**2. Host protein cleavages by coxsackieviral proteases contributing to cardiac dysfunction**

#### **2.1 Viral proteases and dilated cardiomyopathy**

CVB3 encodes two viral proteases, 2A and 3C, both of which are cysteine proteases that have chymotrypsin-like activity and play critical roles in successful viral replication (Chau *et al.*, 2007). First, the viral proteases are required to process the large viral polyprotein, a product of mono-cistronic translation of RNA genome, into the individual functional, structural and non-structural proteins. Second, the viral proteases facilitate viral replication by cleaving a number of host proteins that are involved in various cell functions such as transcription, translation, cell signaling, and cellular structure.

Viral myocarditis is originally thought to be an immune response driven disorder. The initial observation that established the importance of direct virus-mediated myocardial damage in the pathogenesis of viral myocarditis was made in severe combined immunodeficient (SCID) mice, which lack functional T and B lymphocytes and yet developed early and severe myocyte damage upon enterovirus challenge (McManus *et al.*, 1993). The significance of viral proteins in the development of viral myocarditis and DCM was further explored by Dr. Knowlton's research group. First, they showed that the cardiacspecific expression of a replication-restricted CVB3 mutant genome in transgenic mice, which only allows the expression of viral proteins without generating viral progenies and subsequent immune response, results in DCM phenotype (Wessely *et al.*, 1998). Then, they demonstrated that mice with cardiac-restricted expression of viral protease 2A display a severe DCM phenotype (Xiong *et al.*, 2007). These findings suggest that viral proteases play an important role in the development of viral-induced dilated cardiomyopathy.

#### **2.2 Cleavage of dystrophin during CVB3 infection may contribute to dilated cardiomyopathy**

The observation that mouse cardiac expression of viral protease 2A induces DCM has been explained based on the landmark finding that dystrophin, which links the cytoskeleton to

CVB3 infection of myocarditis susceptible mice results in severe heart failure. The disease progression of viral myocarditis in the experimental infection model can be classified into three phases: acute (viremia), subacute (inflammatory), and chronic (remodeling) phases. The acute (viremia) phase is signified by active viral replication and direct virus-induced cardiomyocyte damage. The subacute (inflammatory) phase is characterized by the infiltration of immune cells that helps viral clearance but nonetheless adds to myocardial damage. The chronic (remodeling) phase is featured by the continual efforts of the impaired heart to meet the hemodynamic demand by remodeling the myocardium. Cardiac hypertrophy is triggered during remodeling to compensate for reduced contractile function due to myocyte loss and interstitial fibrosis in the earlier phases. However, such an adaption is unsustainable in the longer term in face of increasingly hostile environments, i.e. reduced blood supply and increased reactive oxidative stress, thus leading to cardiomyocyte death and triggering further inflammation and fibrosis. The pathological remodeling process

This book chapter focuses on the virus-host protein interactions in cardiomyocytes during viral myocarditis. We discuss the role of virus-induced protein cleavage and dysregulation of the host protein degradation systems in the pathogenesis of viral myocarditis and its

**2. Host protein cleavages by coxsackieviral proteases contributing to cardiac** 

CVB3 encodes two viral proteases, 2A and 3C, both of which are cysteine proteases that have chymotrypsin-like activity and play critical roles in successful viral replication (Chau *et al.*, 2007). First, the viral proteases are required to process the large viral polyprotein, a product of mono-cistronic translation of RNA genome, into the individual functional, structural and non-structural proteins. Second, the viral proteases facilitate viral replication by cleaving a number of host proteins that are involved in various cell functions such as

Viral myocarditis is originally thought to be an immune response driven disorder. The initial observation that established the importance of direct virus-mediated myocardial damage in the pathogenesis of viral myocarditis was made in severe combined immunodeficient (SCID) mice, which lack functional T and B lymphocytes and yet developed early and severe myocyte damage upon enterovirus challenge (McManus *et al.*, 1993). The significance of viral proteins in the development of viral myocarditis and DCM was further explored by Dr. Knowlton's research group. First, they showed that the cardiacspecific expression of a replication-restricted CVB3 mutant genome in transgenic mice, which only allows the expression of viral proteins without generating viral progenies and subsequent immune response, results in DCM phenotype (Wessely *et al.*, 1998). Then, they demonstrated that mice with cardiac-restricted expression of viral protease 2A display a severe DCM phenotype (Xiong *et al.*, 2007). These findings suggest that viral proteases play

an important role in the development of viral-induced dilated cardiomyopathy.

**2.2 Cleavage of dystrophin during CVB3 infection may contribute to dilated** 

The observation that mouse cardiac expression of viral protease 2A induces DCM has been explained based on the landmark finding that dystrophin, which links the cytoskeleton to

eventually leads to DCM and heart failure.

**2.1 Viral proteases and dilated cardiomyopathy** 

transcription, translation, cell signaling, and cellular structure.

subsequent progression to DCM.

**dysfunction** 

**cardiomyopathy** 

the extracellular matrix (ECM) by forming the dystrophin glycoprotein complex (DGC), is cleaved during CVB3 infection by protease 2A (Badorff *et al.*, 1999; Badorff & Knowlton, 2004) (Fig. 1A). Dystrophin has three domains that serve different purposes. Its N-terminal domain anchors to the actin cytoskeleton and its rod domain provides the linkage to the Cterminal domain, which binds to -dystroglycan that in turn connects to the sarcolemma and the extracellular matrix (Fig. 1A). Furthermore, dystrophin-deficient mice have been shown to have an increased susceptibility to viral myocarditis and develop severe cardiomyopathy (Xiong *et al.*, 2002). Human genetic mutations of the dystrophin gene cause Duchenne Muscular Dystrophy (DMD) (Nigro *et al.*, 1990). Approximately 20% of DMD patients suffer and die from a resultant cardiomyopathy. Other mutations of the dystrophin gene also cause X-linked DCM (Ferlini *et al.*, 1999). CVB3-induced dystrophin cleavage occurs at its 3' hinge and therefore breaks its connection to the ECM. As a result, the sarcolemmal integrity is compromised and force transmission is reduced. This can further lead to cardiomyocyte necrosis due to the increased sarcolemmal permeability (Fig. 1A). Thereafter, dystrophin cleavage has been viewed as a major mechanism in enteroviral cardiomyopathy. However, dystrophin knockout mice (Mdx) display only mild cardiomyopathy phenotype, due to the compensatory upregulation of utrophin, a dystrophin homologue (Deconinck *et al.*, 1997; Grady *et al.*, 1997). Thus, other mechanisms may also contribute to the severe cardiomyopathy phenotype in viral protease 2A expressing mice.

#### **2.3 Cleavage of serum response factor by viral protease 2A is associated with impaired cardiac function**

Recent efforts have demonstrated for the first time that serum response factor (SRF) is cleaved in CVB3-infected mouse hearts and cultured murine cardiomyocytes (unpublished). SRF, which belongs to the MADS-box (MCM1, Agamous, Deficiens, and SRF) protein superfamily, is a muscle-enriched transcription factor that regulates the expression of contractile and regulatory genes, as well as microRNAs (miRNAs) (Miano, 2003; Niu *et al.*, 2007; Oka *et al.*, 2007) (Fig. 1B). SRF interacts with tissue specific cofactors such as myocardin, Nkx2.5, c-Fos, and binds to the serum response element (SRE) of its target genes. It contains two major domains: the N-terminal DNA binding and dimerization domain and the C-terminal transactivation domain (Fig. 1B). Genomic studies have identified over 1200 SRE containing genes and more than 250 of these have been verified (Sun *et al.*, 2006). Cardiac contractile genes under SRF regulation include cardiac -actin, -myosin heavy chain, myosin light chain, cardiac troponin I, etc. SRF is indispensable for mesoderm formation and plays a central role in cardiac development and function (Niu *et al.*, 2007). Therefore, SRF knockout results in embryonic lethality (Parlakian *et al.*, 2004). The construction of cardiac-specific inducible SRF knockout transgenic mice overcomes this problem and helps illustrate the importance of SRF in cardiac function (Parlakian *et al.*, 2005). It was shown that SRF knockout in the adult mouse heart results in damaged cardiac function, and subsequent progression to DCM. Genetic mutations of SRF in humans have not been described likely due to the associated lethality. However, the expression of alternatively spliced SRF isoforms, which was shown to have inhibitory effects on wild-type SRF, is increased in failing human and animal hearts (Davis *et al.*, 2002). Furthermore, a cleaved form of SRF lacking the transactivation domain was also found in human failing hearts as a result of caspase-3 activation during cardiomyocyte apoptosis (Chang *et al.*,

Impaired Cardiac Function in Viral Myocarditis 297

Fig. 1. Host protein cleavages by coxsackieviral proteases in viral myocarditis.

the heart

force transmission and increasing sarcolemmal permeability.

A. Coxsackieviral protease 2A cleaves dystrophin at its 3' hinge. Dystrophin is a component of the dystrophin-glycoprotein complex that links the cytoskeleton to the extracellular matrix. Dystrophin cleavage contributes to myocyte dysfunction by reducing contractile

B. Viral protease 2A also cleaves serum response factor (SRF). SRF is a muscle-enriched transcription factor that regulates the expression of cardiac regulatory proteins, sarcomere contractile proteins, as well as cardiac-specific microRNAs (miRNAs). SRF associates with co-factors such as GATA4, Nkx2.5, MEF2, and myocardin and binds to serum response element (SRE) (also known as CArG box) to activate gene transcription. SRF cleavage results in myocyte dysfunction by the dissociation of the N-terminal DNA binding/dimerization domain from the C-terminal transcriptional activation domain, thus abolishing SRFmediated gene expression. Furthermore, the N-terminal fragment (SRF-N) exhibits a dominant-negative effect on endogenous SRF function by competing for DNA binding.

2003). This caspase-cleaved SRF fragment functions as a dominant-negative mutant that inhibits SRF-dependent activation of cardiac genes. On the other hand, SRF overexpression in transgenic mouse hearts leads to the development of cardiac hypertrophy and subsequent cardiomyopathy (Zhang *et al.*, 2001).

During CVB3 infection, SRF is cleaved into a ~47kD N-terminal fragment (SRF-N) and a ~20kD C-terminal fragment (SRF-C) by viral protease 2A (unpublished). As a result, the DNA-binding domain is detached from the transactivation domain, thus abolishing SRF function. In addition, the SRF-N fragment generated from the cleavage can compete with wild-type SRF for target gene binding, and thereby exhibits dominant-negative suppression of SRF target gene expression (Fig. 1B). This study suggests another important mechanism by which CVB3 damages cardiac function and leads to subsequent DCM. Further research using knock-in transgenic mice expressing non-cleavable SRF will help clarify the relative contribution of SRF cleavage in the pathogenesis of viral myocarditis.

#### **2.4 Caspase activation by viral proteases**

Both viral proteases 2A and 3C lead to late onset of host cell apoptosis through the direct cleavage of caspase-8 and the activation of caspase-3, as well as the cleavage of antiapoptotic protein Bid that leads to mitochondrial cytochrome c release and subsequent initiation of the caspase cascade (Chau *et al.*, 2007). Cardiomyocyte apoptosis is a common phenomenon in various cardiac diseases. Apoptosis, also known as programmed cell death, is the self-destruction pathway activated when host cells decide to commit suicide. Apoptosis during viral myocarditis caused by viral protease-induced caspase activation, however, is "switched on" intentionally by viral mechanisms for the release of mature viral progenies. In cell culture, caspase activation results in cytoplasmic proteolysis and DNA fragmentation. However, despite ultrastructural evidence of cytochrome c release detected in many cardiomyocytes of heart failure patients, intact nuclei are seen in all of these myocytes (Narula *et al.*, 1998). This suggests that the terminally differentiated cardiomyocytes have evolved strategies to resist nuclear fragmentation despite ongoing cytoplasmic apoptosis. In fact, the number of apoptotic myocytes in a cardiomyopathic heart ranges only from 0.07% to 0.7% as compared to 0.003% in the normal myocardium (Narula *et al.*, 1998). Nevertheless, cytoplasmic apoptosis initiated during viral myocarditis may compromise mitochondrial ATP generation, as well as cause destruction to contractile proteins which add to systolic dysfunction in the disease pathogenesis.

#### **2.5 Other mechanisms**

Other mechanisms that viral proteases use to impair cardiomyocyte function include the interference of host gene transcription by the cleavage of cyclic AMP response element binding protein (CREB) (Yalamanchili *et al.*, 1997), the disruption of host protein translation through the cleavage of eukaryotic translation initiation factor 4 (eIF4) (Chau *et al.*, 2007) and eIF5B (de Breyne *et al.*, 2008), the interception of cell signaling pathways via the cleavage of RasGAP (Huber *et al.*, 1999), and the weakening of the cytoskeletal network by the cleavage of cytokeratin-8 (Seipelt *et al.*, 2000). Although there are no known cardiac diseases associated with the aforementioned proteins, their cleavages exert additive effects to the final detriment of the infected cardiomyocytes.

2003). This caspase-cleaved SRF fragment functions as a dominant-negative mutant that inhibits SRF-dependent activation of cardiac genes. On the other hand, SRF overexpression in transgenic mouse hearts leads to the development of cardiac hypertrophy and subsequent

During CVB3 infection, SRF is cleaved into a ~47kD N-terminal fragment (SRF-N) and a ~20kD C-terminal fragment (SRF-C) by viral protease 2A (unpublished). As a result, the DNA-binding domain is detached from the transactivation domain, thus abolishing SRF function. In addition, the SRF-N fragment generated from the cleavage can compete with wild-type SRF for target gene binding, and thereby exhibits dominant-negative suppression of SRF target gene expression (Fig. 1B). This study suggests another important mechanism by which CVB3 damages cardiac function and leads to subsequent DCM. Further research using knock-in transgenic mice expressing non-cleavable SRF will help clarify the relative

Both viral proteases 2A and 3C lead to late onset of host cell apoptosis through the direct cleavage of caspase-8 and the activation of caspase-3, as well as the cleavage of antiapoptotic protein Bid that leads to mitochondrial cytochrome c release and subsequent initiation of the caspase cascade (Chau *et al.*, 2007). Cardiomyocyte apoptosis is a common phenomenon in various cardiac diseases. Apoptosis, also known as programmed cell death, is the self-destruction pathway activated when host cells decide to commit suicide. Apoptosis during viral myocarditis caused by viral protease-induced caspase activation, however, is "switched on" intentionally by viral mechanisms for the release of mature viral progenies. In cell culture, caspase activation results in cytoplasmic proteolysis and DNA fragmentation. However, despite ultrastructural evidence of cytochrome c release detected in many cardiomyocytes of heart failure patients, intact nuclei are seen in all of these myocytes (Narula *et al.*, 1998). This suggests that the terminally differentiated cardiomyocytes have evolved strategies to resist nuclear fragmentation despite ongoing cytoplasmic apoptosis. In fact, the number of apoptotic myocytes in a cardiomyopathic heart ranges only from 0.07% to 0.7% as compared to 0.003% in the normal myocardium (Narula *et al.*, 1998). Nevertheless, cytoplasmic apoptosis initiated during viral myocarditis may compromise mitochondrial ATP generation, as well as cause destruction to contractile

Other mechanisms that viral proteases use to impair cardiomyocyte function include the interference of host gene transcription by the cleavage of cyclic AMP response element binding protein (CREB) (Yalamanchili *et al.*, 1997), the disruption of host protein translation through the cleavage of eukaryotic translation initiation factor 4 (eIF4) (Chau *et al.*, 2007) and eIF5B (de Breyne *et al.*, 2008), the interception of cell signaling pathways via the cleavage of RasGAP (Huber *et al.*, 1999), and the weakening of the cytoskeletal network by the cleavage of cytokeratin-8 (Seipelt *et al.*, 2000). Although there are no known cardiac diseases associated with the aforementioned proteins, their cleavages exert additive effects

contribution of SRF cleavage in the pathogenesis of viral myocarditis.

proteins which add to systolic dysfunction in the disease pathogenesis.

to the final detriment of the infected cardiomyocytes.

cardiomyopathy (Zhang *et al.*, 2001).

**2.4 Caspase activation by viral proteases** 

**2.5 Other mechanisms** 

Fig. 1. Host protein cleavages by coxsackieviral proteases in viral myocarditis. A. Coxsackieviral protease 2A cleaves dystrophin at its 3' hinge. Dystrophin is a component of the dystrophin-glycoprotein complex that links the cytoskeleton to the extracellular matrix. Dystrophin cleavage contributes to myocyte dysfunction by reducing contractile force transmission and increasing sarcolemmal permeability.

B. Viral protease 2A also cleaves serum response factor (SRF). SRF is a muscle-enriched transcription factor that regulates the expression of cardiac regulatory proteins, sarcomere contractile proteins, as well as cardiac-specific microRNAs (miRNAs). SRF associates with co-factors such as GATA4, Nkx2.5, MEF2, and myocardin and binds to serum response element (SRE) (also known as CArG box) to activate gene transcription. SRF cleavage results in myocyte dysfunction by the dissociation of the N-terminal DNA binding/dimerization domain from the C-terminal transcriptional activation domain, thus abolishing SRFmediated gene expression. Furthermore, the N-terminal fragment (SRF-N) exhibits a dominant-negative effect on endogenous SRF function by competing for DNA binding.

Impaired Cardiac Function in Viral Myocarditis 299

docking of polyubiquitinated target protein. 19S also serves to detach and hence recycle the Ub by its deubiquitinating enzyme (DUB) activity. Furthermore, 19S unfolds the target

The immunoproteasome is an alternative version of the proteasome expressed to accommodate inflammatory responses upon stimulation with interferon- (Rivett & Hearn, 2004). The immunoproteasome has a 20S core that substitutes the constitutive catalytic βsubunits with inducible β-counterparts (β1i, β2i, and β5i), which offer different proteolytic function and activity to generate small peptides suitable for antigen presentation by major histocompatibility complex (MHC) class I (Griffin *et al.*, 1998) (Fig. 2A). In addition to the 19S proteasome, the immunoproteasome can also have a different lid(s) – the 11S proteasome, also known as PA28 (proteasome activator 28). Different compositions of 11S exist: heteroheptamer of PA28 and PA28 that are induced by interferon- under intensified immune response (Murray *et al.*, 2000) and homoheptamer of PA28 that resides in the nucleus and assists ATP- and ubiquitin-independent proteasomal activity (Mao *et al.*, 2008). Sometimes, hybrid proteasomes with both 11S and 19S lids are also observed.

UPS dysregulation is a common phenomenon in heart diseases. It is accentuated with the accumulation of Ub-protein conjugates and is associated with markedly reduced proteasome proteolytic activity in failing human hearts as compared to non-failing hearts (Predmore *et al.*, 2010). This suggests that ubiquitinated proteins in hearts are not degraded due to impaired proteasomal function. While no changes were noted in protein expression of proteasome subunits (i.e. 20S, 19S, 11S), elevated levels of protein carbonyls and 4-hydroxynonenylated proteins were observed in failing hearts. Also, oxidative modification to the 19S ATPase subunit Rpt5 was found in these failing hearts. Together, these oxidative modifications to proteasome subunits and substrate proteins may lead to impaired proteasomal function. On the other hand, microarray studies demonstrate reduced transcript levels of some 20s α- and β-subunits in the failing hearts as compared to controls (Hwang *et al.*, 2002; Kaab *et al.*, 2004). The incongruence between protein and mRNA expression of proteasome subunits may be

Animal models of cardiac diseases also have an increased ubiquitinated protein expression, but are acommpanied with changes in their proteasome expression profile. Upregulation in protein expression of proteasome subunits was observed in a left ventricular hypertrophy mouse model (Depre *et al.*, 2006). Post-translational modifications of the proteasome subunits were also reported in these hypertrophic hearts (Depre *et al.*, 2006). Treatment with proteasome inhibitor effectively prevents cardiac hypertrophy development, suggesting that the upregulation of proteasome expression is central to this physiological adaptation. Similar beneficial effects of proteasome inhibition in the regression of cardiac hypertrophy were observed in other studies (Meiners *et al.*, 2008; Stansfield *et al.*, 2008). Besides hypertrophic cardiomyopathy, the accumulation of Ub-conjugated proteins was observed in hyperglycemia-induced cardiomyopathy mouse model (Powell *et al.*, 2008). A parallel drop in the basal ATP-dependent proteasomal activity was observed in these mice. However, an increased ATP-independent chymotryptic proteasomal activity was observed, which is accompanied by an increased expression of 11S lid subunits PA28 and PA28, as well as

protein and feeds it to the 20S core for degradation.

However, their functions remain to be explored.

attributed to myocyte loss and fibrosis in the failing hearts.

**3.2.1 The UPS and heart diseases** 
