocation, which suggested it as a pathological mediator (11,12), the relevance of G␣ q overexpression to naturally occurring cardiac disease is uncertain. Additional studies are therefore ongoing to determine the effects of ⑀V1 and ⑀RACK on cardiac hypertrophy and function in hemodynamically stressed hearts.
In conclusion, the current studies contradict the generally accepted notion that ⑀PKC is a pathophysiological mediator of cardiac disease. Rather, they support the notion that ⑀PKC translocation, either by direct administration of ⑀RACK peptide, by gene therapy, or through novel pharmacological means, could be used to treat heart failure in the same manner that induction of "physiological" hypertrophy by exercise or growth factors has proven to be therapeutic (17).