Hold me tight: The role of the HSP family of chaperones in cardiac disease

Introduction

During the development of cardiac hypertrophy, heart failure, and ischemia reperfusion challenge, the heart accumulates misfolded proteins as a result of cellular stresses ^1–4. While the compensatory increases in chaperones/co-chaperones work to prevent misfolding, refold denatured proteins and/or target them for degradation, this system can become overwhelmed, leading to worsening of cardiac function. In fact, recent studies have demonstrated experimentally that increasing the burden of misfolded proteins in the heart can contribute to the development of cardiac dysfunction ⁵. In this review, we discuss the role of heat shock proteins in common cardiac diseases including cardiac hypertrophy, heart failure, and ischemia/reperfusion injury. Furthermore, we delineate the many specific mechanisms by which these chaperones, co-chaperones, and heat shock factor (HSF) transcription factors have been found to be cardioprotective in experimental models. Lastly, we review recent studies involving drugs that are being developed (and currently used) to increase the expression (and presumably function) of chaperone/co-chaperone systems that may be applicable to the treatment of common cardiac diseases as well as familial cardiac diseases whose etiology includes a major component of misfolded proteins (e.g. desminopathies).

Chaperones enhance productive protein folding and refolding and prevent protein aggregation

There are several general families of molecular chaperones in the cytoplasm of mammalian cells, including heat shock protein (HSP) 90, HSP70, CCT (also called TRiC), and small HSP (sHSP) family proteins (Figure 1A). Members of the HSP90 family of chaperones are the most abundant chaperones located in the cytosol. They form dimers consisting of HSP90α and HSP90β subunits and are inducible with stress, although are quite abundant without stress also ^{8, 9}. HSP90 assists many proteins involved with signal transduction, including more than 40 kinases and many steroid hormone receptors, with a supporting role in conformational changes involved in ATP hydrolysis ^{10, 11}. The HSP70 chaperone family consists of 6 member proteins that are found in the cytosol ¹², including HSP70 and the cognate of HSP70 (HSC70). Like HSP90, HSP70 proteins are inducible with stress, however they are also highly abundant without stress. HSP70 family members are functionally highly homologous, recognizing hydrophobic surfaces of unfolded proteins and partially folded intermediates. Their activity is controlled by their hydrolysis activity as well as by their ability to bind ATP ¹³. HSP90 and HSP70 proteins both inhibit protein aggregation, thereby promoting productive folding of proteins (for comprehensive reviews, see Young et al.^14–16).

Figure 1. The regulation of the protein quality control and chaperone systems in the heart that protect against the toxicity of misfolded proteins.

A. Both NF-kB and HSF transcription factors regulate the expression of key molecular chaperones in the mitochondria and cytosol, respectively. Stress-activated HSF1 and HSF2 also increase the expression of co-chaperones and ubiquitin expression. B. The upregulation of both chaperones and ubiquitin help maintain protein quality control by either refolding (chaperone) the protein or ubiquitinating misfolded proteins (co-chaperone, e.g. CHIP). This is a dynamic process particularly important during cardiac stress. Adapted from: Kubota 2009 ⁶ and Wang et al. ⁷

Other molecular chaperones present in the cytosol include TRiC (TCP1-ring complex, or chaperonin containing TCP1 (CCT)) ^{17, 18}. The central cavity of TRiC uses a lid-like structure to encapsulate substrate proteins to allow it to trap and fold target proteins. This encapsulation prevents aggregation and allows changes in conformation that ensure correct folding (in an ATP-dependent manner) before substrates are released. The family of sHSPs also maintain protein confirmation in an ATP-independent manner ¹⁹. Over 10 sHSP chaperone proteins have been characterized and all function by maintaining an equilibrium between the dimer and large oligomer states of their target proteins ²⁰. In contrast to HSP90 and 70 though, only a few sHSPs (including HSP27, HSP22, and alphaB-crystallin) are increased in response to stress ²¹. HSP27 and alphaB-crystallin (CryAB) are abundant in cardiac and skeletal muscles and increase in response to stress in order to protect against insults such as ischemia ²². Both these proteins associate with actin and are vital to muscle development and assembly ²³. CryAB also interacts with several other cytoskeletal proteins such as desmin to maintain protein folding and prevent aggregation ^24–27.

In addition to the molecular chaperones that are found predominantly in the cytosol, there are also chaperones that are known to maintain proteins in other compartments of the cell, in particular the mitochondria. HSP60 is a chaperone originally identified in the mitochondria ²⁸ but which is also found in the cytosol ²⁹. HSP60 is responsible for refolding and transportation of proteins between the mitochondrial matrix and the cytoplasm of the cell ³⁰ and associates with a number of cytosolic proteins involved in apoptosis such as B-cell lympohoma-1-assocated X protein (BAX), B-cell lymphoma (BCL)-xl and BCL-2 homologous antagonist/killer (BAK) ^{31, 32}. HSP60 is believed to be a homolog of the bacterial heat shock protein groEL and, as such, it is believed that HSP60 assists in folding linear amino acids into their three dimensional structures in ways that have been described for groEL ³³.

Co-chaperones control the activity of chaperones: DnaJ, BAGs, Hop, CHIP, Immunophilins, and Prefoldin

As their name implies, co-chaperones are proteins that assist chaperone functions, including in protein folding. There are over 40 genes in the HSP40/DnaJ protein family, with the different members playing diverse roles in regulating protein folding, assembly, translocation, and even degradation. DnaJ proteins bind to the ATPase binding domain of HSP70 proteins to enhance their ATPase activity ³⁴. They also bind substrate proteins to modulate folding in a substrate-specific manner. Bcl2-associated athanogene (BAG) proteins all have a conserved BAG region that also binds the ATPase domain of HSP70 proteins to affect the rate of substrate binding and release ³⁵. Conversely, Hop, carboxy terminus of HSP70-binding protein (CHIP) and immunophilins all bind both HSP70 and HSP90 via tetra-tricopeptide repeat (TPR) domains, which allow the transfer of substrates between them ^{8, 11}. CHIP is critical to quality control processes and ubiquitinates misfolded proteins when correct folding cannot be achieved (see Figure 1B) ^{36, 37}, while immunophilins are necessary for the functions of the p23 steroid aporeceptor-associated protein ¹¹. Lastly, prefoldin, also known as GimC, helps TriC/CCT-dependent folding of tubulin and actin by way of its 6 tentacle-like processes that trap unfolded substrates and assist with folding in collaboration with HSP70/HSP90^{38, 39}.

The transcription factors HSF1 and HSF2 regulate chaperone and co-chaperone expression in the cell

The number of misfolded proteins increases during times of cellular stress, including oxidative stress and proteasome inhibition. Therefore, many chaperones are induced at the transcriptional level in the presence of these conditions to protect against the toxicity of misfolded proteins. There are at least 4 transcription factors (HSF1-4) that regulate heat shock proteins. HSF1 is the primary transcription factor involved in this process and binds heat shock response elements (HSE) in the promoter regions of stress-induced genes. HSF1 is found throughout the cytosol as a monomer that binds HSP90 to inhibit its chaperone activity. During stress, denatured proteins competitively bind HSP90 ⁴⁰, effectively releasing HSF1, thereby allowing it to translocate to the nucleus as trimers ⁴¹. In the nucleus, increased HSF1 upregulates the expression of chaperones, including HSP70. HSP70 then binds HSF1, which in turn, attenuates the HSF response, indicating a feedback mechanism ⁴². Interestingly, part of the stress response induced by HSF1 includes the upregulation of ubiquitin, suggesting that it regulates the ubiquitin proteasome system to enhance the cell’s capacity to degrade proteins during stress (recently reviewed by Willis et al.⁴³). The roles of the remaining three HSFs in the stress response are less well studied. Studies have shown that HSF2 contributes to the inducible expression of heat shock protein genes by interacting with HSF1 ⁴⁴. The roles of HSF3 and HSF4 as factors that regulate the expression of non-classical and sHSPs are just beginning to be understood ^{45, 46}.

Cells have a diverse array of molecular chaperones available to them in the cytosol and mitochondrial compartments that are regulated by the transcription factors NF-κB and HSF1-4. While these chaperones primarily play a role in refolding proteins, some co-chaperones cross talk with the ubiquitin proteasome system to directly ubiquitinate proteins for subsequent degradation by the proteasome. While much of the work on molecular chaperones has been in non-cardiovascular systems, their role in the heart during health and disease is becoming more greatly appreciated.

Cardiac chaperones are regulated in the pathophysiology of a number of common cardiac diseases

Cardiac hypertrophy

Cardiac chaperones such as HSP70, CryAB, and HSP22 (alphaC-crystallin) increase in expression during the development of cardiac hypertrophy (Table 1). HSP70 can be induced by a variety of hypertrophic stimuli, including aortic banding, angiotensin II infusion, isoproterenol infusion and swimming ⁵⁷. CryAB can be induced in rat cardiomyocytes in response to endothelin-1 induced cardiomyocyte hypertrophy, resulting in a 2-fold increase in CryAB expression. The DNA binding activity of HSF1 is heightened as a result of increasing pre-load in heart ⁷¹. Since cardiac hypertrophy is a response to a number of stress stimuli, this general increase in cardiac chaperone expression in cardiac hypertrophy is not surprising. However, the purpose behind this increased chaperone expression is much less well understood. Mice with transgenic overexpression of CryAB gene challenged with 2 weeks of aortic banding to induce pathologic cardiac hypertrophy exhibit a significant reduction in NFAT (nuclear factor of activated T cells) transactivation and attenuated hypertrophy ⁸⁶. This contrasts to CryAB null mice, which display an enhanced NFAT transactivation at baseline, and an accelerated development of heart failure in response to pressure overload induced by aortic banding ⁸⁶. These studies demonstrate that cardiac CryAB plays a role in suppressing cardiac hypertrophic response, possibly through inhibiting NFAT signaling.

Table 1. A summary of the known regulation of chaperones, co-chaperones, and transcription factors involved in the heat shock protein response in cardiac hypertrophy, heart failure, and ischemia reperfusion injury.

N.D.=not determined.

	Example substrate recognition/Other	Cardiac Hypertrophy	Heart Failure	Ischemia reperfusion injury
Chaperones
1. HSP90	Actin, tubulin, TGFbeta1, CHIP 47, 48	N.D.	HSP90 in DCM=IHD=Normal Controls 60	Rat cardiac ischemia-reperfusion results in increased HSP90 65 Rat cardiac ischemia-reperfusion results in increased HSP70 65
2. HSP70	Calcineurin, HDAC6, CHIP49, 50	Hypertrophic stimuli (aortic banding, isoproterenol, AngII, swimming) induce HSP70 expression57 HSP70 support pro-hypertrophic signaling via interaction with Hdac2 57	HSP72 not increased in heart failure induced by coronary artery ligation in rats for 9–12 weeks 7 No difference in HSP72 expression8 weeks post-coronary artery occlusion compared to controls.61 HSP70 and HSC70 in DCM=IHD=Normal controls 60	HSP72 increased at 1 week post-coronary artery occlusion 61
3. CCT (TRiC)	TRiC binds Actin 51	N.D.	N.D.	N.D.
4. Small HSPs
-AlphaB-Crystallin (=HSPB5)	AlphaB-Crystallin binds desmin 52	AlphaB-Crystallin increases 2 fold in ET-1 treated neonatal rat cardiomyocytes 58	N.D.	N.D.
-HSP20 (=HSPB6)	HSP20 binds actin and alpha-actinin53	N.D.	N.D.	HSP20 increased in I/R, regulated, in part by miR-32066.
-HSP22 (=HSPB8, H11 kinase, alphaC-Crystallin)	HSP22 binds lipid membranes 54	HSP22 expression increases ~3 fold in cardiac hypertrophy induced in dogs by aortic banding 59	N.D.	HSP22 expression is increased ~3 fold 1 hour after reperfusion following ischemia in a pig model 67 In human hibernating myocardium and swine model of is hibernating myocardial, HSP22 increased 68
-HSP27 (=HSPB1, HSPB2)	HSP27 binds IκB 55	N.D.	No change in HSP27 expression at 8 weeks post-coronary artery occlusion (rat)/Increased HSP27 at 1 week post-coronary artery occlusion 61 HSP27 increased in human DCM compared to controls 60	N.D.
5. Mitochondrial HSPs
-HSP60	HSP60 binds Bax and Bak 56	HSP60 decreased 13 fold in ET-1 treated neonatal rat cardiomyocytes 58	HSP60 increased 8+ weeks (but not at 1 week) post-coronary artery occlusion (rat) 7, 61, 62 There is increased expression of cardiac HSP60 in heart failure (coronary artery ligation in rats for 9–12 weeks); may be driven by NF-κB activation 7. HSP60 doubled in human DCM and IHD compared to controls60 HSP60 moves from cytoplasm to mitochondria in DCM and IHD 63 Serum HSP60 levels are associated with the severity of heart failure in patients 64	N.D.??
Transcription Factors regulating Chaperones
1. HSF1	HSF1 binds the heat shock elements (HGAAN) of HSP72; HSP90 interacts to repress.	Increased per-load/mechanical stress increases HSF1 activity 71	HSF1 levels increased, without an increase in activity in heart failure (8 weeks post-coronary artery ligation in rats) 7	Rat cardiac ischemia-reperfusion results in increased HSF1 activity, but not HSF2 65. HSF1 induction in ischemia mediated by ROS and ATP levels 72, 73
2. HSF2	HSF-2 binds the heat shock elements of HSP90, HSP27, c-Fos69; interacts with HSF1 and Nucleoporin62 70	N.D.	HSF2 levels increased, without an increase in activity in heart failure (8 weeks post-coronary artery ligation in rats) 7
Co-Chaperones
1. DnaJ	DnaJ binds ribosome bound nascent polypeptides 74	N.D.	N.D.	DnaJ-like pDJA1 increased 4 fold after reperfusion in a pig model of I/R82
2. BAG-1	BAG-1 binds Bcl-2, Raf1 75	N.D.	N.D.	BAG-1 protects against I/R injury 83, 84
3. Hop	Hop binds HSP70, HSP90 76	N.D.	N.D.	N.D.
4. carboxyl terminus of HSP70-interacting protein (CHIP)	CHIP binds HSP70, HSP90, HIF1-α77	CHIP increases in response to high glucose and regulates pro-hypertrophic GATA4 in cardiomyocytes in vitro 81	N.D.??	CHIP protects against ischemia reperfusion injury 85
5. Immunophilins	FKBP38 binds mTOR complex 78	N.D.	N.D.	N.D.
6. Prefoldin	Prefoldin binds Nascent chain of actin and tubulin Chaperonin 79, 80	N.D.	N.D.	N.D.

In contrast to CryAB’s ability to attenutae cardiac hypertrophy, HSP70 is necessary for the induction of cardiac hypertrophy. One potential mechanism by which HSP70 may accomplish this is through its association with the activated form of histone deacetylase2 (HDAC2). Overexpressing a dominant negative form of HSP70 or decreasing HDAC2 with siRNA blunts the hypertrophic response in the heart. Furthermore, cardiac hypertrophy induced by isoproterenol infusion or aortic banding in mice lacking HSP70, results in a blunting of HDAC2 activity ⁵⁷. This suggests a role for HSP70 in the induction of cardiac hypertrophy, possibly through its stabilization of HDAC2. This HSP70-dependence of HDAC2 activity is interesting because of the importance of HDAC2 in cardiac hypertrophy signaling. The HDACs are instrumental in regulating hypertrophic gene expression in pathological settings ⁴⁹. Class II HDACs (HDAC 4, 6, 7, 9) negatively regulate hypertrophy by repressing MEF/GATA/NFAT-mediated gene transcription ⁸⁷. Conversely, the Class I HDAC 2 has been implicated in having acetylase-dependent pro-hypertrophic activity, possibly by releasing repression of IGF-1 signaling ^{49, 57, 88}.

Another cardiac chaperone, HSP22, has also been shown to increase during the development of cardiac hypertrophy. HSP22 has several names, the most common one used in the literature is H11 kinase, but alphaC-crystallin is sometimes used and indicates its relationship to other sHSPs (see recent review ⁸⁹). Increasing HSP22 expression results in the activation of signaling pathways involved in survival and cell growth, including the PI3K/Akt pathway, AMPK, PKC epsilon, nitric oxide, and mTOR⁸⁹. These signaling pathways induce pre-conditioning, growth, and protection against apoptosis among other cardioprotective pathways ⁸⁹. Recently, studies have identified that HSP22 increases during the development of cardiac hypertrophy in a variety of animal models. Inducing a slowly progressive cardiac hypertrophy in puppies through use of aortic banding (that parallels the gradual progression of human disease to a much greater extent than acute aortic banding models), causes HSP22 to increase ~3 fold ⁵⁹. Increasing HSP22 expression in cultured cardiomyocytes and the intact mouse heart (~ 7 fold) results in the development of a spontaneous hypertrophy, characterized by the re-expression of the fetal gene program ^{59, 90}. This suggests that the many cardioprotective pathways HSP22 induces ⁸⁹ may be detrimental if expressed chronically.

Physiologic versus Pathologic Hypertrophy

The differences in the underlying signaling between physiologic and pathologic cardiac hypertrophy is only superficially understood. Discovering these differences is critical to our understanding of what makes patients undergoing pathologic hypertrophy so susceptible to heart failure, whereas patients undergoing physiologic hypertrophy (for example through exercise) are not. Pathologic cardiac hypertrophy is induced by persistent pressure or volume overload as a result of hypertension or valvular heart disease. Physiologic hypertrophy, on the other hand, is induced by exercise. Both stimuli result in increases in cardiomyocyte size, however, pathologic hypertrophy is limited in its ability to maintain cardiac function, eventually resulting in heart failure. In contrast, physiologic cardiac hypertrophy maintains and improves function, as illustrated in athletes. The difference in cellular signaling between these two processes has been of great interest to researchers. A number of recent studies have lead to the hypothesis that HSF1 may regulate some of the differences in the development of physiologic and pathologic cardiac hypertrophy (recently reviewed by Toko et al.⁹¹). This is based largely on gene expression studies which have shown a differential expression of ~100 genes ^92–95. Among these differentially expressed genes are a number of HSF1-regulated genes such as HSP70 and HSP27 as well as increases in HSF1 itself. HSF1 may be one of many differences regulating the differential signaling of physiologic and pathologic cardiac hypertrophy.

Heart Failure

In order to protect cardiomyocytes from injury, heat shock proteins within the cell increase in response to externally applied stressors, including oxidative stress and inflammation. Beginning in the late 1990s, researchers have been investigating the expression of heat shock proteins in the failing human heart (Table 1). To this end, investigators have examined the expression of HSP90, HSP72, HSC70, HSP27, and HSP60 from dilated cardiomyopathy (DCM) patients, ischemic cardiomyopathy patients, and normal controls ⁶⁰. HSP72, HSC70, and HSP90 are not significantly changed between the 3 groups ⁶⁰. In contrast, DCM patients exhibit a two-fold increase in HSP27 expression in the heart compared to healthy control patients. This is in addition to a doubling of the HSP60 level, which also occurs in the hearts from ischemic heart disease patients ⁶⁰. The fact that HSP72 protein doesn’t increase in heart failure (even though it is cardioprotective ⁹⁶), while levels of the HSP60 protein are doubled ⁶⁰, suggests that there is differential regulation between HSP60 and either HSF1 or HSF-2-regulated heat shock proteins. Indeed, following induction of heart failure by placement of a permanent high left anterior descending coronary artery ligation in rats, no differences can be seen between HSF1 and HSF-2 activity as determined by EMSA ⁷. Additionally, HSP72 mRNA levels are not increased. In contrast, HSP60 mRNA increases and appears to be due to increased binding of NF-κB to both of the NF-κB binding elements in the HSP60 gene. Since HSP60 contains NF-κB binding elements, but HSP72 does not ⁷, this may explain why HSP60, but not other heat shock proteins, are increased during heart failure.

In more acute studies, where Wistar rats undergo a permanent left anterior descending coronary ligation to induce heart failure, acute increases in HSP72 and HSP27 are seen, whereas HSP60 expression remains unaffected ⁶¹. However, after 8 weeks, at which time the development of heart failure is significant, a decrease in HSP72 and HSP27 expression is observed, which appears to be somewhat contradictory when compared to the results seen in human heart failure patients (see above). Induction of heart failure in these rats also results in a parallel increase in cardiac HSP60 levels. Additional studies have determined that these increases in HSP60 correlate with a decrease in mitochondrial oxygen consumption rate as well as an increase in markers for reactive oxygen species (determined by thiobarbiturate reacting substance Toga, 2007 ⁶²). The differences in HSP regulation found in these Wistar rat studies compared to the studies in human disease may be due to the relatively acute nature of these studies compared to human studies, or possibly a result of species and even strain dependent effects.

HSP60 in heart failure

Although HSP60 is predominantly found in the mitochondria, approximately 15% of total cellular HSP60 normally resides within the cytoplasm ²⁹. However, in both DMC and ischemic heart disease hearts the distribution of HSP60 changes, with cytosolic HSP60 translocating to the mitochondria ⁶³. In other studies of heart failure using animal models and human explanted failing hearts, HSP60 has been found localized to the plasma membrane where it is detectable on the cell surface by both flow cytometry and confocal microscopy. Interestingly, localization of HSP60 to the plasma membrane of a cell correlates with an increase in apoptosis of the affected cell, possibly due to the fact that the cell surface HSP60 may be able to interact with other cells to trigger the innate immune response, resulting in the release of proinflammatory cytokines such as TNF-α. This would make HSP60 an early signal inducing myocyte loss and contributing to heart failure ⁵⁶. These few studies indicate a number of potentially conflicting data that might be due to differences resulting from species or strain variations, or from experimental design. Considerably more work is needed to delineate the regulation of HSP60 in heart failure.

HSP60’s involvement in heart failure is made even more complicated by the fact that it is also released from the cells and can be found circulating in plasma during early in heart failure ⁵⁶. Circulating HSP60 has been hypothesized to play a role in atherosclerosis by inducing inflammation and autoimmune mechanisms (see recent reviews^{97, 98}). The presence of HSP60 in the blood of normal patients was first described in 1999 ⁹⁹. Recent studies have investigated the relationship between chronic heart failure severity and serum HSP60 levels ⁶⁴. In 112 patients with CHF and 62 control subjects, serum HSP60 levels were higher in patients with CHF compared to controls ⁶⁴. CHF patients with advancing New York Heart Association functional classes had higher levels of HSP60 as well ⁶⁴. Likewise, patients with cardiac events during the average 569 days of follow up had higher serum HSP60 levels compared to event-free patients ⁶⁴. These findings demonstrate a relationship between serum HSP60 levels, the severity of CHF, and a high risk for adverse cardiac events in patients with heart failure. The role of circulating HSP60 in the underlying pathophysiology of heart failure has not been delineated.

Cardiac Ischemic Injury

Most studies investigating heat shock proteins in cardiac ischemia/reperfusion injury have reported on their cardioprotective effects. However, a few studies have concentrated on the regulation of chaperone protein expression and activity during the course of ischemia/reperfusion injury (Table 1). Reperfusion following 20 minutes of ischemia results in increases to both HSP70 and HSP90 mRNA levels, with the increase in HSP70 being much higher than that of HSP90 ( ~75 fold and HSP90 ~16 respectively) ⁶⁵. This increase in HSP70 and HSP90 expression is most likely due to a concurrent increase in the transcription factor HSF1 (but not HSF2), which in turn appears to be driven by an accumulation of reactive oxygen species during ischemia/reperfusion injury ⁷². In addition though, other studies have identified that HSF1 activation can be modulated by ATP concentrations within the cell. Moderate decreases in intracellular ATP correlate with HSF1 activation, while severe ATP depletion results in an attenuated HSF1 response, which can subsequently be rescued upon ATP restoration ⁷³. Studies investigating differential expression of genes in a pig model of ischemia/reperfusion injury reveal that HSP22 significantly increases ~ 3 fold after 1 hour of reperfusion ⁶⁷. HSP22 is also significantly increased in cases of human hibernating myocardium and pig models of hibernating myocardium ⁶⁸.

HSP70/HSP72

Both HSP70 and HSP72 have proven to be beneficial to the outcome of cardiac ischemia/reperfusion injury. Knocking-down HSP72 expression in isolated feline cardiomyocytes increases their susceptibility to cell death in response to hypoxia and reoxygenation ¹⁰⁰. In addition, increasing HSP72 in adult male rats by successive bouts of endurance exercise improves the outcomes of ischemia/reperfusion injury, illustrated by a decrease in cardiac infarct size as well as the amount of cardiac apoptosis in endurance-trained rats compared to sedentary controls⁹⁶. In the case of HSP70, adenovirus-mediated gene transfer into rabbit hearts results in a reduction in injury after ischemia/reperfusion injury ¹⁰¹. Furthermore, at least 4 studies have demonstrated that the transgenic overexpression of HSP70 in the heart of mice significantly protects against ischemia/reperfusion injury ^102–105. Since all of these studies increase HSP70 prior to the ischemia/reperfusion insult, it is not clear what the clinical utility of increased HSP70 at therapeutically plausible time points (after ischemia/reperfusion injury) would be.

Small heat shock proteins: HSP20, HSP22, HSP27, alphaB-crystallin (CryAB) (Table 1)

Increasing the HSP20 expression in isolated cardiomyocytes improves their function ¹⁰⁶ and protects against apoptosis induced by beta-agonist stimulation ¹⁰⁷. Cardiac-specific overexpression of HSP20 in mouse models (~10 fold) protects against ischemia/reperfusion injury. When HSP20 transgenic hearts are challenged with ischemia/reperfusion injury ex vivo, they exhibit an improved contractile performance, a decrease in indices of myocyte cell death, and a significant decrease in infarct size compared to wild type hearts ¹⁰⁸. This protective effect of HSP20 appears to be due to HSP20’s role in activating autophagy, a critical mechanism for dealing with ischemia/reperfusion injury ¹⁰⁹. Transgenic mice in which serine 16 on HSP20 is mutated such that it is nonphosphorylatable are more susceptible to ischemia/reperfusion injury than wild-type mice, in part, due to the inability of the mutant HSP20 to activate autophagic pathways ¹⁰⁹. HSP20 can protect against ischemia/reperfusion injury via other mechanisms also. HSP20 protects not only against oxidative stress due to ischemia/reperfusion injury, but recent studies have also found that it can protect against other injuries due to increased oxidative stress such as doxorubicin therapy ¹¹⁰. Recent studies have demonstrated that HSP20 expression is regulated, at least in part, by the mircoRNA miR-320. Down-regulation of miR-320 using an antagomir has been shown to be cardioprotective in ischemia reperfusion, in part, by its up-regulation of HSP20⁶⁶. In the case of HSP22, transgenic mice that have increased expression of HSP22 are protected against ischemia/reperfusion injury. After 45 minutes of coronary artery occlusion and reperfusion, HSP22 transgenic mice have an 82% reduction in infarct size compared to controls ⁹⁰, with HSP22 transgenic hearts exhibiting significant activation of a number of survival kinases, including Akt and AMPK to which HSP22 binds directly ⁹⁰. The cardioprotective effect of HSP22 appears to be mediated specifically through BMP signaling via activation of the PI3K/Akt pathway¹¹¹. The small heat shock protein HSP27 has also been shown to protect against ischemia/reperfusion injury using dog cardiomyocytes, with just minimal (2–3 fold) increases in expression ¹¹².

Previously we described the protective nature of CryAB in inhibiting cardiac hypertrophy. However, this small heat shock protein is also cardioprotective against ischemia/reperfusion injury when its expression is increased prior to insult. Transgenic mice in which CryAB is overexpressed, suffer less cardiac oxidative stress, decreased extent of infarction, and attenuated apoptosis and necrosis when challenged with ischemia/reperfusion injury ¹¹³. Likewise, mice lacking CryAB and HSP27, both of which are highly expressed in the heart, subjected to ischemia/reperfusion challenge exhibit a nearly 2 fold decrease in contractility recovery, with parallel increases in necrosis and apoptosis measures compared to controls ¹¹⁴. These studies indicate that, while CryAB and HSP27) are not necessary for cardiac development (CryAB/HSP27 mice develop normally and have no discernable differences in heart structure from wild type), they do play a key role in anti-oxidative mechanisms during ischemia/reperfusion injury ¹¹⁴.

Heat shock factor proteins (HSF1)

As the heart senses stress, it induces heat shock proteins by a number of mechanisms. Studies have identified that many of the heat shock proteins are regulated by the HSF family of transcription factors (Table 1). In the context of cardiac ischemia/reperfusion, HSF1 expression can up-regulate heat shock protein expression to protect against subsequent ischemia/reperfusion injury. Mice with cardiac overexpression of HSF1 challenged with ischemia/reperfusion injury recover faster, have smaller infarct sizes and decreased cardiomyocyte cell death compared to wild-type mice ¹¹⁵. In addition, Akt is enhanced while Jun N-terminal kinase and caspase 3 (apoptotic mediators) are less activated than wild type mice ¹¹⁵.

The cardioprotective role of HSF1 has been studied by using experimental models known to induce HSF1. Specifcially, cardiac HSF1 has been induced by whole body hyperthermia (WBH) or by transgenic over-expression of CaMKII-delta B. Upon HSF1 induction, these models were then challenged with ischemia/reperfusion injury. Preconditioning mice with WBH for 48 hours, then subjecting the isolated hearts to 20 minutes of normothermic ischemia and 30 minutes of reperfusion results in an increase in both HSF1 mRNA and protein and overall cardioprotection ¹¹⁶ The increase in HSF1expression is directly related to the cardioprotection as this effect is abolished with siRNA HSF1 ¹¹⁶. Inhibiting HSF1 with siRNA in the face of WBH results in an inhibition of HSP32, HSP47, and HSP60 and increased thermal intolerance, resulting in a higher mortality rate ¹¹⁶. The Ca2+/calmodulin-dependent kinase (CaMK)II is also instrumental in protecting the heart against ischemia/reperfusion injury. CaMKII is a multi-functional kinase that regulates Ca2+ handling and regulates cell death in response to ischemia/reperfusion injury. Increasing CaMKII-deltaB expression protects against oxidative stress, hypoxia and angiotensin II-induced apoptosis ¹¹⁷. Recent studies have determined that this cardioprotection is due, in part, to increasing inducible HSP70 through phosphorylation of HSF1 ¹¹⁷. These studies suggest that HSF1 may be a common mechanism by which cardiomyocytes induce a number of heat shock proteins to protect against cardiac/ischemia reperfusion injury.

At least 13 chaperones and co-chaperones regulated by at leat 2 HSF transcription factors have been described identified in the heart (Table 1). Quite predictably, of the proteins identified, their expression increases in cardiac disease and is generally cardioprotective. While most studies have focused on the regulation of these chaperones, co-chaperones, and transcriptions factors in heart failure, a growing number of studies have demonstrated more broadly their regulation in cardiac hypertrophy and ischemia reperfusion injury (Table 1). This cardioprotection includes a host of mechanisms that regulate growth and inhibit apoptosis through a variety of systems including the PI3K/Akt pathway, AMPK, PKC epsilon, nitric oxide, and mTOR. Pharmacologic enhancement of these endogenous cardioprotective mechanisms may prove to be simple yet effective strategies for reducing the morbidity and mortality associated wth common cardiac diseases.

Co-chaperones in the heart: Chaperone assistants and Protein Triage

Co-chaperones have many functions in the heart including assisting chaperones with protein folding and/or assisting with other functions, including targeting damaged proteins for degradation by the ubiquitin proteasome system in a process called protein triage ¹¹⁸. Increased co-chaperone expression in the heart has been found to be cardioprotective in ischemia and necessary to regulate proteins involved in Long-QT syndrome. A number of co-chaperones have been identified that control the activity of chaperones, including DnaJ, BAGs, Hop, CHIP, Immunophilins, and Prefoldin (Figure 1). Of these 6 general types of co-chaperones, only DnaJ, BAG-1, Hop and CHIP have been described in the heart, and our understanding of their role is preliminary (Table 1).

Using a pig model of ischemia/reperfusion to identify genes participating in mechanisms of cell survival, the Dna J-like co-chaperone (pDJA1) was identified by microarray using subtractive hybridization ⁸². pDJA1 is restricted to cardiomyocytes and is not present in skeletal muscle, liver, lung, kidney, aorta, stomach, or spleen ⁸². pDJA1 increases somewhat during ischemia, but increases 4 fold following ischemia and is protective against staurosporine-induced apoptosis in isolated rat cardiomyocytes ⁸². Since the identification of pDJA1 in 2003, little more has been reported on it despite its potential role in limiting damage in the post-ischemic myocardium.

Studies on the role of BAG-1 in cardiac ischemia reperfusion injury have demonstrated the ability of BAG-1 to inhibit apoptosis and induce autophagy in order to protect cardiomyocytes. BAG-1 interacts with HSC70 and HSP70 and promotes cell survival by coordinating the function of these chaperones with the degradation of proteins by the proteasome. Both BAG-1 isoforms (BAG-1S and BAG-1L) are rapidly induced after ischemia challenge in rat cardiomyocytes, with the increase in BAG-1 being sustained after subsequent reperfusion ⁸³. The interaction of BAG-1 with HSC70 increases after ischemia/reperfusion injury ⁸³, and increasing BAG-1S and BAG-1L in cardiomyocytes reduces apoptosis after ischemia/reperfusion injury. When BAG-1S or BAG-1L are fused to a nuclear localization sequence to force their nuclear localization, they fail to protect cardiomyoctyes, similar to BAG-1 deletion mutants that are unable to bind HSC70/HSP70 ⁸³. BAG-1 deletion constructs missing the N-terminal ubiquitin-like domain, however, do not affect the proteins ability to protect against ischemia/reperfusion injury ⁸³. These studies demonstrate a novel cardioprotective role for BAG-1, with a critical component related to its interaction with HSC70/HSP70 and cytoplasmic localization. In addition, subsequent studies have identified that autophagy plays an important role in the adaptation to ischemia-reperfusion injury in association with BAG-1 ⁸⁴. BAG-1 associates with the autophagosomal membrane protein LC3-II and may induce autophagy using HSC70 ^{84, 119}. Intracardial injection of BAG-1 siRNA attenuates the induction of LC3-II and abolishes the cardioprotection achieved by adaption ¹¹⁹. The BAG-3 isoform participates in the induction of macroautophagy in association with HSP22 ⁸⁴, demonstrating how BAG family members may shuttle damaged or oxidized proteins into the autophagy pathway to improve cell survival⁸⁴.

The co-chaperone CHIP (carboxy terminus of HSP70 interacting protein) is one protein that plays a key role in both the folding system (as a co-chaperone regulating HSP70) and in the UPS as a ubiquitin ligase. CHIP directs the degradation of aggregate prone proteins^{120, 121}, such as poly-glutamine proteins, which are prevalent in conformation diseases such as Alzheimer or Huntington’s disease^122–124. Although CHIP binds HSP70, it can also target it for degradation in the absence of cargo, possibly as a feedback mechanism to adjust chaperone levels needed for the number of misfolded proteins (Figure 1B) ⁵⁰. Recent studies have identified BAG-2, a specific inhibitor of CHIP-dependent ubiquitin ligase activity, as a common component of CHIP holocomplexes in vivo ¹²⁵. CHIP plays an important role in the heart in response to ischemia/reperfusion injury. When CHIP −/− mice are challenged with ischemia/reperfusion injury in vivo, the ratio of the infarct area to the area of risk is 50% greater than that found in sibling wild type mice ⁸⁵. CHIP −/− hearts are more prone to cell death indicating a critical role of CHIP in ischemia/reperfusion injury. These studies parallel the role of BAG proteins described above, indicating a critical role of CHIP in shuttling damaged and oxidized proteins into autophagic pathways after ischemia/reperfusion injury. The specific role of CHIP in autophagy has yet to be reported. Cardiomyocyte CHIP increases in response to high glucose and is responsible for the degradation of the pro-hypertrophic transcription factor GATA4 ⁸¹. The significance of these findings in cardiac disease has yet to be reported.

The co-chaperone FKBP38 is an immunophilin-type small heat shock protein that has recently been implicated in the maturation of HERG (human Ether-à-go-go Related Gene, also known as KCNH2 in newer nomenclature)¹²⁶. The HERG gene encodes the voltage-dependent delayed rectifier potassium channel (I_KR) and mutations in HERG are among the most common underlying cause of hereditary Long QT syndrome. Recent studies have utilized proteomic screens to identify that HSC70, HSP90, HDJ2 (HSP-organizing protein) and BAG-2 are differentially expressed in models of Long QT syndrome caused by mutations in HERG ¹²⁶. However, the most relevant findings of these studies are that the co-chaperone FKBP38 immunoprecipitates and co-localizes with HERG ¹²⁶. Additionally, siRNA knock-down of FKBP38 causes a reduction in HERG trafficking and overexpression of wild type FKBP38 partially rescues HERG trafficking in the presence of F805C disease-causing KCNH2 mutation¹²⁶. These studies suggest an important role for the co-chaperone FKBP38 in rescuing mutations in KCNH2 that lead to the Long QT syndrome.

A picture of the heat shock protein system as a mediator of protein quality control is emerging. Specific co-chaperones such as CHIP have the ability to ubiquitinate proteins that the chaperone/co-chaperone complex is unable to refold (Figure 1B). The ubiquitination of key structural proteins, such as sarcomere proteins and transcription factors are critical to the long term health of the heart (see recent reviews^{43, 51, 127, 128}). These co-chaperones represent a system of triage whereby protein quality is maintained, and in the long run, the health of the cardiomyocyte is maintained.

Future directions: The role of drug therapies in cardiac health and pathophysiology

Drugs that induce heat shock proteins

A number of studies reviewed here have tested the hypothesis that increasing heat shock proteins improves the outcome of cardiac diseases experimentally, particularly in the context of ischemia/reperfusion injury. While these studies were primarily proof of concept that increasing heat shock proteins were cardioprotective, it is interesting to note that the fold increase in these proteins was as little as 2. From a clinical standpoint, there are several drugs and herbal products that increase heat shock proteins that may be beneficial in the treatment of cardiac diseases. However, their clinical utility has yet to be tested experimentally in the context of their regulation of heat shock proteins.

Geranylgeranylacetone (GGA)

GGA is a cyclic polyisoprenoid gastric ulcer drug that protects the gastric mucosa by inducing HSF1 and HSP70 mRNA ¹²⁹. It has recently been shown experimentally to be cardioprotective by inducing HSP72 ^{130, 131}. It has also been shown experimentally to suppress poly-glutamine toxicity (see recent review ¹³²).

Arimoclomol

Arimoclomol, developed by CytRx, is a small molecule that acts by inducing HSF1 resulting in downstream increases of HSP70 and HSP90 ¹³³. Experimentally, arimoclomol increases HSP70 and HSP90 approximately 5 fold in an experimental model of ALS ¹³⁴. Arimoclomol is currently in phase II/II clinical trials as a treatment for ALS.

Celastrol

Celastrol is a triterpenoid compound with a retinoid skeleton extracted from Tripterygium wilfordii that is used in traditional Chinese medicine. It potently induces HSF1 and HSP70 expression, having both anti-oxidant and anti-inflammatory activities ¹³⁵. Celastrol has been shown to ameliorate the neurodegeneration of SOD1 mutant mice; however, it has not been determined if this is through its anti-inflammatory effects and/or its affect on HSF1 and HSP70 induction ¹³⁶.

Statins

In addition to their ability to decrease cholesterol synthesis via inhibition of HMG-CoA, statins have been shown to have many additional activities, including modulation of the immune system, reduction in apoptosis, and an affect on nitric oxide production ^137–139. Both simvastatin and lovastatin induce HSP27, but not HSP70 and HSP90 in an osteoblast-like cell line ¹⁴⁰. Simvastatin induces HSF1 in vascular endothelial cells, to induce nuclear translocation and the transcription of HSP70 and HSP90 ¹⁴¹. Simvastatin induces HSP27 in axotomized retinal ganglion cells to enhance their survival after optic nerve transaction ¹⁴² Statins increase HSF1 and HSF-2 in retinal ganglion cells in vivo ¹⁴³. Statins act to increase HSFs, HSP70, HSP90, and sHSPs, possibly in a cell dependent manner. Their effect on cardiac heat shock proteins has not been identified to date. With the discovery of statin’s ability to induce heat shock proteins, studies have identified a decreased prevalence of Alzheimer’s disease in patients taking statins ^{144, 145} and a decrease in neurofibrulary tangles ¹⁴⁵. These studies suggest that a several drugs, including the widely prescribed statins, have the potential to be cardioprotective due to their ability to prime critical heat shock proteins in the heart. The use of most of these drugs has yet to be determined in human studies.

Conclusion

During the development of cardiac hypertrophy, heart failure, and ischemia/reperfusion injury, there is a general increase in a number of chaperones, co-chaperones, and the transcription factors that regulate them. In this review, we discuss the their uniformly protective mechanisms and the possibility that therapeutic regulation may enhance both acute and chronic health of the heart. With the recent discovery that increases in soluble pre-amyloid oligomers play a significant role in cardiac disease and are able to induce cardiomyopathy experimentally⁵, there is a need for a rational way to increase chaperone/co-chaperone function to combat the accumulation of misfolded proteins. A number of drugs with the potential to increase heat shock proteins are being developed for neurodegenerative diseases caused by misfolded proteins (i.e. poly-glutamine diseases such as Huntington’s disease). Given the parallel mechanisms found in cardiac diseases and the overwhelming evidence that increasing chaperones/co-chaperones is cardioprotective against the most common cardiac diseases, there may be future clinical applicability in cardiology of these drugs.

Non-standard abbreviations

BAG

Bcl2-associated athanogene

CHIP

carboxy terminus of HSP70-binding protein

CryAB

alphaB-crystallin

ERAD

endoplasmic reticulum-associated degradation

HDAC

histone de-acetylase

HSC

cognate of heat shock protein

HSF

heat shock factor

HSP

heat shock protein

PAO

preamyloid oligomers

PolyQ

poly-glutamine

TRiC

(TCP1-ring complex, or chaperonin containing TCP1 (CCT)

UPS

Ubiquitin proteasome system