Control of Cardiac Growth and Function by Calcineurin Signaling*

The central role of calcium in the control of cardiac function is well established. The release and reuptake of calcium by the sarcoplasmic reticulum (SR) controls not only the contraction of the cardiac myocyte but, ultimately, the heartbeat. However, the role of calcium as a key second messenger in signal transduction pathways that control growth of the heart has only recently been recognized. The calcium-dependent protein phosphatase calcineurin, also called protein phosphatase 2B, is a critical transducer of calcium signals that govern cardiac growth during development and disease. The actions of calcineurin are dependent on an array of effector proteins that influence its enzymatic activity, subcellular distribution, and stability. Additional proteins transmit calcineurindependent signals to the nucleus with consequent changes in gene transcription. Calcineurin signaling affects the functions of a wide range of cell types and although many of its effectors are ubiquitous, others are restricted to cardiac (and skeletal) muscle, providing muscle specificity to calcineurin signaling. The diversity of calcineurin effectors provides entry points into the signaling pathways that govern cardiac growth and function and provides opportunities for pharmacological and genetic modification of these processes. Here we discuss the functions of calcineurin and its effectors, the possibilities and pitfalls involved in its modulation as a therapeutic target in the heart, and questions for the future.

hypertrophy in response to pathological stimuli is thought to be an initial salutary response to normalize ventricular wall stress and sustain cardiac output, chronic hypertrophy can progress to dilated cardiomyopathy with associated fibrosis, arrhythmias, and sudden death. Indeed, hypertrophy is a strong predictor for morbidity and mortality and is the single most important risk factor for heart failure in humans (3). Whether physiological hypertrophy and pathological hypertrophy are controlled by the same or different pathways is a central issue in the field with significant clinical implications because enhancing the former and suppressing the latter could have profound therapeutic consequences.

Calcium-dependent Activation of Calcineurin Signaling
Calcineurin is a calcium/calmodulin-dependent serine/threonine protein phosphatase composed of a catalytic A subunit (CnA) and a regulatory B subunit (CnB) (reviewed in Ref. 4). It can be distinguished from the more abundant protein phosphatases 1 and 2A by its sensitivity to inhibition by the immunosuppressants cyclosporin A (CsA) and FK-506, and insensitivity to okadaic acid and calyculin A. Calcineurin is also unique in its specific responsiveness to sustained, low frequency calcium signals. The activation of calcineurin occurs through the binding of calcium/calmodulin, which displaces an autoinhibitory domain of the CnA subunit (Fig. 2).
The most well characterized substrate of calcineurin is the transcription factor nuclear factor of activated T-cells (NF-AT), so named because of its role in the activation of T-cells during the immune response (5). Calcineurin dephosphorylates multiple serine residues near the N termini of NF-AT proteins leading to their translocation from the cytoplasm to the nucleus where they engage a variety of transcription factors and activate calcineurin-responsive genes (6 -8). Rephosphorylation of the same sites by glycogen synthase kinase-3␤ (GSK3␤) and other kinases promotes nuclear export of NF-AT and terminates the calcineurin signal to the nucleus.

Calcineurin and Cardiac Hypertrophy
Pathological cardiac hypertrophy is coupled to the activation of a "fetal" gene program in which proteins involved in contractility, calcium handling, and energy metabolism of the fetal heart are up-regulated with consequent changes in cardiac function. Members of the GATA and myocyte enhancer factor-2 (MEF2) families of transcription factors regulate, either directly or indirectly, fetal cardiac genes before birth and in response to stress signals in the adult heart (9,10). The link between calcium and calcineurin as a key signal transduction pathway in the heart was first recognized by the discovery that NF-ATc4, one of four NF-AT proteins, associates with GATA4, a zinc finger transcription factor (11), thereby providing a potential connection between calcium-dependent signaling and fetal gene activation.
The initial description of calcineurin as a mediator of cardiac hypertrophy demonstrated that overexpression of an activated CnA subunit under control of the cardiac-specific ␣-myosin heavy chain (␣-MHC) promoter in transgenic mice led to severe cardiac hypertrophy and eventually to heart failure and sudden death (11). This phenotype could be largely recapitulated by cardiac expression of a constitutively nuclear form of NF-ATc4. Subsequent studies demonstrated that inhibition of calcineurin activity with CsA and FK-506 suppressed hypertrophy in response to a variety of pathological stimuli, although conflicting results have been reported (summarized in Ref. 12). Targeted disruption of the gene encoding the CnA␤ isoform has also provided support for the role of calcineurin in pathological hypertrophy (13). CnA␤ null mice have smaller hearts than normal and show a markedly impaired hypertrophic response to infusion of angiotensin II or isoproterenol as well as to aortic constriction.
The measurable activity of calcineurin is elevated in response to several hypertrophic stimuli including ␤-adrenergic infusion, pressure overload, and exercise (14 -16). Increased calcineurin activity has also been shown to accompany hypertrophy in response to overexpression of K v 4.2N, a cardiac calcium channel (17). Increases in calcineurin activity often correlate with increases in calcineurin protein levels (18); whether this reflects enhanced stability of the calcineurin protein or transcriptional activation of the calcineurin gene, or both, has not been determined.
Sustained intracellular calcium concentrations rather than transient pulses are required to maintain nuclear localization of NF-AT (19). Given that intracellular calcium levels in cardiomyocytes change by 10-fold with every heartbeat, a pertinent question is how calcineurin activity might be chronically regulated in vivo. This may not occur through a global intracellular rise in calcium but perhaps through increases in calcineurin protein levels as discussed above, local changes in intracellular calcium pools, or through the actions of modulatory proteins.
Despite the wealth of information linking calcineurin and the progression of cardiac hypertrophy, little is known of the exact mechanism by which calcineurin promotes growth of the cardiac myocyte, as few calcineurin substrates other than NF-AT that may contribute to the etiology of hypertrophy have been identified. Although activation of NF-AT appears to be sufficient to promote hypertrophy and NF-ATc3 null mice have a blunted hypertrophic response (20), NF-AT gene targets in the cardiac myocyte also remain to be identified. NF-AT may cooperate with GATA4 in the activation of hypertrophic gene markers (10, 21). There is also evidence from other cell types that NF-AT may directly interact with MEF2 (22, 23), a MADS (MCMI, agamous, deficiens, serum response factor) box transcription factor activated during cardiac hypertrophy (9).

Calcineurin and Calcium Handling in the Myocardium
During heart failure, signaling through the ␤-adrenergic pathway decreases. Normally, ␤-adrenergic signaling results in phosphorylation via protein kinase A of PLB, an inhibitor of SERCA2a. Phosphorylation of PLB relieves inhibition of SERCA2a and increases contractility (24). Calcineurin has been shown to dephosphorylate PLB in vitro (25). In addition, SERCA2a activity was decreased in human cardiac lysates in the presence of exogenous calcineurin. Paradoxically, however, the levels of phosphorylated PLB have been reported to be increased in the hearts of ␣-MHCcalcineurin transgenic mice (26); this was correlated with an increased calcium uptake by the SR and improved contractility in isolated cardiac myocytes. Calcineurin has also been shown to interact with the RyR in the SR in a calcium-dependent manner (27), although it was not detected by another group in an isolated RyR protein complex (28).
CsA treatment suppresses cardiac hypertrophy in response to a variety of hypertrophic stimuli. However, CsA treatment of a mouse model with a mutation in ␣-MHC resembling human familial hypertrophic cardiomyopathy resulted in decreased survival with a concomitant increase in heart size (29,30). The mutant MHC protein has been proposed to act as an ion trap, decreasing SR calcium concentrations. If calcineurin promotes calcium uptake by the SR as discussed below, calcineurin inhibition by CsA may further decrease cardiac contractility of these mutant mice resulting in the observed decrease in survival.

Connections between Calcineurin and Other
Hypertrophic Signaling Pathways Connections between calcineurin and other signal transduction pathways governing cardiac myocyte growth may also contribute to the growth-promoting effects of calcineurin. Protein kinase C ␣, ␤ 1 , and ␥ activation was found to be increased in the hearts of ␣-MHC-CnA transgenic mice (31). Furthermore, CsA prevented protein kinase C activation in rats following aortic constriction. Calcineurin also affects signaling through the mitogen-activated protein kinase pathway. Activation of c-Jun NH 2 -terminal kinase is enhanced whereas stress-responsive mitogen-activated protein kinase p38 activity is decreased by calcineurin (31,32). Lower p38 activity correlated with increased protein levels of the dual specificity phosphatase MKP1, a protein thought to be responsible for p38 inactivation. Levels of ERK1/2 activation appear to be largely unaffected by calcineurin. In addition, p38 inhibition in vivo enhanced signaling through the calcineurin/NF-AT pathway suggesting cross-talk between the two pathways (33).
Calcineurin has also been shown to increase the activity of an unidentified kinase that phosphorylates class II histone deacetylases (HDACs) (34). Class II HDACs act as signal-responsive repressors of cardiac hypertrophy, at least in part by interacting with MEF2 and suppressing its ability to activate the fetal gene program. Phosphorylation of class II HDACs leads to their inactivation through nuclear export and promotion of the cardiac hypertrophic program. It is unclear, however, if calcineurin directly regulates the activity of this HDAC kinase or whether activation of the hypertrophic growth program leads indirectly to its enhanced activity, through autocrine feedback loops, for example.

Endogenous Modulators of Calcineurin
Recently, several endogenous protein inhibitors of calcineurin have been described. These include AKAP-79, cabin-1/cain, and the calcineurin B homology protein (35)(36)(37). In addition to their interest as biological modulators of calcineurin activity, these molecules have proven useful in studying the role of calcineurin in cardiac disease. Although these proteins are not expressed specifically in the heart, cardiac overexpression of the calcineurin inhibitory portions of the Cain and AKAP-79 proteins blunted the hypertrophic response to isoproterenol infusion and aortic banding (16). Calcineurin modulators with defined functions in the cardiomyocyte are discussed below.

MCIPs
Modulatory calcineurin interacting protein 1 (MCIP1) was originally identified as the product of the Down syndrome critical region-1 gene on chromosome 21 (38). Its connection to calcineurin was discovered through identification of the yeast homolog Rcn1p (39), which functions as a high copy suppressor of calcineurin activity. MCIP1 is distinct from other protein modulators of calcineurin because of its enriched expression in striated muscles and its regulation by calcineurin signaling, which up-regulates its expression through a series of NF-AT binding sites in its gene promoter (40). This observation led to the hypothesis that MCIP1 participated in a negative feedback inhibition loop to control calcineurin activity (Fig. 3). Consistent with this notion, overexpression of MCIP1 in the heart suppresses hypertrophy in response to activated calcineurin, as well as to isoproterenol infusion, aortic banding, and voluntary wheel running (41,42). Overexpression of MCIP1 also improves the mortality rate of transgenic mice that develop dilated cardiomyopathy in response to cardiac overexpression of activated MEK5, 2 suggesting that inhibition of calcineurin during later stages of heart failure may be beneficial as well (43).
The finding that mice with a targeted disruption of the MCIP1 gene show an exaggerated hypertrophic response when crossed to ␣-MHC-CnA transgenic mice also supported the notion that MCIP acts as a suppressor of pathological levels of calcineurin signaling (44). However, MCIP1-null mice display a diminished cardiac hypertrophic response to isoproterenol infusion and transverse aortic banding, suggesting that MCIP1 has both inhibitory and facilitative effects on calcineurin activity. Such a dual role of MCIP is analogous to the function of Rcn1 in yeast, because yeast with a disrupted Rcn1 gene also have defects in calcineurin signaling (39). How MCIP potentiates calcineurin activity is unclear. Perhaps it serves a chaperone-like function or is required to properly localize calcineurin within the cell. In addition, because calcineurin bound by MCIP is catalytically inactive, mechanisms must exist to relieve this repression. Nevertheless, the above findings suggest that the stoichiometry between calcineurin and MCIP1 is likely to be critical in determining whether MCIP1 acts as an inhibitor or facilitator of calcineurin activity.

Calsarcins
In addition to directly regulating calcineurin activity, other proteins may serve to properly localize calcineurin to discreet subcellular locations. A recently identified family of sarcomeric proteins, the calsarcins, appears to localize calcineurin to the Z-line of the sarcomere in skeletal and cardiac muscle (45,46). Calsarcin-1 expression is limited to cardiac and skeletal muscle, whereas calsarcin-2 and -3 are skeletal muscle-specific. The interaction of calsarcin with ␣-actinin, an integral component of the Z-line, appears to be responsible for this localization. Both ␣-actinin and CnA binding domains have been mapped on the calsarcin protein.
The functional consequence of this subcellular localization of calcineurin is unknown; however, mutations in several Z-line proteins have been implicated in the development of dilated cardiomyopathy.

Glycogen Synthase Kinase-3
As described above, GSK3␤ antagonizes the actions of calcineurin by directly phosphorylating NF-AT and stimulating its nuclear export (47). Several hypertrophic stimuli have been shown to inhibit GSK3␤ activity, which would have the effect of augmenting calcineurin signaling. Transgenic mice that express a constitutively active form of GSK3␤ in the heart are resistant to hypertrophy in response to calcineurin activation, ␤-adrenergic infusion, and pressure overload (46). Whether all the anti-hypertrophic effects of GSK3␤ are mediated by NF-AT phosphorylation or whether other GSK3␤ substrates contribute to this effect remains to be determined.

Protein Kinase G
Recent studies have implicated the natriuretic peptide signaling pathway as a negative regulator of cardiac hypertrophy (48). Signaling by the natriuretic peptide receptors activates protein kinase G (PKG), which may represent a cardioprotective mechanism to suppress pathological signaling (49). Expression of activated PKG in cardiac myocytes also prevents hypertrophy in response to calcineurin activation. At least a portion of this effect can be ascribed to suppression of calcium entry through L-type calcium channels (50). PKG prevents nuclear translocation of NF-AT in response to hypertrophic agonists but not in response to constitutively activated calcineurin; this suggests that PKG acts both upstream and downstream of calcineurin in the hypertrophic signaling cascade.

Calcineurin and Cardiac Myocyte Apoptosis
Apoptosis of cardiac myocytes is triggered by a number of different pathological stimuli leading to heart failure (reviewed in Ref. 51), including ischemia/reperfusion injury, dilated cardiomyopathy, and hypertrophic cardiomyopathy. Calcineurin promotes apoptosis through dephosphorylation of the pro-apoptotic protein BAD (52). Inhibition of calcineurin by CsA or FK506 has been shown to limit the infarct size caused by ischemia/reperfusion injury to the brain (53). Furthermore, it was reported that calcineurin had proapoptotic properties in the heart (54). However, in a separate report calcineurin inhibition blocked the protective actions of endothelin-1 in response to hydrogen peroxide-induced apoptosis (55). Expression of the active form of calcineurin has also been shown to partially block the apoptotic effects of staurosporine and 2-deoxyglucose in cardiac myocytes (56). Most recently, a selective NFAT inhibitory peptide, VIVIT, was shown inhibit cardiac hypertrophy in neonatal cardiac myocytes in culture and to increase the degree of apoptosis observed in response to phenylephrine (57). The exact cardioprotective mechanism of calcineurin against myocyte apoptosis is unknown although it appears to require NFAT activation.

Calcineurin and Cardiac Arrhythmias
As previously discussed, overexpression of calcineurin in mice leads to extensive cardiac hypertrophy, heart failure, and sudden death (11). Electrocardiograms from these transgenic mice show recurrent episodes of sustained pleomorphic tachycardia prior to sudden death (58). Cardiac arrhythmias are an important mechanism contributing to the high mortality and sudden death of patients with cardiac ventricular hypertrophy. Many animal models of hypertrophy show a similar phenotype of prolonged action potential duration because of an increase in depolarization currents or a decrease in repolarizing currents (59). The molecular mechanisms that trigger these electrical changes are not known. Studies of transgenic mice overexpressing calcineurin show a decrease in the density of potassium channels (58). Inhibition of calcineurin activity in transgenic mice by administration of CsA not only reduced hypertrophy but also prevented a decrease in a rapidly 2 B. Rothermel, personal communication. activating and inactivating, fast transient outward current. However, because of an observed decrease of the other outward currents by CsA treatment in wild-type animals, a reversal of the decrease in other outward currents was not evident. More work is needed to examine the effects of CsA treatment on action potential. A potential mode of action linking intracellular signaling to electrical remodeling and myocardial hypertrophy may involve calcineurin activity in phosphorylation modifications of ion channels. If this hypothesis is supported, therapeutic approaches to decreasing calcineurin activity may be applied to human electrophysiological pathologies.

Therapeutic Prospects and Pitfalls
The roles of calcineurin in cardiac disease including hypertrophy and failure make it an attractive therapeutic target. Pharmacological or genetic inhibition of calcineurin by overexpressing peptide inhibitors has a demonstrated benefit in preventing cardiac hypertrophy and failure in short term studies in animals. Whether inhibition of calcineurin activity is sufficient to suppress long term changes in the heart in response to stress signaling remains to be determined. The many physiological roles of calcineurin signaling, such as in the immune response, also pose significant challenges to its systemic inhibition as a means of blocking or reversing cardiac disease. Recent advancements in understanding the connections between the calcineurin and other hypertrophic signaling pathways, as well as new discoveries of effectors of calcineurin activity with cardiac specificity, should offer additional entry points into the cellular circuitry underlying cardiac hypertrophy and heart failure and will undoubtedly yield new approaches to therapeutic drug design for the treatment of heart disease.