Keywords

1 Introduction

Stroke refers to a range of conditions caused by interruption of the blood supply to the brain, either by occlusion of the cerebral vessels, or by rupture leading to hemorrhage (Kriz and Lalancette-Hebert 2009). Following stroke, the brain undergoes a coordinated stress response which seems to protect it from injury. This cellular response includes the induction of a variety of stress proteins among which, one of the most widely studied are the heat shock proteins (HSP). HSP were originally described because they were highly induced in cells exposed to sublethal heat stress. HSP are a highly conserved family of stress proteins believed to play a role in preventing the aggregation of harmful protein as well as facilitate the synthesis of new proteins via various chaperone function mechanisms such as translocation, degradation, folding, and complex protein assembly. (Giffard and Yenari 2004). Universally expressed HSP family members appear within all subcellular compartments, and play an essential role in normal development and cellular function. After a variety of insults including brain ischemia, inducible forms increase, and appear to be part of a larger orchestrated stress response (Giffard et al. 2008). Over the past few decades, work has shown that some HSP also act as a cytoprotectants. They have also been shown to affect cellular signaling, and have been extensively studied to provide protection against different experimental brain injury models.

2 Heat Shock Proteins

At stroke onset, synthesis of most cellular proteins is reduced; however, different proteins seem to be upregulated, and it is becoming increasingly clear that many of these proteins serve to protect the cell from injury. HSP were among the first of these proteins to be identified and studied. Mammalian HSP have been divided into several families in accordance with molecular weight. These include HSP110, HSP90, HSP70/HSP80, HSP60, and small molecular HSP (smHSP). Constitutive HSP, such as HSP90, HSP40 and cognate HSP70 (HSC70), function together within cells (Lianos et al. 2015). In brain cells, heat stress acts a catalyst to increase inducible HSPs, including HSP27, HSP32 and HSP70 (Lianos et al. 2015). In brain cells, heat stress triggers a robust expression of inducible HSP, such as HSP70, HSP32 and HSP27 (Brown 2007). HSP70, or the 70-kDa class which includes an inducible form also known as HSP72, HSP70i, or simply HSP70 is probably the most studied of this class of protein. It binds to hydrophobic regions of substrate polypeptides in an ATP-dependent manner. ATP hydrolysis then induces a conformational change in HSP70 and triggers substrate folding followed by release (Mayer and Bukau 2005). HSP70 has an N-terminal ATPase domain and a C-terminal substrate-binding domain that helps facilitate protein folding by switching states between open and closed ATP-binding with low and high substrate affinity respectively (Giffard and Yenari 2004). In studies of ischemic stroke, HSP70 was first documented to be induced in brain regions that were relatively resistant to ischemic insults. Hence, the notion of a ‘molecular penumbra’ was introduced, and raised questions as to whether this expression was an epiphenomenon of injury, or an active participant in cell survival (Sharp et al. 2000). A growing body of research has demonstrated repeatedly that HSP70 acts to preserve the brain following experimental insults such as ischemic stroke, neurodegenerative disease, epilepsy, and trauma. HSP70 appears to decreases protein aggregates and intracellular inclusions via chaperone functions. (Giffard et al. 2008). The stress proteins HSP27 and HSP32, or heme oxygenase 1 (HO-1), have also been studied in brain ischemia (Sharp et al. 2013). HSPs, in addition to their role in protein processing, have been shown to preserve the brain via multiple immune response and cell death mechanisms (Giffard and Yenari 2004; Kelly and Yenari 2002).

3 Mechanism of HSP Induction

HSPs interact with various signaling cascades and are responsible for cell growth and differentiation under normal circumstances. Cell stress, including brain injury, causes HSPs to be quickly induced. Insults such as heat, ischemia and other causes of accumulation of unfolded proteins are thought to serve as stimuli for stress protein induction. These unfolded proteins activate heat shock factor (HSF) within the cytosol by dissociating other HSP that are normally bound to HSF (Kim et al. 2012). Once liberated, HSFs are phosphorylated and form trimers. The trimers then enter the nucleus and bind to heat shock elements within the promoters of different heat shock genes, leading to more HSP70 generation (Kim et al. 2018) (Fig. 6.1).

Fig. 6.1
figure 1

Induction HSP mechanism. Under non-stressful conditions, heat-shock factor (HSF) monomers are associated with a chaperone complex that consists at least of HSP70, HSP90 and HSP40. After ischemic stroke, dissociation of the complex thus phspholytaion HSF monomers able to move into the nucleus and bind to the heat-shock element, leading to a decrease in HSP70 gene transcription

Full size image

New HSP70 protein, in conjunction with ATP, HSP40 and HSP90, acts to bind denatured proteins and serves as a chaperone at the molecular level by assisting with the repair, refolding and trafficking of damaged proteins in the cell. Inside of the cell, this chaperone complex undergoes multiple iterations to attempt to refold the proteins with binding of the Hip protein to the N-terminus and the Hop protein to the C-terminus of HSP70. HSP70 ultimately helps fold nascent proteins and refold denatured proteins (Lanneau et al. 2010). When refolding does not occur, Bag-1 protein binds to the N-terminus of HSP70, and the E3-ubiquitin ligase CHIP (C-terminus of HSP70/HSC70 interacting protein) binds to the C terminus of HSP70. This complex then interacts with the denatured protein and recruits it to the proteasome where it is ubiquitnated and degraded (Demand et al. 2001). Thus, HSP70 assists in refolding or degradation of damaged proteins.

4 HSP in Ischemic Stroke

Laboratory studies have most commonly used rodent models of middle cerebral artery occlusion (MCAO) and global cerebral ischemia to model the clinical conditions of ischemic stroke and global ischemic brain injury following cardiac arrest, respectively. In both experimental brain ischemia models, induction of HSP has been shown to protect against a variety of acute insults (Sharp et al. 2000). During homeostatic conditions, inducible HSP are present at very low levels; however, its expression is markedly increased following injury. Several studies have reported the relationship between HSP induction and the development of tolerance in ischemic stroke. Direct overexpression of these HSP have been shown to lead to neuroprotection following brain ischemia (Kelly and Yenari 2002).

HSP70 is the most abundant HSP found in cells. After 10 min of focal cerebral ischemia, HSP70 can be detected within the ischemic territory 24 h later. After 1.5 h of middle cerebral artery (MCA) occlusion, much of the MCA territory has undergone infarction, but HSP70 has been detected within the watershed zone between the middle and anterior cerebral arteries. Within these peri-infarct or ‘penumbra’ zones, HSP70 induction occurs primarily in neurons (Sharp et al. 2000). Within zones of infarction and regions adjacent, HSP70 can be observed in glial and endothelial cells (Sharp et al. 2000). Similar observations have been observed in global cerebral ischemia models where HSP70 showed the highest induction within neuronal populations most vulnerable to ischemic injury (Chen and Simon 1997).

Viral vector-mediated HSP70 overexpression has been shown to improve survival of neurons and astrocytes from ischemic and ischemia-like insults, including oxygen glucose deprivation and focal and global cerebral ischemic injury (Giffard and Yenari 2004). Transgenic mice have shown to be protected from ischemic insults when they overexpress HSP70. Similarly, a deficiency in HSP70 in transgenic mice, has shown negative effects. (Kim et al. 2016; Lee et al. 2001). Experimentation utilizing a stroke model demonstrated that intravenous TAT-HSP70, a HSP70 tethered to a TAT motif to enhance BBB penetration, decreased infarct volumes, enhanced neurological outcomes, and increased the survival rate of neural progenitors (Doeppner et al. 2009). The effects of this neuroprotection were associated with fewer apoptotic cells, and elevated levels of anti-apoptotic proteins.

HSP27 has many attributes similar to HSP70, except that ATP is not needed for HSP27’s actions (Stetler et al. 2009). Transgenic mice overexpressing HSP27 subjected to cerebral ischemia demonstrated neuroprotective benefits due to this overexpression (Latchman 2004). protection against ischemic brain and kainate induced neuronal cell death was demonstrated by viral vector-mediated HSP27 overexpression both in vitro and in vivo. (Sharp et al. 2013). HSP27’s capacity to impede apoptosis has been attributed to this protective nature. Research shows that HSP70 and HSP27 prevent mitochondrial dispersion of cytochrome c and the formation of the apoptosome. It is possible that, HSP27 directly interacts with pro-caspase-3 and may work to prohibit Bax translocation to the cell’s mitochondria. HSP27’s neuroprotective properties against cell death have been demonstrated in hippocampal pyramidal neurons against ischemic stroke when the PEP-1-HSP27 fusion protein, a construct to penetrate the BBB, was administered intraperitoneally (An et al. 2008).

HO-1(heme oxygenase-1, HSP32) is an inducible enzyme involved in heme catabolism. Because it contains a HSF in its promoter, it is included as a member of the stress protein family. It acts to transform heme into biliverdin, carbon dioxide and ferrous iron. It shares induction factors with many HSPs (Kelly and Yenari 2002); yet, studies of HO-1 relating to brain ischemia and similar conditions is inconsistent. HO-1 knockout mice have been shown to have worse outcomes in ischemic stroke models, but better outcomes in brain hemorrhage models (Wang and Dore 2007). Circumstances surrounding the location where HO-1 is active and the differential effects of its metabolites may both be possible reasons for these conflicting results.

5 HSP in Neuronal Cell Death Pathways

HSP70 directly or indirectly interferes with cell death pathways such as apoptosis. HSP70 affects several factors of the programmed cell death machinery, both upstream (Stankiewicz et al. 2005; Steel et al. 2004) and downstream of mitochondrial events (Ravagnan et al. 2001). HSP70 has been shown to interrupt cytochrome c release in ischemic stroke models (Lee et al. 2004b; Tsuchiya et al. 2003) and inhibit nuclear translocation of apoptosis inducing factor (AIF) (Matsumori et al. 2005) while decreasing brain infarct size. Overexpression of HSP70 in transgenic mice impedes the translocation of procaspase-9 into the apoptosome, and sequester AIF (Beere et al. 2000). HSP70 also inhibited release of the proapoptotic protein Smac/DIABLO from myocyte mitochondria (Jiang et al. 2005). Mitochondrial HSP70, also known as HSP75 or mortalin assistants act to maintain mitochondrial membrane potential, may and help preserve mitochondrial function and mitochondrial protein import (Geissler et al. 2000) (Voloboueva et al. 2008). HSP27 disrupts the formation of the apoptosome and interacts with cytochrome c, prohibiting cytochrome c release from mitochondria, resulting in a reduction of procaspase-9 activation (Stetler et al. 2009). HSP27 can directly interact with procaspase-3 and inhibit caspase-3 activation (Stetler et al. 2009). When HSP70 is induced in astrocytes, it reduced the cell’s vulnerability to in vitro ischemia-like injury (oxygen glucose deprivation, OGD) and preserved higher ATP levels in stressed cells (Voloboueva et al. 2008). These outcomes were related to reduced reactive oxygen species (ROS) formation and better maintained mitochondrial membrane potential in in vitro model of ischemic stroke (Ouyang et al. 2006) and with better preservation of glutathione levels (Xu and Giffard 1997).

Bcl-2 plays major role in preventing apoptosis. It inhibits the release of cytochrome c and AIF to which are essential to caspase activation. HSP70 overexpression can decrease induction of apoptosis upstream of mitochondria in ischemic stroke models. HSP70 overexpression by viral vectors led to improved survival amongst targeted hippocampal neurons, and this protective effect was associated with increased levels of Bcl-2 protein (Kelly and Yenari 2002). Whether Bcl-2 family members inhibit or promote apoptosis depends on the balance between pro- and anti-apoptotic members which can regulate the mitochondrial membrane permeability transition (Yuan and Yankner 2000); thus, HSP70, through increasing anti-apoptotic Bcl-2 protein expression, may inhibit apoptosis by altering this balance so as to favor cell survival. HSP70 has also been shown to reduce heat-induced apoptosis primarily by blocking mitochondrial translocation of the pro-apoptotic Bcl-2 family member Bax, thereby preventing the release of pro-apoptotic factors (Stankiewicz et al. 2005). Previous studies of HSP27 and its anti-apoptotic activity established that HSP27 can indirectly suppress stress-induced Bax oligomerization and translocation to the mitochondria (Havasi et al. 2008). HSP27 has also been shown to phosphorylate the survival kinase Akt/PKB (Rane et al. 2003) or to inactivate pro-death c-Jun N-terminal kinase (JNK) (Schepers et al. 2005). HSP70 also reduces the activity of apoptosis protease activating factor-1 (Apaf-1), which forms the apoptosome and leads to subsequent activation of caspase-9 (Beere et al. 2000), although others have reported that it does not directly interact with Apaf-1 (Steel et al. 2004).

HSP70 also has the potential to interrupt extrinsic, or receptor-mediated apoptosis. Among various death receptors, Fas is perhaps the most studied. Fas, when bound by its ligand FasL, initiates apoptosis by recruiting caspase-8, which then leads to a cascade that ultimately culminates in caspase-3 activation followed by apoptotic cell death. Through protein profiling, our group discovered that in ischemic brains of HSP70 overexpressing mice exposed to experimental stroke, dynamin was one of the most downregulated proteins. Dynamin is a GTPase typically associated with endocytosis. However, Ivanov and colleagues reported that dynamin also trafficks Fas from the endoplasmic reticulum to the cell’s surface (Ivanov et al. 2006) where it becomes accessible to FasL. In experimental stroke, HSP70 seems to prevent Fas trafficking to the cell surface through interactions with dynamin (Kim et al. 2016). Thus, HSP70 also prevented the extrinsic or receptor-mediated apoptotic pathway through specific chaperone interactions.

6 Role of HSP in Neuroinflammation

HSP are also known to modulate inflammatory responses through both pro- and anti-inflammatory mechanisms. Following acute stroke, immune responses are thought to be mediated through innate pathways, of which HSP have been shown to participate (Srivastava 2002). In the extracellular environment, HSP have been well studied in terms of their role in both innate and adaptive immunity where they appear to assist in and potentiate these responses. HSP70, perhaps the most studied of the HSP with respect to its role in inflammation, appears to play dual roles depending on the nature of the stimulus and the ensuing immune response. As an innate immune modulator, HSP70 can interact with macrophages, microglia, and dendritic cells through Toll-like receptors (TLRs) leading to activation of nuclear factor-kappaB (NF-kB), which induces pro-inflammatory molecules such as cytokines and inducible nitric oxide synthase (iNOS) (Giffard et al. 2008; Srivastava 2002). HSP60 and HSP70 are both thought to interact with TLR 2 and TLR4 (Asea 2008); however, some of this work has been questioned, since some preparations of recombinant HSP may contain low levels of endotoxin, which is the classic ligand for TLR4 (Gaston 2002). With regard to the adaptive immune system, extracellular HSP70 complexed with peptides elicit CD8+ T-cell responses after exogenous administration. Immunization of mice with these same complexes can elicit CD4 responses, indicating that HSP can act as an adjuvant. These HSP70-peptide complexes can also interact with the macrophage/dendritic cell CD 40, CD91, or lectin-like oxidized low density lipoprotein receptor-1 (LOX-1) and aid in antigen presentation.

HSP70 has also been reported to have anti-inflammatory effects. It has been shown to decrease the release of pro-inflammatory factors such as matrix metalloproteinases (MMPs), reactive oxygen species (ROS) and inhibit NF-kB activation. Intracellular overexpression of HSP70 or its intracellular induction through heat stress has been shown to decrease inflammatory cell production of nitric oxide (NO) and iNOS expression while blocking NF-kB activation in glial cells of the brain (Feinstein et al. 1996). Heat stress has also been correlated to the reduction of tumor necrosis factor-alpha (TNF-α) and ROS generation. Heat shock-induced HSP70 also correlated to reduced expression inflammatory cytokines such TNF-α and interleukin-6 (IL-6) (Van Molle et al. 2002), while induction of HSP70 in macrophages blocked LPS-induced TNF, IL-1, IL-10 and IL-12 expression (Ding et al. 2001). In an experimental model of intracerebral hemorrhage, overexpression of HSP70 decreased TNF-α expression and interrupted blood brain barrier (BBB) disruption, edema formation, and neurological dysfunction (Manaenko et al. 2010).

HSP70 overexpression by heat stress also blocks NADPH oxidase activity in neutrophils and promotes superoxide dismutase (SOD) in phagocytes (Polla et al. 1995). In heat-pretreated astrocytes, HSP70 overexpression also interrupts the phosphorylation of IkB, JNK and p38 and blunts DNA binding of transcription factors, such as NF-kB, activator portien-1 (AP-1) and signal transducer and activator of transcription factor 1 (STAT-1), effectively downregulating the expression of pro-inflammatory genes (Kim et al. 2015) (Fig 6.2). In another study, prior-heat stress lead to interruption of the inflammatory response, and this was associated with inhibition of NF-kB translocation to the nucleus (Guzhova et al. 1997; Heneka et al. 2000). HSP70 has also been shown to interrupt NF-kB dissociation by preventing the phosphorylation of the inhibitor of kB (IkB) (Feinstein et al. 1996). A few studies have shown that HSP70 can also bind to and inhibit NF-kB and/or its regulatory proteins (Ran et al. 2004; Zheng et al. 2008), although how it does this may depend on the nature of the stimulus. In a cell death model induced by TNF-α, HSP70 inhibited IkB kinase (IKK) activity directly, whereas in a model of ischemic stroke, HSP70 appeared to associate with NF-kB and IkB, thus preventing IkB phosphorylation by IKK. The inhibition of NF-kB by HSP70 was thus shown to have a neuroprotective effect in a stroke model by preventing transcription of several immune genes (Zheng et al. 2008).

Fig. 6.2
figure 2

Influence of HSP in innate immunity. Following ischemic stroke, HSP have been shown to inhibit the activation of transcription factors and their nuclear translocation, thus interrupting the activation of various pro-inflammatory factors expressed following ischemic stroke. HSP70 induction pharmacologically is also possible through inhibitors of HSP90 (Geldanamycin, 17-AAG)

Full size image

HSP70 also appears to inhibit generation of matrix metalloproteinases (MMPs), a family of extracellular proteases thought to contribute to reperfusion injury and brain hemorrhage in experimental stroke models. In an in vitro study of HSP70 overexpression astrocytes, MMP-9 expression, also under the transcriptional control by NF-kB, was inhibited following exposure to ischemia-like insults (Lee et al. 2004a). HSP70 overexpression also seemed to similarly decrease MMP-2 (Lee et al. 2004a). Interestingly, MMP-9 expression is regulated by NF-kB, whereas MMP-2 is not. However, our prior study in cultured astrocytes (Kim et al. 2015) and studies in alveolar macrophages indicate that HSP70 can inhibit STAT-1 (Howard et al. 2010), and STAT-1 has been linked to MMP-2 expression (Johnston et al. 2000). Further, HSP70 also appears to prevent MMP processing from its pro, or inactive form to its cleaved or active form. Thus, it is clear that HSP have a myriad of roles, some of which modulate immune responses toward both pro-and anti-inflammatory phenotypes, although the effects in acute stroke models suggests that the overall response is anti-inflammatory.

7 Clinical Translation of HSP

After ischemic stroke, strategies to increase intracellular HSP70 might be significant in many neurological conditions related to cell death or pro-inflammatory processes. While heat stress may be impractical to translate, pharmacological induction of HSP70 is not impossible. A few studies have now shown that pharmacological induction of HSP70 can protect the brain in experimental stroke and brain injury models. The best studied HSP70 inducers are the ansamycins, geldanamycin (GA) and 17-(Allylamino) geldanamycin (17-AAG). These compounds induce HSP70 through their ability to inhibit HSP90, and have been shown to protect the brain from brain injury (Kim et al. 2012) (Fig 6.2). Additionally, GA and 17-AAG have already been studied in clinical trials, albeit for other indications. Clinical trials of GA were halted due to liver toxicity (Supko et al. 1995), but the less toxic GA analogue 17-AAG with similar HSP90-antagonistic functions have been studied in phase 3 clinical trials as cancer therapy (Porter et al. 2010). With its solubility in aqueous solution, 17-(2-dimethylaminoethyl) amino-17-demethoxygeldanamycin(17-DMAG), was developed for cancer treatment, but was also found to improve outcome following experimental stroke, while decreasing activation of microglia and NF-kB through inhibition of IκB phosphorylation (Qi et al. 2014).

Also, other HSP70-inducing pharmaceuticals include the purine-based compounds, the resorcinols, and other novel chemotypes. The purine series is a synthetic class that was modeled after the way ansamycins co-opt ADP to bind the HSP90 ATP binding site (Jones et al. 2011). BIIB021 (also named CNF-2024) is the best-known purine based HSP70-inducer so far, and is a potent HSP90-inhibitor (Porter et al. 2010). It can be given orally, and has been studied up to the phase II level in cancer studies with an acceptable safety profie (Jhaveri et al. 2012). The resorcinols are compounds based on radicicol, but to date, there have been no studies evaluating resorcinols’ effect on brain injury (Porter et al. 2010), although considering related studies to inhibit HSP90, there would be good reason to study this class of compounds in stroke models. Since radicicol itself tends to degrade in vivo, there are some variants designed to overcome this pharmaceutical limit, such as NVP-AUY922 and AT-13387. There is also growing interest in another HSP70 inducer, geranylgeranylacetone (GGA) known for its antiulcer properties. Administration of GGA after middle cerebral infarction showed that it anti-inflammatory effect was due to up-regulation of HSP70 through protein kinase C induction (Yasuda et al. 2005). Zhao et al. found that pretreatment with GGA led to neuroprotection by reducing neuronal cell apoptosis and microglial activation in a traumatic brain injury model (Zhao et al. 2013). No clinical studies of HSP70 inducers/HSP90 inhibitors have been carried out for neuroprotection against acute neurological insults, but considering robust laboratory data, ample clinical experience in cancer patients, and renewed interest in neuroprotection considering a prominent role for acute revascularization in stroke patients (Mizuma et al. 2018), this may be area ripe for investigation.

8 Conclusions

Several studies have investigated the beneficial effects of HSPs after ischemic stroke. Several laboratories have shown that HSPs lead to neuroprotective effects in experimental stroke and should thus be explored as a therapeutic target. One translatable approach may be through pharmacological induction of HSPs through several already available HSP90 inhibitors, some of which have already been tested in humans. There are a few laboratory studies in preclinical animal model that have shown the beneficial effects of induced HSP70 for treatment of ischemic stroke and brain trauma. Multiple protective mechanisms of HSP70 make it a particularly attractive target, as it has the potential to target several aspects of ischemic injury. Considering recent clinical advances in the treatment of acute stroke, the therapeutic value of HSP should be considered for further investigation.