**2. Links between Hsp70 and the pathogenesis of PD**

PD is currently thought to involve different pathogenic mechanisms that eventually lead to neurodegeneration, as discussed elsewhere in this book series.. There is substantial evidence supporting a prominent role in PD-related cell death of αSyn toxic oligomers. One hypothesis postulates that certain αSyn species can affect the homeostasis of cell membranes (Gupta et al., 2008; Lashuel et al., 2002b; Volles & Lansbury, 2002), and produce ER and oxidative stress, UPS and mitochondrial dysfunction (Gupta et al., 2008; Jellinger, 2010), as well as neuroinflammation (Jellinger, 2010; Roodveldt et al., 2008) processes, all of which have been linked, to a lower or larger extent, to the amyloid-like aggregation of αSyn. As will be described below, the Hsp70 system has been found to be a key player in counteracting most of these processes, not only by physically interacting with αSyn, but also by promoting aggregation clearance (Figure 1).

Fig. 1. Links between Hsp70 and the multiple αSyn-mediated processes in PD pathogenesis. CMA: chaperone-mediated autophagy; UPS: ubiquitin-proteasome system; LB: Lewy bodies; LN: Lewy neurites. Discontinuous lines depict the possible 'sequestration' of certain chaperones by αSyn aggregating species.

The Hsp70 Chaperone System in Parkinson's Disease 225

inhibition of amyloid assembly can occur in the absence of collaborating co-chaperones and in an ATP-independent manner (Ahmad, 2010; Dedmon et al., 2005; Huang et al., 2006; Luk et al., 2008; Roodveldt et al., 2009), while promoting formation of small oligomeric species of moderate toxicity (Roodveldt et al., 2009). However, our recent study demonstrates that, even though Hsp70 is indeed able to control αSyn fibrillation and its associated toxicity in an ATP-independent manner, it does so more efficiently in the presence of ATP (Roodveldt et al., 2009), consistent with the results obtained with a cellular model (Klucken et al., 2004a). Nevertheless, this ATP-dependent activity was found to require Hip co-chaperone activity, and not Hsp40 as observed in the case of mutant huntingtin protein (Lotz et al., 2010), suggesting that Hsp70 can function through different molecular mechanisms, depending on

Interestingly, our study also showed that Hsp70 in the presence of ATP is prone to coaggregate with aggregating αSyn, presumably caused by the formation of a highly insoluble (ADP)Hsp70/αSyn complex. Surprisingly, this co-aggregation can be prevented by the addition of the co-chaperone Hip (St13), which had been found to be under-expressed in serum of PD patients since the early stages of the disease (Scherzer et al., 2007). The relevance of Hip in assisting the suppression of αSyn aggregation in an Hsp70-dependent manner was further supported by a study we performed in a *C. elegans* model of PD (Roodveldt et al., 2009), in which knock-down of Hip produced a much stronger PD phenotype than knock-down of Hsp70, shown by higher levels of αSyn aggregation in the former condition. This finding indicates that Hip co-chaperone could be important not only in modulating the chaperone's molecular mechanism, but also in guarding the functionality

One relevant question relates to the nature of the species along the aggregation pathway of αSyn that are specifically recognized and targeted by Hsp70. Initially, Hsp70 was reported to bind to αSyn filaments *in vitro* (Lindersson et al., 2004). However, it is currently understood that Hsp70 does not disaggregate or alter the structural properties of mature αSyn fibrils (Dedmon et al., 2005), but rather inhibits fibril formation via interactions with soluble pre-fibrillar forms of αSyn. A few studies performed with cell extracts (Zhou et al., 2004) and with live cells (Klucken et al., 2006), in addition to experiments with purified proteins (Dedmon et al., 2005; Huang et al., 2006; Roodveldt et al., 2009), demonstrate the existence of molecular interactions between Hsp70 and αSyn oligomeric species. Even though previous attempts to isolate such complexes by co-immunoprecipitation or pulldown experiments had failed (Luk et al., 2008), we have been able to probe the formation of a complex between Hsp70 and αSyn oligomeric species by FRET (Roodveldt et al., 2009), suggesting that these interactions are transient (Luk et al., 2008; Roodveldt et al., 2009),

In addition, we have recently shown by FRET and nuclear magnetic resonance (NMR) analyses that, contrary to what was previously thought, Hsp70 can also interact with αSyn monomeric species (Roodveldt et al., 2009) with an affinity constant lying within the low micromolar range. Moreover, the fluorimetric study revealed the existence of diverse Hsp70/αSyn complexes that are formed depending on the nucleotide state of the chaperone, either nucleotide-free, or ATP- or ADP-bound, that are sampled along the ATPase cycle. Based on these structural studies, we have proposed that a particularly 'compact' Hsp70/αSyn complex in the ADP-state of Hsp70 that arises during the aggregation process of αSyn, leads to Hsp70 being entrapped or sequestered by the oligomers. However, we

the nature of the aggregating client protein.

and availability of Hsp70 under certain conditions.

besides being highly dynamic (Luk et al., 2008; Roodveldt et al., 2009).

#### **2.1 Hsp70 in modulation of** α**Syn aggregation and cytotoxicity**

Heat-shock proteins (HSPs) prevent and reverse the misfolding and aggregation of proteins, and the Hsp70 family in particular is known to play important roles in protecting neurons from protein aggregation-derived stress (Lu et al., 2010). Therefore, it might not seem surprising that Hsp70 has been found to be linked to several neurodegenerative processes and 'conformational' disorders, including PD (Witt, 2009). Notably, Hsp70 has been shown to colocalize with aggregated αSyn within LBs in brains from PD patients (Auluck et al., 2002), strongly suggesting a role for this chaperone in managing αSyn aggregates in the context of PD (Figure 1). It is then that a substantial portion of the research in the field has focused on the effects of this major cytosolic chaperone on αSyn aggregation and cytotoxicity. Following the discovery that Hsp70 can abrogate the neurotoxicity of abnormal polyglutamine proteins (Warrick et al., 1999), it was shown that Hsp70 can also prevent dopaminergic neuronal loss associated with αSyn in a *Drosophila* PD model (Auluck et al., 2002). Numerous studies that followed have reported that over-expression of Hsp70 is able to reduce αSyn aggregation and/or toxicity in various cellular models (Danzer et al., 2011; Klucken et al., 2004b; McLean et al., 2004; Opazo et al., 2008; Outeiro et al., 2008; Zhou et al., 2004). In particular, McLean and co-workers (Outeiro et al., 2008) have found that Hsp70 rescues αSyn-linked toxicity by promoting the cellular clearance of αSyn oligomers, rather than monomers. Another study (Opazo et al., 2008) found that Hsp70 manages αSyn intracellular aggregation by increasing the clearing of aggregates primarily through the aggresome, and the subsequent removal of small aggregates and aggresomes from the cytosol. An interesting study by McLean and colleagues (Danzer et al., 2011) has recently shown that Hsp70 can also inhibit the formation of extracellular αSyn oligomers and rescue the cytotoxicity produced by such secreted oligomers, in a cellular model. Moreover, their data also indicates that Hsp70 is released to the extracellular medium together with secreted αSyn, adding to the accumulating evidence that Hsp70 can be released from cells by an active mechanism, with functionally relevant consequences.

Intriguingly, while it was found that over-expression of Hsp70 can prevent αSyn aggregation in a Wt αSyn transgenic mouse model (Klucken et al., 2004b), a recent work based on A53T-αSyn transgenic mice has failed to observe this effect (Shimshek et al., 2010), seeding some controversy. Could these findings reflect a difference in the nature of the aggregates generated by Wt and mutant αSyn, and therefore a differential ability of Hsp70 to cope with those aggregates? Even though there is compelling evidence demonstrating the important role of Hsp70 under physiological and pathological scenarios in modulating fibril formation based on *in vivo* and cellular models, the molecular mechanism underlying such anti-aggregation properties in the context of PD, is still not fully understood.

Unlike the many research works performed with PD cellular and animal models that have contributed to our understanding on Hsp70 function under physiological or pathological conditions, only a handful of studies have focused on the molecular mechanism and interactions that underlie the modulation of αSyn aggregation exerted by Hsp70. An in-cell fluorescence resonance energy transfer (FRET) study indicated that Hsp70 alters the conformation of αSyn, inducing it to adopt a more 'open' conformation, without affecting the αSyn-αSyn intermolecular interactions (Klucken et al., 2006). Similarly to what had been found *in vitro* for Hsp70 and the HD-related huntingtin protein (Muchowski et al., 2000; Wacker et al., 2004), several studies with αSyn have shown that Hsp70 is able to suppress αSyn fibril assembly. In this case, a variety of *in vitro* studies have shown that this efficient

Heat-shock proteins (HSPs) prevent and reverse the misfolding and aggregation of proteins, and the Hsp70 family in particular is known to play important roles in protecting neurons from protein aggregation-derived stress (Lu et al., 2010). Therefore, it might not seem surprising that Hsp70 has been found to be linked to several neurodegenerative processes and 'conformational' disorders, including PD (Witt, 2009). Notably, Hsp70 has been shown to colocalize with aggregated αSyn within LBs in brains from PD patients (Auluck et al., 2002), strongly suggesting a role for this chaperone in managing αSyn aggregates in the context of PD (Figure 1). It is then that a substantial portion of the research in the field has focused on the effects of this major cytosolic chaperone on αSyn aggregation and cytotoxicity. Following the discovery that Hsp70 can abrogate the neurotoxicity of abnormal polyglutamine proteins (Warrick et al., 1999), it was shown that Hsp70 can also prevent dopaminergic neuronal loss associated with αSyn in a *Drosophila* PD model (Auluck et al., 2002). Numerous studies that followed have reported that over-expression of Hsp70 is able to reduce αSyn aggregation and/or toxicity in various cellular models (Danzer et al., 2011; Klucken et al., 2004b; McLean et al., 2004; Opazo et al., 2008; Outeiro et al., 2008; Zhou et al., 2004). In particular, McLean and co-workers (Outeiro et al., 2008) have found that Hsp70 rescues αSyn-linked toxicity by promoting the cellular clearance of αSyn oligomers, rather than monomers. Another study (Opazo et al., 2008) found that Hsp70 manages αSyn intracellular aggregation by increasing the clearing of aggregates primarily through the aggresome, and the subsequent removal of small aggregates and aggresomes from the cytosol. An interesting study by McLean and colleagues (Danzer et al., 2011) has recently shown that Hsp70 can also inhibit the formation of extracellular αSyn oligomers and rescue the cytotoxicity produced by such secreted oligomers, in a cellular model. Moreover, their data also indicates that Hsp70 is released to the extracellular medium together with secreted αSyn, adding to the accumulating evidence that Hsp70 can be released from cells by an

Intriguingly, while it was found that over-expression of Hsp70 can prevent αSyn aggregation in a Wt αSyn transgenic mouse model (Klucken et al., 2004b), a recent work based on A53T-αSyn transgenic mice has failed to observe this effect (Shimshek et al., 2010), seeding some controversy. Could these findings reflect a difference in the nature of the aggregates generated by Wt and mutant αSyn, and therefore a differential ability of Hsp70 to cope with those aggregates? Even though there is compelling evidence demonstrating the important role of Hsp70 under physiological and pathological scenarios in modulating fibril formation based on *in vivo* and cellular models, the molecular mechanism underlying such

Unlike the many research works performed with PD cellular and animal models that have contributed to our understanding on Hsp70 function under physiological or pathological conditions, only a handful of studies have focused on the molecular mechanism and interactions that underlie the modulation of αSyn aggregation exerted by Hsp70. An in-cell fluorescence resonance energy transfer (FRET) study indicated that Hsp70 alters the conformation of αSyn, inducing it to adopt a more 'open' conformation, without affecting the αSyn-αSyn intermolecular interactions (Klucken et al., 2006). Similarly to what had been found *in vitro* for Hsp70 and the HD-related huntingtin protein (Muchowski et al., 2000; Wacker et al., 2004), several studies with αSyn have shown that Hsp70 is able to suppress αSyn fibril assembly. In this case, a variety of *in vitro* studies have shown that this efficient

anti-aggregation properties in the context of PD, is still not fully understood.

**2.1 Hsp70 in modulation of** α**Syn aggregation and cytotoxicity** 

active mechanism, with functionally relevant consequences.

inhibition of amyloid assembly can occur in the absence of collaborating co-chaperones and in an ATP-independent manner (Ahmad, 2010; Dedmon et al., 2005; Huang et al., 2006; Luk et al., 2008; Roodveldt et al., 2009), while promoting formation of small oligomeric species of moderate toxicity (Roodveldt et al., 2009). However, our recent study demonstrates that, even though Hsp70 is indeed able to control αSyn fibrillation and its associated toxicity in an ATP-independent manner, it does so more efficiently in the presence of ATP (Roodveldt et al., 2009), consistent with the results obtained with a cellular model (Klucken et al., 2004a). Nevertheless, this ATP-dependent activity was found to require Hip co-chaperone activity, and not Hsp40 as observed in the case of mutant huntingtin protein (Lotz et al., 2010), suggesting that Hsp70 can function through different molecular mechanisms, depending on the nature of the aggregating client protein.

Interestingly, our study also showed that Hsp70 in the presence of ATP is prone to coaggregate with aggregating αSyn, presumably caused by the formation of a highly insoluble (ADP)Hsp70/αSyn complex. Surprisingly, this co-aggregation can be prevented by the addition of the co-chaperone Hip (St13), which had been found to be under-expressed in serum of PD patients since the early stages of the disease (Scherzer et al., 2007). The relevance of Hip in assisting the suppression of αSyn aggregation in an Hsp70-dependent manner was further supported by a study we performed in a *C. elegans* model of PD (Roodveldt et al., 2009), in which knock-down of Hip produced a much stronger PD phenotype than knock-down of Hsp70, shown by higher levels of αSyn aggregation in the former condition. This finding indicates that Hip co-chaperone could be important not only in modulating the chaperone's molecular mechanism, but also in guarding the functionality and availability of Hsp70 under certain conditions.

One relevant question relates to the nature of the species along the aggregation pathway of αSyn that are specifically recognized and targeted by Hsp70. Initially, Hsp70 was reported to bind to αSyn filaments *in vitro* (Lindersson et al., 2004). However, it is currently understood that Hsp70 does not disaggregate or alter the structural properties of mature αSyn fibrils (Dedmon et al., 2005), but rather inhibits fibril formation via interactions with soluble pre-fibrillar forms of αSyn. A few studies performed with cell extracts (Zhou et al., 2004) and with live cells (Klucken et al., 2006), in addition to experiments with purified proteins (Dedmon et al., 2005; Huang et al., 2006; Roodveldt et al., 2009), demonstrate the existence of molecular interactions between Hsp70 and αSyn oligomeric species. Even though previous attempts to isolate such complexes by co-immunoprecipitation or pulldown experiments had failed (Luk et al., 2008), we have been able to probe the formation of a complex between Hsp70 and αSyn oligomeric species by FRET (Roodveldt et al., 2009), suggesting that these interactions are transient (Luk et al., 2008; Roodveldt et al., 2009), besides being highly dynamic (Luk et al., 2008; Roodveldt et al., 2009).

In addition, we have recently shown by FRET and nuclear magnetic resonance (NMR) analyses that, contrary to what was previously thought, Hsp70 can also interact with αSyn monomeric species (Roodveldt et al., 2009) with an affinity constant lying within the low micromolar range. Moreover, the fluorimetric study revealed the existence of diverse Hsp70/αSyn complexes that are formed depending on the nucleotide state of the chaperone, either nucleotide-free, or ATP- or ADP-bound, that are sampled along the ATPase cycle. Based on these structural studies, we have proposed that a particularly 'compact' Hsp70/αSyn complex in the ADP-state of Hsp70 that arises during the aggregation process of αSyn, leads to Hsp70 being entrapped or sequestered by the oligomers. However, we

The Hsp70 Chaperone System in Parkinson's Disease 227

Beyond the clear genetic association between the UPS and PD, several studies have demonstrated a reduced proteolytic activity in the SN of PD patients when compared with aged-matched controls (McNaught et al., 2003; McNaught & Jenner, 2001; McNaught & Olanow, 2006). In culture of rat primary neurons, treatment with proteasomal inhibitors has shown to lead to the formation of αSyn- and ubiquitin- positive proteinaceous inclusions, analogous to LBs found in PD patients (McNaught et al., 2002a; Rideout et al., 2005). Moreover, systemic administration of proteasomal inhibitors in rats produced a behavioural and pathological phenotype strongly reminiscent of PD (McNaught et al., 2004), although recently some controversy arose in this matter (Manning-Bog et al., 2006). Further support to the active role played by proteasomal impairment in PD progression originated from studies with pesticides, such as rotenone and paraquat (Cook & Petrucelli, 2009). Mice treated with such environmental toxins display a strong reduction in proteolytic activity that is dependent on the presence of αSyn (Fornai et al., 2005). All together, these data

undoubtedly link αSyn aggregation with impairment of the UPS in PD progression.

the relentless, progressive nature of the disease (Vila & Przedborski, 2003).

Alpha-synuclein extracted from LBs in PD brains has been found to be mono- and diubiquitinated (Hasegawa et al., 2002), while soluble αSyn is mono-ubiquitinated by SIAH-1 and -2, but not by parkin (Liani et al., 2004; Rott et al., 2008). Instead, a modified, Oglycosylated, version of αSyn is a substrate for parkin-induced ubiquitination (Shimura et al., 2001), and it was shown that interaction of parkin with αSyn is mediated by synphilin-1 (Chung et al., 2001) as well as by the protein 14-3-3-η (Sato et al., 2006). Importantly, parkin has been shown to be able to rescue primary neurons from the toxic effects of αSyn (Petrucelli et al., 2002), suggesting that the two proteins share a common pathway that may determine the fate of dopaminergic neurons in PD. In addition, UCH-L1 may be involved in regulating the cytoplasmic abundance of αSyn, as it displays unexpected ubiquitin ligase activity that is also able to polyubiquitinate mono- and diubiquitinated αSyn (Liu et al., 2002). Since attachment of at least four ubiquitin molecules is known to be required for

In relation to protein misfolding and aggregation, it has been proposed that failure of the UPS to adequately remove misfolded or abnormal proteins may underlie demise of nigral cells in sporadic PD (McNaught et al., 2001). Furthermore, deficits in the 26/20S proteasome pathways are accompanied by protein accumulation and aggregation, which may also cause neurodegeneration (Chung et al., 2001), in line with recent findings that general intracellular aggregation of proteins into aggresomes can can inhibit the UPS (Bence et al., 2001). Moreover, chaperones of the heat-shock families, including HSP27, 40, 70, 60, 90, and 110, as well as components of the UPS, such as ubiquitin, UCH-L1 and parkin, are found in LBs extracted from PD patients *post mortem* (Shults, 2006). These combined pieces of evidence have attracted much attention lately as they imply that LBs could originate from ubiquitinrich aggresomes that the proteasomal components may not be able to process (McNaught et al., 2002b). One plausible mechanism would involve αSyn adopting abnormal protein conformations and overwhelming the cellular protein degradation systems (Wong & Cuervo, 2010), whereas deficits in the UPS machinery would challenge the cell's ability to detect and degrade misfolded proteins that can result in the formation of toxic early aggregates (McNaught et al., 2002b). The common outcome of this failure at different levels is thus expected to be a cellular build-up of unwanted toxic species that should have been cleared in otherwise healthy conditions. Minimal defects in the crucial protein turnover machinery may suffice to cause a slow demise of dopaminergic neurons, which may explain

propose that this compact complex can be stabilized by the Hip co-chaperone and therefore the co-aggregation of the chaperone, and ultimately the formation of αSyn fibrils, can be prevented (Roodveldt et al., 2009). In other words, our findings indicated that a decreased expression of Hip could facilitate depletion of Hsp70 by amyloidogenic polypeptides, impairing chaperone proteostasis. Interestingly, another case of chaperoneinhibition/depletion by αSyn oligomers, but not by unstructured monomers, was recently reported for the Hsp70/Hsp40 system (Hinault et al., 2010). In this case, the authors found that the inhibition of the chaperone system was predominantly caused by the sequestration or incapacitation, by off-pathway αSyn oligomers, of the J-domain (Hsp40) co-chaperone (Hinault et al., 2010).

What is the region on the αSyn molecule thought to be recognized and bound by Hsp70? One study mapped this region as the broad segment between residues 21 and 110 (Luk et al., 2008). Based on a predictive algorithm for Hsp70-binding regions (Rudiger et al., 1997) and FRET analyses to probe Hsp70-αSyn interactions (Roodveldt et al., 2009), our results indicate that Hsp70 can bind to the N-terminus and the central NAC region of the protein. The first binding region is involved in functional lipid interactions, while the second one comprises the stretch of hydrophobic residues that readily forms fibrils *in vitro* and is generally assumed to be involved in initiating the fibrillation process (Giasson et al., 2001). In addition, our results suggest that Hsp70 also interacts with the negatively charged Cterminus of the αSyn molecule, especially in the nucleotide-free state. Taken together, the studies carried out thus far demonstrate that Hsp70 modulates αSyn aggregation by interacting with the protein at different stages of aggregation, by recognizing essentially two or three regions in the αSyn molecule, and forming different transient complexes with the substrate. The strong binding versatility displayed by Hsp70 with aggregating αSyn might be possible thanks to the large structural flexibility conferred by the lid subdomain within the SBD, that has been recently described by Mayer and colleagues (Schlecht et al., 2011).

#### **2.2 Cooperation of Hsp70 with the ubiquitin-proteasome system (UPS) 2.2.1 Involvement of the ubiquitin-proteasome system in PD**

The ubiquitin-proteasome system (UPS) is responsible for the degradation of vital regulatory proteins that control almost every cellular function (Hershko & Ciechanover, 1998). The UPS system is composed of three classes of ubiquitinating enzymes (E1, E2 and E3) that activate, transfer and attach the small protein ubiquitin to the Lys residues of proteins that are targeted for degradation. Conjugation of at least four ubiquitin moieties acts as the degradation signal in a process that is initiated by the recognition of the ubiquitin linkage by the proteasome, which first catalyzes the unfolding and then the proteolysis of the targeted protein (Hershko & Ciechanover, 1998).

Impairment in the UPS has been linked to many neurodegenerative disorders and indeed to contribute to disease progression in PD (Ciechanover & Brundin, 2003; Cook & Petrucelli, 2009; McNaught et al., 2001). The involvement of UPS in PD arouse major interest with the identification of mutations in the E3 ubiquitin ligase parkin protein as a cause of autosomal recessive PD (Kitada et al., 1998). Both the loss of E3 activity and the possibility of incomplete or aberrant ubiquitination are proposed as causes of parkin-related PD (Giasson & Lee, 2003). A second member of the UPS involved in PD is the ubiquitin carboxy-terminal hydrolase-L1 (UCH-L1), and mutations in the uchl-1 gene cause dysfunction of this enzyme and lead to accumulation of toxic products (Leroy et al., 1998).

propose that this compact complex can be stabilized by the Hip co-chaperone and therefore the co-aggregation of the chaperone, and ultimately the formation of αSyn fibrils, can be prevented (Roodveldt et al., 2009). In other words, our findings indicated that a decreased expression of Hip could facilitate depletion of Hsp70 by amyloidogenic polypeptides, impairing chaperone proteostasis. Interestingly, another case of chaperoneinhibition/depletion by αSyn oligomers, but not by unstructured monomers, was recently reported for the Hsp70/Hsp40 system (Hinault et al., 2010). In this case, the authors found that the inhibition of the chaperone system was predominantly caused by the sequestration or incapacitation, by off-pathway αSyn oligomers, of the J-domain (Hsp40) co-chaperone

What is the region on the αSyn molecule thought to be recognized and bound by Hsp70? One study mapped this region as the broad segment between residues 21 and 110 (Luk et al., 2008). Based on a predictive algorithm for Hsp70-binding regions (Rudiger et al., 1997) and FRET analyses to probe Hsp70-αSyn interactions (Roodveldt et al., 2009), our results indicate that Hsp70 can bind to the N-terminus and the central NAC region of the protein. The first binding region is involved in functional lipid interactions, while the second one comprises the stretch of hydrophobic residues that readily forms fibrils *in vitro* and is generally assumed to be involved in initiating the fibrillation process (Giasson et al., 2001). In addition, our results suggest that Hsp70 also interacts with the negatively charged Cterminus of the αSyn molecule, especially in the nucleotide-free state. Taken together, the studies carried out thus far demonstrate that Hsp70 modulates αSyn aggregation by interacting with the protein at different stages of aggregation, by recognizing essentially two or three regions in the αSyn molecule, and forming different transient complexes with the substrate. The strong binding versatility displayed by Hsp70 with aggregating αSyn might be possible thanks to the large structural flexibility conferred by the lid subdomain within the SBD, that has been recently described by Mayer and colleagues (Schlecht et al., 2011).

**2.2 Cooperation of Hsp70 with the ubiquitin-proteasome system (UPS)** 

The ubiquitin-proteasome system (UPS) is responsible for the degradation of vital regulatory proteins that control almost every cellular function (Hershko & Ciechanover, 1998). The UPS system is composed of three classes of ubiquitinating enzymes (E1, E2 and E3) that activate, transfer and attach the small protein ubiquitin to the Lys residues of proteins that are targeted for degradation. Conjugation of at least four ubiquitin moieties acts as the degradation signal in a process that is initiated by the recognition of the ubiquitin linkage by the proteasome, which first catalyzes the unfolding and then the proteolysis of

Impairment in the UPS has been linked to many neurodegenerative disorders and indeed to contribute to disease progression in PD (Ciechanover & Brundin, 2003; Cook & Petrucelli, 2009; McNaught et al., 2001). The involvement of UPS in PD arouse major interest with the identification of mutations in the E3 ubiquitin ligase parkin protein as a cause of autosomal recessive PD (Kitada et al., 1998). Both the loss of E3 activity and the possibility of incomplete or aberrant ubiquitination are proposed as causes of parkin-related PD (Giasson & Lee, 2003). A second member of the UPS involved in PD is the ubiquitin carboxy-terminal hydrolase-L1 (UCH-L1), and mutations in the uchl-1 gene cause dysfunction of this enzyme

**2.2.1 Involvement of the ubiquitin-proteasome system in PD** 

the targeted protein (Hershko & Ciechanover, 1998).

and lead to accumulation of toxic products (Leroy et al., 1998).

(Hinault et al., 2010).

Beyond the clear genetic association between the UPS and PD, several studies have demonstrated a reduced proteolytic activity in the SN of PD patients when compared with aged-matched controls (McNaught et al., 2003; McNaught & Jenner, 2001; McNaught & Olanow, 2006). In culture of rat primary neurons, treatment with proteasomal inhibitors has shown to lead to the formation of αSyn- and ubiquitin- positive proteinaceous inclusions, analogous to LBs found in PD patients (McNaught et al., 2002a; Rideout et al., 2005). Moreover, systemic administration of proteasomal inhibitors in rats produced a behavioural and pathological phenotype strongly reminiscent of PD (McNaught et al., 2004), although recently some controversy arose in this matter (Manning-Bog et al., 2006). Further support to the active role played by proteasomal impairment in PD progression originated from studies with pesticides, such as rotenone and paraquat (Cook & Petrucelli, 2009). Mice treated with such environmental toxins display a strong reduction in proteolytic activity that is dependent on the presence of αSyn (Fornai et al., 2005). All together, these data undoubtedly link αSyn aggregation with impairment of the UPS in PD progression.

In relation to protein misfolding and aggregation, it has been proposed that failure of the UPS to adequately remove misfolded or abnormal proteins may underlie demise of nigral cells in sporadic PD (McNaught et al., 2001). Furthermore, deficits in the 26/20S proteasome pathways are accompanied by protein accumulation and aggregation, which may also cause neurodegeneration (Chung et al., 2001), in line with recent findings that general intracellular aggregation of proteins into aggresomes can can inhibit the UPS (Bence et al., 2001). Moreover, chaperones of the heat-shock families, including HSP27, 40, 70, 60, 90, and 110, as well as components of the UPS, such as ubiquitin, UCH-L1 and parkin, are found in LBs extracted from PD patients *post mortem* (Shults, 2006). These combined pieces of evidence have attracted much attention lately as they imply that LBs could originate from ubiquitinrich aggresomes that the proteasomal components may not be able to process (McNaught et al., 2002b). One plausible mechanism would involve αSyn adopting abnormal protein conformations and overwhelming the cellular protein degradation systems (Wong & Cuervo, 2010), whereas deficits in the UPS machinery would challenge the cell's ability to detect and degrade misfolded proteins that can result in the formation of toxic early aggregates (McNaught et al., 2002b). The common outcome of this failure at different levels is thus expected to be a cellular build-up of unwanted toxic species that should have been cleared in otherwise healthy conditions. Minimal defects in the crucial protein turnover machinery may suffice to cause a slow demise of dopaminergic neurons, which may explain the relentless, progressive nature of the disease (Vila & Przedborski, 2003).

Alpha-synuclein extracted from LBs in PD brains has been found to be mono- and diubiquitinated (Hasegawa et al., 2002), while soluble αSyn is mono-ubiquitinated by SIAH-1 and -2, but not by parkin (Liani et al., 2004; Rott et al., 2008). Instead, a modified, Oglycosylated, version of αSyn is a substrate for parkin-induced ubiquitination (Shimura et al., 2001), and it was shown that interaction of parkin with αSyn is mediated by synphilin-1 (Chung et al., 2001) as well as by the protein 14-3-3-η (Sato et al., 2006). Importantly, parkin has been shown to be able to rescue primary neurons from the toxic effects of αSyn (Petrucelli et al., 2002), suggesting that the two proteins share a common pathway that may determine the fate of dopaminergic neurons in PD. In addition, UCH-L1 may be involved in regulating the cytoplasmic abundance of αSyn, as it displays unexpected ubiquitin ligase activity that is also able to polyubiquitinate mono- and diubiquitinated αSyn (Liu et al., 2002). Since attachment of at least four ubiquitin molecules is known to be required for

The Hsp70 Chaperone System in Parkinson's Disease 229

degrade its cargo clients. A ternary complex involving BAG-1, Hsp70 and CHIP has been isolated, and BAG-1 is capable of stimulating CHIP mediated degradation of some proteins (Alberti et al., 2002). The co-chaperone HspBP1, in turn, competes with BAG-1 for binding to the ATPase domain of Hsp70 (Alberti et al., 2004). Notably, when in complex with HspBP1 bound Hsp70, CHIP shows a much reduced ubiquitin ligase activity and chaperone

In the context of neurodegeneration, CHIP has been shown to intervene in the degradation of misfolded aggregation-prone proteins associated with AD, PD and HD, as well as with spinocerebellar ataxia and spinal bulbar muscle atrophy (Adachi et al., 2007; Al-Ramahi et al., 2006; Hatakeyama et al., 2004; Shin et al., 2005; Urushitani et al., 2004). Indeed, CHIP appears to play an active role in modulating αSyn aggregation and degradation (Kalia et al., 2011; Shin et al., 2005; Tetzlaff et al., 2008). It was shown that CHIP interacts with both soluble and aggregated αSyn, and its over-expression reduces αSyn abundance and aggregation (Shin et al., 2005). Interestingly, CHIP induces αSyn clearance via two alternative pathways, one involving Hsp70 and proteasomal degradation, and another Hsp70-independent route that targets αSyn to lysosomes (Shin et al., 2005). A fluorescencecomplementation assay demonstrated that CHIP reduced significantly the abundance of αSyn toxic oligomers in cell culture, suggesting that this aberrant species is preferentially recognized by the co-chaperone (Tetzlaff et al., 2008). More recently, it was discovered that CHIP is an E3 ubiquitin ligase of αSyn since it efficiently conjugates ubiquitin moieties to this protein (Kalia et al., 2011). CHIP-catalyzed ubiquitinated forms of αSyn include monoand poly-ubiquitinated species, and the activity of CHIP depends on the presence of Hsp70 and the co-chaperone BAG5. Contrary to the enhancer activity reported for BAG1, BAG5 reduces the ability of CHIP to ubiquitinate αSyn in and Hsp70-dependent manner (Kalia et

Additional pathogenic mechanisms in PD that involve the activity of CHIP relate to the ubiquitin ligase activity of parkin and the kinase activity of the leucine-rich repeat kinase-2 (LRRK2). It has been found that CHIP, Hsp70 and parkin form a ternary complex that promotes ubiquitination and degradation of the Pael receptor, a protein localized in the ER and whose accumulation has been linked to dopaminergic neuronal death (Imai et al., 2002), and CHIP has been proposed to enhance the activity of parkin, even in the absence of Hsp70 (Imai et al., 2002). Concerning LRRK2, CHIP regulates the ubiquitination, degradation, and toxicity mediated by pathogenic mutations of this kinase (Ko et al., 2009). Moreover, it has been shown that CHIP binds to LRRK2 via its TPR motifs and formation of this complex is protective in cell culture models, while in the presence of mutant LRKK2, knock-down of

Macroautophay and chaperone-mediated autophagy (CMA) are the two main lysosomal proteolytic systems in mammalian cells for the degradation of intracellular proteins (Xilouri & Stefanis, 2011). CMA is the process of degradation of intracellular components by lysosomes which selectively degrades cytotolic proteins containing a KFERQ-like motif (Koga & Cuervo, 2010). This process is known to involve binding of a complex of constitutive cytosolic Hsp70 (Hsc70) and co-chaperones (including Hsp40, Hip, Hsp90 and BAG1) to substrate proteins, and their subsequent targeting to lysosomes via the lysosomal surface receptor LAMP-2A (Xilouri & Stefanis, 2011). The substrate protein is subsequently

substrates are not longer ubiquitinated and targeted to the UPS (Alberti et al., 2004).

al., 2011).

CHIP leads to cell death (Ko et al., 2009).

**2.3 Hsp70 in chaperone-mediated autophagy (CMA)** 

protein degradation via the UPS (Hershko & Ciechanover, 1998), it is likely that mono-, di-, and aberrant poly-ubiquitinated αSyn could have a pathogenic impact (Rott et al., 2008; Sun & Chen, 2004).

### **2.2.2 Role of the C-terminus Hsp70-interacting protein (CHIP) in PD**

Hsp70 and Hsp90 family members as well as small HSPs all take part in the degradation of protein substrate and are able to cooperate with the UPS towards this goal (Patterson & Hohfeld, 2006). Notably, protein homeostasis thus appears to be tightly controlled by interplay between the protein folding and protein degradation systems. Hsp70 takes part in the degradation of immature and aberrant forms of certain proteins, particularly ER-bound membrane proteins (Taxis et al., 2003), but also some cytosolic and nuclear proteins (Bercovich et al., 1997). For example, Hsp70 assists in the folding of the aggregation prone cystic fibrosis transmembrane conductance regulator (CFTR); however, Hsp70 is also able to present CFTR to the UPS and thus to control precisely the abundance of this protein, which is known to accumulate in aggresome and to cause cystic fibrosis (Zhang et al., 2001). Indeed, Hsp70 is actively recruited to aggresome and it has been proposed that this chaperone can reduce aggresome formation by stimulating proteasomal degradation of misfolded proteins (Dul et al., 2001; Garcia-Mata et al., 1999). The proposed mechanism for chaperone-UPS cooperation is that both systems compete for the same misfolded and aggregation-prone substrate proteins, and that the efficiency of chaperones in maintaining these proteins in solution increments the probability of the UPS to degrade aberrant polypeptides. Conversely, failure of chaperones to keep misfolded proteins in a soluble state gives rise to aggregates that are not efficiently degraded by the UPS and, moreover, may inhibit UPS activity towards other protein targets, altering protein homeostasis (Figure 1).

Insights into the mechanism that enables the cooperation of protein chaperones with the UPS have been obtained from the identification and functional characterization of the Cterminal Hsp70 interacting protein (CHIP) co-chaperone (Ballinger et al., 1999; Hohfeld et al., 2001; McDonough & Patterson, 2003). CHIP contains three tetratricopeptide repeat (TPR) domains that recognize and bind to the EEVD motif in both Hsp70 and Hsp90 (Ballinger et al., 1999). CHIP-complexed Hsp70 displays reduced ATP hydrolysis *in vitro*, suggesting that the co-chaperone diminishes the on-rate of binding and release cycles in Hsp70 (Ballinger et al., 1999). In its C-terminus, CHIP possesses a ubiquitin ligase domain (U box) that is capable of targeting proteins to degradation, in a homologous manner to RING finger domains found in E3 ubiquitin ligase enzymes (Connell et al., 2001; Meacham et al., 2001).

CHIP has been shown to efficiently act as E3 ligase for several Hsp70 and Hsp90 substrates, such as the glucocorticoid receptor and Erb2 (Dickey et al., 2007; McDonough & Patterson, 2003; Murata et al., 2001). Proteins ubiquitinated by CHIP are efficiently targeted to the proteasome and subsequently degraded. Notably, the UbcH4/UbcH5 proteins are E2 conjugating enzymes (Demand et al., 2001) that are stress-activated, suggesting that upon stress CHIP furnishes the cell with chaperone-dependent ubiquitin ligases capable of ubiquitinating misfolded and aggregation-prone substrates.

The ubiquitinating activity of CHIP is regulated by two co-chaperones, BAG-1, which is an enhancer, and the Hsp70 binding protein 1 (HspBP1), which acts as a repressor. BAG-1 associates with the ATPase domain of Hsp70 and, in addition, possesses a ubiquitin-like domain that is efficiently recognized and bound by the proteasome (Luders et al., 2000). Thus, BAG-1 provides the proteasome with the capability to recruit Hsp70 and consequently

protein degradation via the UPS (Hershko & Ciechanover, 1998), it is likely that mono-, di-, and aberrant poly-ubiquitinated αSyn could have a pathogenic impact (Rott et al., 2008; Sun

Hsp70 and Hsp90 family members as well as small HSPs all take part in the degradation of protein substrate and are able to cooperate with the UPS towards this goal (Patterson & Hohfeld, 2006). Notably, protein homeostasis thus appears to be tightly controlled by interplay between the protein folding and protein degradation systems. Hsp70 takes part in the degradation of immature and aberrant forms of certain proteins, particularly ER-bound membrane proteins (Taxis et al., 2003), but also some cytosolic and nuclear proteins (Bercovich et al., 1997). For example, Hsp70 assists in the folding of the aggregation prone cystic fibrosis transmembrane conductance regulator (CFTR); however, Hsp70 is also able to present CFTR to the UPS and thus to control precisely the abundance of this protein, which is known to accumulate in aggresome and to cause cystic fibrosis (Zhang et al., 2001). Indeed, Hsp70 is actively recruited to aggresome and it has been proposed that this chaperone can reduce aggresome formation by stimulating proteasomal degradation of misfolded proteins (Dul et al., 2001; Garcia-Mata et al., 1999). The proposed mechanism for chaperone-UPS cooperation is that both systems compete for the same misfolded and aggregation-prone substrate proteins, and that the efficiency of chaperones in maintaining these proteins in solution increments the probability of the UPS to degrade aberrant polypeptides. Conversely, failure of chaperones to keep misfolded proteins in a soluble state gives rise to aggregates that are not efficiently degraded by the UPS and, moreover, may inhibit UPS activity towards other protein targets, altering protein homeostasis (Figure 1). Insights into the mechanism that enables the cooperation of protein chaperones with the UPS have been obtained from the identification and functional characterization of the Cterminal Hsp70 interacting protein (CHIP) co-chaperone (Ballinger et al., 1999; Hohfeld et al., 2001; McDonough & Patterson, 2003). CHIP contains three tetratricopeptide repeat (TPR) domains that recognize and bind to the EEVD motif in both Hsp70 and Hsp90 (Ballinger et al., 1999). CHIP-complexed Hsp70 displays reduced ATP hydrolysis *in vitro*, suggesting that the co-chaperone diminishes the on-rate of binding and release cycles in Hsp70 (Ballinger et al., 1999). In its C-terminus, CHIP possesses a ubiquitin ligase domain (U box) that is capable of targeting proteins to degradation, in a homologous manner to RING finger domains found in E3 ubiquitin ligase enzymes (Connell et al., 2001; Meacham et al., 2001). CHIP has been shown to efficiently act as E3 ligase for several Hsp70 and Hsp90 substrates, such as the glucocorticoid receptor and Erb2 (Dickey et al., 2007; McDonough & Patterson, 2003; Murata et al., 2001). Proteins ubiquitinated by CHIP are efficiently targeted to the proteasome and subsequently degraded. Notably, the UbcH4/UbcH5 proteins are E2 conjugating enzymes (Demand et al., 2001) that are stress-activated, suggesting that upon stress CHIP furnishes the cell with chaperone-dependent ubiquitin ligases capable of

**2.2.2 Role of the C-terminus Hsp70-interacting protein (CHIP) in PD** 

ubiquitinating misfolded and aggregation-prone substrates.

The ubiquitinating activity of CHIP is regulated by two co-chaperones, BAG-1, which is an enhancer, and the Hsp70 binding protein 1 (HspBP1), which acts as a repressor. BAG-1 associates with the ATPase domain of Hsp70 and, in addition, possesses a ubiquitin-like domain that is efficiently recognized and bound by the proteasome (Luders et al., 2000). Thus, BAG-1 provides the proteasome with the capability to recruit Hsp70 and consequently

& Chen, 2004).

degrade its cargo clients. A ternary complex involving BAG-1, Hsp70 and CHIP has been isolated, and BAG-1 is capable of stimulating CHIP mediated degradation of some proteins (Alberti et al., 2002). The co-chaperone HspBP1, in turn, competes with BAG-1 for binding to the ATPase domain of Hsp70 (Alberti et al., 2004). Notably, when in complex with HspBP1 bound Hsp70, CHIP shows a much reduced ubiquitin ligase activity and chaperone substrates are not longer ubiquitinated and targeted to the UPS (Alberti et al., 2004).

In the context of neurodegeneration, CHIP has been shown to intervene in the degradation of misfolded aggregation-prone proteins associated with AD, PD and HD, as well as with spinocerebellar ataxia and spinal bulbar muscle atrophy (Adachi et al., 2007; Al-Ramahi et al., 2006; Hatakeyama et al., 2004; Shin et al., 2005; Urushitani et al., 2004). Indeed, CHIP appears to play an active role in modulating αSyn aggregation and degradation (Kalia et al., 2011; Shin et al., 2005; Tetzlaff et al., 2008). It was shown that CHIP interacts with both soluble and aggregated αSyn, and its over-expression reduces αSyn abundance and aggregation (Shin et al., 2005). Interestingly, CHIP induces αSyn clearance via two alternative pathways, one involving Hsp70 and proteasomal degradation, and another Hsp70-independent route that targets αSyn to lysosomes (Shin et al., 2005). A fluorescencecomplementation assay demonstrated that CHIP reduced significantly the abundance of αSyn toxic oligomers in cell culture, suggesting that this aberrant species is preferentially recognized by the co-chaperone (Tetzlaff et al., 2008). More recently, it was discovered that CHIP is an E3 ubiquitin ligase of αSyn since it efficiently conjugates ubiquitin moieties to this protein (Kalia et al., 2011). CHIP-catalyzed ubiquitinated forms of αSyn include monoand poly-ubiquitinated species, and the activity of CHIP depends on the presence of Hsp70 and the co-chaperone BAG5. Contrary to the enhancer activity reported for BAG1, BAG5 reduces the ability of CHIP to ubiquitinate αSyn in and Hsp70-dependent manner (Kalia et al., 2011).

Additional pathogenic mechanisms in PD that involve the activity of CHIP relate to the ubiquitin ligase activity of parkin and the kinase activity of the leucine-rich repeat kinase-2 (LRRK2). It has been found that CHIP, Hsp70 and parkin form a ternary complex that promotes ubiquitination and degradation of the Pael receptor, a protein localized in the ER and whose accumulation has been linked to dopaminergic neuronal death (Imai et al., 2002), and CHIP has been proposed to enhance the activity of parkin, even in the absence of Hsp70 (Imai et al., 2002). Concerning LRRK2, CHIP regulates the ubiquitination, degradation, and toxicity mediated by pathogenic mutations of this kinase (Ko et al., 2009). Moreover, it has been shown that CHIP binds to LRRK2 via its TPR motifs and formation of this complex is protective in cell culture models, while in the presence of mutant LRKK2, knock-down of CHIP leads to cell death (Ko et al., 2009).

#### **2.3 Hsp70 in chaperone-mediated autophagy (CMA)**

Macroautophay and chaperone-mediated autophagy (CMA) are the two main lysosomal proteolytic systems in mammalian cells for the degradation of intracellular proteins (Xilouri & Stefanis, 2011). CMA is the process of degradation of intracellular components by lysosomes which selectively degrades cytotolic proteins containing a KFERQ-like motif (Koga & Cuervo, 2010). This process is known to involve binding of a complex of constitutive cytosolic Hsp70 (Hsc70) and co-chaperones (including Hsp40, Hip, Hsp90 and BAG1) to substrate proteins, and their subsequent targeting to lysosomes via the lysosomal surface receptor LAMP-2A (Xilouri & Stefanis, 2011). The substrate protein is subsequently

The Hsp70 Chaperone System in Parkinson's Disease 231

SNAP-25, a pre-synaptic SNARE protein, and prevent its aggregation, thus promoting the formation of the SNARE complex involved in neuronal synapse. Interestingly, another recent study using a transgenic mouse model of αSyn aggregation with associated neurodegeneration, has shown that SNAP-25 is redistributed within synaptic terminals and the protein was found to colocalize with αSyn within intraneuronal aggregates (Garcia-Reitbock et al., 2010). These data raise the question whether aggregating αSyn might actually be sequestering CSPα involved in the CSPα-Hsc70-SGT complex, analogously to what was observed for aggregating huntingtin (Miller et al., 2003), and therefore enabling

The unfolded protein response (UPR) is a mechanism activated within the cell when the endoplasmic reticulum (ER) function is impaired and, as a result, unfolded proteins accumulate in the ER lumen (a process called 'ER stress'). The ER-resident Hsp70 family member GRP78 (HSPA5/BiP) recognizes and binds to such unfolded proteins, which causes the release of UPR-activating factors (Rutkowski & Kaufman, 2004). To restore ER function, the UPR reduces protein translation and enhances the folding and processing capacities within ER. However, if the stress overwhelms the cell restoring capacity, the UPR induces apoptosis (Paschen & Mengesdorf, 2005). A few years ago, a mutation in the parkin gene was found to be related to familiar PD due to the impairment of its ubiquitin ligase function which results in protein accumulation within the ER lumen and leads to ER stress and apoptosis (Imai et al., 2001). More recently, a strong correlation between UPR activation and PD pathogenesis in PD patients was established (Hoozemans et al., 2007), suggesting a functional connection between αSyn and ER stress. In addition, it was found that overexpression of (Wt) αSyn triggers UPR in yeast (Cooper et al., 2006). A recent study has shown that αSyn is also found aggregated and accumulated within the ER lumen and induces UPR by binding to GRP78/BiP in cells, which could lead to apoptosis (Bellucci et al., 2011). Clearly, supplementary investigations are needed to determine the exact role of

In both familial and sporadic forms of PD, several mitochondrial alterations and increase of oxidative species are well recorded (Jellinger, 2010). The important role of mitochondrial pathology in PD is reflected by the specific and selective loss of mitochondrial complex I activity in the SN of PD patients (Valente et al., 2004). Results from a study with neuronal cell cultures indicate that this impairment is dependent on αSyn mitochondrial import and accumulation (Devi et al., 2008). Interestingly, a recent proteomic study revealed that expression of mortalin (HSPA9/mtHSP70/GRP75), a mitochondrial stress protein and member of the Hsp70 chaperone family which binds to DJ-1 and αSyn (Jin et al., 2007), is significantly decreased in PD brains (De Mena et al., 2009; Jin et al., 2006) as well as in a cellular model of PD (Jin et al., 2006). Moreover, specific coding mutational variants of the mortalin gene have been recently discovered in a few PD patients (Burbulla et al., 2010; De Mena et al., 2009). Finally, differential levels of mitochondrial mortalin were measured in Wt and A53T cellular models of PD (Pennington et al., 2010), suggesting a possible involvement

of αSyn aggregation in PD-related mitochondrial dysfunction (Xie et al., 2010).

the aggregation of SNAP-25 in a PD scenario.

the UPR in the pathophysiology of PD.

**2.4.3 Mitochondrial dysfunction in PD** 

**2.4.2 The unfolded protein response (UPR) in PD** 

degraded after unfolding and translocation into the lysosomal lumen, in a process involving lysosomal Hsc70 (Figure 1). It has been estimated that about 30% of cytosolic proteins could be subjected to degradation via CMA (Dice, 2007), and furthermore, this pathway may be a major route by which αSyn is degraded in neurons (Witt, 2009). Even though it is currently accepted that dysregulation of autophagy plays a role in neurodegeneration (Bandyopadhyay & Cuervo, 2007; Nixon, 2006; Rubinsztein, 2006; Xilouri & Stefanis, 2011), including the PD neurodegenerative process (Martinez-Vicente et al., 2008; Yang et al., 2009), the mechanism by which CMA modulates neuronal survival or death, is still unclear. Given that pathologic accumulation of αSyn is a hallmark of PD, several recent studies have addressed the possible link between αSyn degradation, CMA dysfunction and the neurodegenerative process. Indeed, αSyn, which contains a pentapeptide sequence (95VKKDQ99) consistent with Hsc70 binding (Dice, 1990), has been shown to be degraded via CMA using isolated lysosomal preparations (Cuervo et al., 2004) and neuronal cells (Alvarez-Erviti et al., 2010; Martinez-Vicente et al.; Vogiatzi et al., 2008). On the contrary, the A30P and A53T αSyn variants were observed to bind strongly to LAMP-2A receptors but were not internalized, thus inhibiting the CMA degradation of other substrates (Cuervo et al., 2004). This CMA dysfunction was later shown to mediate αSyn toxicity in cellular models (Xilouri et al., 2009). Moreover, a recent study using both αSyn transgenic- and paraquat- PD mouse models (Mak et al., 2010), showed that αSyn can be degraded in the lysosome through CMA, *in vivo*. The study also revealed an up-regulation of LAMP-2A and lysosomal Hsc70 and an increase in Hsc70-αSyn interactions in brain lysosomes, relative to controls. On the other hand, a recent work work revealed a significant reduction of both LAMP-2A and Hsc70 levels in the SN and amygdala of PD brains, relative to age-matched Alzheimer's disease (AD) and healthy, brain controls (Alvarez-Erviti et al., 2010). Even though their results might initially appear contradictory, these findings support a key role for Hsc70 and the CMA system in maintaining intracellular general proteostasis, especially within the αSyn-overload scenario that is typically associated to PD and other αsynucleinopathies.

#### **2.4 Emerging links between the Hsp70 system and neurodegeneration in PD 2.4.1 The CSP**α**-Hsc70-SGT complex and neurodegeneration in PD**

Cysteine-string protein α (CSPα) is an abundant protein localized in synaptic vesicles that ameliorates neurodegeneration in cellular and animal models (Johnson et al., 2010). It contains a Dna-J domain and has been shown to interact with Hsc70 and to increase its ATPase activity (Braun et al., 1996). On the other hand, CSPα was shown to assemble into an enzymatically active ternary complex with Hsc70 and SGT (small glutamine-rich tetratricopeptide repeat domain protein) with a likely regulatory function in secretory vesicles (Tobaben et al., 2001). CSPα has also been reported to interact with other chaperones, including Hsp90 (Sakisaka et al., 2002), Hip, Hop (Rosales-Hernandez et al., 2009), and Hsp40 (Gibbs et al., 2009).

CSPα dysfunction has been implicated in various pathologies, including memory impairment, type-2 diabetes, cystic fibrosis, and HD (reviewed in (Johnson et al., 2010)). A few years ago, a link between CSPα and PD was also established (Chandra et al., 2005). The results of this study, performed using a transgenic mouse model, indicate that αSyn cooperates with CSPα in preventing neurodegeneration (Chandra et al., 2005). In a recent study (Sharma et al., 2011), the CSPα-Hsc70-SGT complex was found to bind to monomeric

degraded after unfolding and translocation into the lysosomal lumen, in a process involving lysosomal Hsc70 (Figure 1). It has been estimated that about 30% of cytosolic proteins could be subjected to degradation via CMA (Dice, 2007), and furthermore, this pathway may be a major route by which αSyn is degraded in neurons (Witt, 2009). Even though it is currently accepted that dysregulation of autophagy plays a role in neurodegeneration (Bandyopadhyay & Cuervo, 2007; Nixon, 2006; Rubinsztein, 2006; Xilouri & Stefanis, 2011), including the PD neurodegenerative process (Martinez-Vicente et al., 2008; Yang et al., 2009), the mechanism by which CMA modulates neuronal survival or death, is still unclear. Given that pathologic accumulation of αSyn is a hallmark of PD, several recent studies have addressed the possible link between αSyn degradation, CMA dysfunction and the neurodegenerative process. Indeed, αSyn, which contains a pentapeptide sequence (95VKKDQ99) consistent with Hsc70 binding (Dice, 1990), has been shown to be degraded via CMA using isolated lysosomal preparations (Cuervo et al., 2004) and neuronal cells (Alvarez-Erviti et al., 2010; Martinez-Vicente et al.; Vogiatzi et al., 2008). On the contrary, the A30P and A53T αSyn variants were observed to bind strongly to LAMP-2A receptors but were not internalized, thus inhibiting the CMA degradation of other substrates (Cuervo et al., 2004). This CMA dysfunction was later shown to mediate αSyn toxicity in cellular models (Xilouri et al., 2009). Moreover, a recent study using both αSyn transgenic- and paraquat- PD mouse models (Mak et al., 2010), showed that αSyn can be degraded in the lysosome through CMA, *in vivo*. The study also revealed an up-regulation of LAMP-2A and lysosomal Hsc70 and an increase in Hsc70-αSyn interactions in brain lysosomes, relative to controls. On the other hand, a recent work work revealed a significant reduction of both LAMP-2A and Hsc70 levels in the SN and amygdala of PD brains, relative to age-matched Alzheimer's disease (AD) and healthy, brain controls (Alvarez-Erviti et al., 2010). Even though their results might initially appear contradictory, these findings support a key role for Hsc70 and the CMA system in maintaining intracellular general proteostasis, especially within the αSyn-overload scenario that is typically associated to PD and other α-

**2.4 Emerging links between the Hsp70 system and neurodegeneration in PD** 

Cysteine-string protein α (CSPα) is an abundant protein localized in synaptic vesicles that ameliorates neurodegeneration in cellular and animal models (Johnson et al., 2010). It contains a Dna-J domain and has been shown to interact with Hsc70 and to increase its ATPase activity (Braun et al., 1996). On the other hand, CSPα was shown to assemble into an enzymatically active ternary complex with Hsc70 and SGT (small glutamine-rich tetratricopeptide repeat domain protein) with a likely regulatory function in secretory vesicles (Tobaben et al., 2001). CSPα has also been reported to interact with other chaperones, including Hsp90 (Sakisaka et al., 2002), Hip, Hop (Rosales-Hernandez et al.,

CSPα dysfunction has been implicated in various pathologies, including memory impairment, type-2 diabetes, cystic fibrosis, and HD (reviewed in (Johnson et al., 2010)). A few years ago, a link between CSPα and PD was also established (Chandra et al., 2005). The results of this study, performed using a transgenic mouse model, indicate that αSyn cooperates with CSPα in preventing neurodegeneration (Chandra et al., 2005). In a recent study (Sharma et al., 2011), the CSPα-Hsc70-SGT complex was found to bind to monomeric

**2.4.1 The CSP**α**-Hsc70-SGT complex and neurodegeneration in PD** 

synucleinopathies.

2009), and Hsp40 (Gibbs et al., 2009).

SNAP-25, a pre-synaptic SNARE protein, and prevent its aggregation, thus promoting the formation of the SNARE complex involved in neuronal synapse. Interestingly, another recent study using a transgenic mouse model of αSyn aggregation with associated neurodegeneration, has shown that SNAP-25 is redistributed within synaptic terminals and

the protein was found to colocalize with αSyn within intraneuronal aggregates (Garcia-Reitbock et al., 2010). These data raise the question whether aggregating αSyn might actually be sequestering CSPα involved in the CSPα-Hsc70-SGT complex, analogously to what was observed for aggregating huntingtin (Miller et al., 2003), and therefore enabling the aggregation of SNAP-25 in a PD scenario.

#### **2.4.2 The unfolded protein response (UPR) in PD**

The unfolded protein response (UPR) is a mechanism activated within the cell when the endoplasmic reticulum (ER) function is impaired and, as a result, unfolded proteins accumulate in the ER lumen (a process called 'ER stress'). The ER-resident Hsp70 family member GRP78 (HSPA5/BiP) recognizes and binds to such unfolded proteins, which causes the release of UPR-activating factors (Rutkowski & Kaufman, 2004). To restore ER function, the UPR reduces protein translation and enhances the folding and processing capacities within ER. However, if the stress overwhelms the cell restoring capacity, the UPR induces apoptosis (Paschen & Mengesdorf, 2005). A few years ago, a mutation in the parkin gene was found to be related to familiar PD due to the impairment of its ubiquitin ligase function which results in protein accumulation within the ER lumen and leads to ER stress and apoptosis (Imai et al., 2001). More recently, a strong correlation between UPR activation and PD pathogenesis in PD patients was established (Hoozemans et al., 2007), suggesting a functional connection between αSyn and ER stress. In addition, it was found that overexpression of (Wt) αSyn triggers UPR in yeast (Cooper et al., 2006). A recent study has shown that αSyn is also found aggregated and accumulated within the ER lumen and induces UPR by binding to GRP78/BiP in cells, which could lead to apoptosis (Bellucci et al., 2011). Clearly, supplementary investigations are needed to determine the exact role of the UPR in the pathophysiology of PD.

#### **2.4.3 Mitochondrial dysfunction in PD**

In both familial and sporadic forms of PD, several mitochondrial alterations and increase of oxidative species are well recorded (Jellinger, 2010). The important role of mitochondrial pathology in PD is reflected by the specific and selective loss of mitochondrial complex I activity in the SN of PD patients (Valente et al., 2004). Results from a study with neuronal cell cultures indicate that this impairment is dependent on αSyn mitochondrial import and accumulation (Devi et al., 2008). Interestingly, a recent proteomic study revealed that expression of mortalin (HSPA9/mtHSP70/GRP75), a mitochondrial stress protein and member of the Hsp70 chaperone family which binds to DJ-1 and αSyn (Jin et al., 2007), is significantly decreased in PD brains (De Mena et al., 2009; Jin et al., 2006) as well as in a cellular model of PD (Jin et al., 2006). Moreover, specific coding mutational variants of the mortalin gene have been recently discovered in a few PD patients (Burbulla et al., 2010; De Mena et al., 2009). Finally, differential levels of mitochondrial mortalin were measured in Wt and A53T cellular models of PD (Pennington et al., 2010), suggesting a possible involvement of αSyn aggregation in PD-related mitochondrial dysfunction (Xie et al., 2010).

The Hsp70 Chaperone System in Parkinson's Disease 233

As described in this chapter, in the last few years it has become evident that HSPs play an important role in the initiation and progression of PD and other neurodegenerative diseases. This cumulative evidence has prompted the development of therapeutic tools based on the Hsp70 machinery. Different strategies have been tested to manipulate Hsp70 as a therapeutic approximation for PD and related neurodegenerative diseases (reviewed in Kalia 2010). Three general approaches have been explored, namely, to increase the intracellular activity of Hsp70, to overexpress Hsp70 and/or other co-chaperones, and to

Theoretically, it should be feasible to control the activity of the cellular chaperone machinery by using different types of drugs. The mechanism of most of such chemical compounds is based on activating HSF-1, a key transcriptional regulator of the heat shock response (HSR) that activates the gene expression of inducible HSPs. One way of activating HSF-1 is by inhibiting Hsp90 activity; as a result, HSF-1 becomes active and increases the expression of inducible chaperones like HSPA1A/Hsp72 and others. Geldanamycin (GA) and its derivative 17-allylamino-17-demethoxygeldanamycin (17-AAG) are antibiotics with high affinity for the ATPase domain of Hsp90, which blocks its interaction with HSF-1 and subsequently allows its activation (Waza et al., 2006; Zou et al., 1998). GA prevents αSyninduced dopaminergic cell loss in cell culture (McLean et al., 2004) and in animal models of PD (Auluck et al., 2005; Shen et al., 2005), while the less toxic 17-AAG has also proven to be neuroprotective in PD cellular models (Danzer et al., 2011; Riedel et al., 2010) and in two animal models of PD-related neurodegenerative diseases (Fujikake et al., 2008; Waza et al., 2005). Even though 17-AAG is currently under phase-II clinical trials as an anti-tumour drug (Pacey et al., 2010; Richardson et al., 2010; Solit et al., 2008), its use in patients with neurodegenerative diseases could be hampered by its toxicity and unavailability for oral administration (Pacey et al., 2010). Another family of inhibitors of Hsp90 activity, is SNX-2112 and its analogues, which are orally available and present improved blood brain barrier (BBB) permeability. In particular, SNX-0723 was shown to prevent αSyn oligomer formation and αSyn-induced toxicity in cell culture (Putcha et al., 2010), and preclinical studies for cancer therapy are proving their safety (Zhai et al., 2011), although similar studies in

Other drugs are able to activate HSF-1 without inhibiting Hsp90 activity, which could represent a less toxic approximation for neurodegenerative pathologies. Arimoclomol, for example, has already been tested in phase I- and IIa-clinical trials for treating ALS, and shown to be safe and tolerable (Cudkowicz et al., 2008; Phukan, 2010). Another example is HSF-1A (Neef et al., 2010), shown to upregulate Hsp70 expression and to reduce poly-Qcitotoxicity in cell and fly models of poly-Q neurodegenerative disorders (Neef et al., 2010). Celastrol, yet a similar drug, and some structural relatives, appear as promising drugs due to their rapid kinetics and low EC50 (Westerheide et al., 2004), although further

Besides these approaches, HSF-1 co-inducers could represent more tolerable drugs for therapy. These are molecules that partially activate HSF-1, reducing its activation threshold and often working in conjunction with secondary stress signals to fully induce HSR. Non steroidal anti-

deliver chaperones or regulatory factors using cell-penetrating peptides (CPPs).

**4. The Hsp70 system as a therapeutic tool for PD** 

**4.1 Increasing the intracellular activity of Hsp70** 

neurodegenerative animal models are still needed.

studies are needed.
