CHAPTER 1: Overview of Molecular Chaperones in Health and Disease

After being synthesized on ribosomes as linear amino acid chains, proteins need to be folded into their native states, a dynamic equilibrium of closely related three-dimensional structures. In addition to this initial process, cells also need protein quality control and the maintenance of proteome homeostasis (known as proteostasis), both of which are crucial for cellular and organismal health. That is why many diseases appear to be caused by misregulation of protein maintenance. Examples of this are the loss-of-function diseases such as cystic fibrosis and the gain-of-toxic-function diseases such as Alzheimer’s, Parkinson’s and Huntington’s diseases. Proteostasis is maintained by a complex regulatory network, which comprises proteins, cofactors and processes that control protein synthesis, folding, trafficking, aggregation, disaggregation and degradation.¹  Molecular chaperones and their regulators are central players of the proteostatic network.² 

Molecular chaperones interact with their targets to provide a temporary stabilization, which facilitates folding into a functionally active conformation, unfolding for degradation, or assembly or disassembly of multi-component complexes.³  Typically, molecular chaperones are not associated with their target proteins once these have acquired their final functional conformation.⁴  Recent findings show that there are exceptions. It was recently demonstrated that the glucorticoid receptor (GR) translocates into the nucleus as an Hsp90 heterocomplex upon stimulation by glucocorticoids before dissociating within the nucleus.⁵  Moreover, the intranuclear dynamics of GR is still Hsp90-dependent,^6,7  which indicates that some molecular chaperones may escort their clients during their entire lifespan.

Molecular chaperones further ensure proteostasis and prevent proteotoxicity by promoting protein degradation and disposal through multiple pathways. For example, Hsp70, Hsp90 and their co-chaperones target unfolded proteins for degradation via the ubiquitin–proteasome system.^8–11  They also assist autophagy. Recent studies show the contribution of Hsp70 to this type of removal of pathogenic proteins. Hsp70 is involved in a certain type of autophagy involving late endosomes known as endosomal microautophagy,¹²  and in a form of macroautophagy mediating the degradation of protein aggregates known as chaperone-assisted selective autophagy.¹³  The more selective chaperone-mediated autophagy (CMA) requires that cytosolic proteins that contain the pentapeptide targeting motif KFERQ¹⁴  are recognized by heat-shock cognate 70 (Hsc70) and delivered to the surface of lysosomes¹⁵  to be translocated and degraded by lysosomal proteases.¹⁶  A fraction of Hsp90 is present at lysosomes, bound to the luminal side of the lysosomal membrane, and it can either increase or decrease CMA activity depending on the cell type.^17,18 

Molecular chaperones are classified into five families according to their molecular size, namely Hsp100, Hsp90, Hsp70 and J proteins, chaperonins and small heat-shock proteins (sHsp).² 

Hsp100 chaperones are members of a large superfamily of AAA+ ATPases. They form oligomeric rings involved in protein refolding, disaggregation and degradation.¹⁹  They are found in bacteria, yeast and plants but not in animal cells. Most members of this family use the energy derived from ATP to unfold substrates and to translocate them for degradation to a protease subunit that can be associated with them.²⁰  Other chaperone machines such as the Hsp70 and J proteins cooperate with the protein disaggregation activity of Hsp100 proteins.²¹ 

This molecular chaperone is highly abundant in the cytosol of bacterial and eukaryotic cells under physiological conditions, and can be further up-regulated by cellular stress. The Hsp90 family in mammalian cells is composed of four major homologs: Hsp90α (inducible form) and Hsp90β (constitutive form) are cytosolic isoforms; the 94 kDa glucose-regulated protein (GRP94) is localized in the endoplasmic reticulum,²²  and TRAP1 resides in the mitochondrial matrix.²³  The cytosolic forms of Hsp90 bind proteins in a metastable native-like state, which they may have acquired with the help of other chaperone machines. Hsp90 acts with a group of co-chaperones that modulate its client recognition, ATPase cycle and chaperone function. Due to the nature of its clients, its proteostatic functions affect several essential cellular activities such as development, transcription, cell cycle, intracellular signaling, apoptosis, protein degradation and innate and adaptive immunity.^24–29 

Hsp70 proteins are highly conserved, present both as constitutively expressed and stress-inducible cytosolic isoforms, and isoforms localized to other cellular compartments such as the endoplasmic reticulum and mitochondria.³⁰  They are important for de novo folding, but also for other functions, including protein trafficking, unfolding and degradation of misfolded proteins. More generally speaking, Hsp70s together with a group of essential cofactors, the J proteins of the Hsp40 family and the large nucleotide-exchange factors (NEFs), are involved in ATP-regulated binding and release of non-native proteins including nascent polypeptide chains.³¹  Binding and release by Hsp70 is achieved through the allosteric coupling of a conserved N-terminal ATPase domain with a separate substrate binding domain. Hydrolysis of ATP to ADP is strongly accelerated by Hsp40 proteins. Hsp40s also interact directly with unfolded polypeptides and can recruit Hsp70 to protein substrates.³²  Binding of Hsp70 to non-native substrates impedes aggregation by rapidly protecting exposed hydrophobic segments thereby reducing the presence of species tending to aggregate. Recently, it was described that the nucleotide exchange factor Hsp110, an Hsp70 homolog in eukaryotes, cooperates with the conventional Hsp70-Hsp40 machinery to disaggregate and to refold aggregated proteins.³³  Hsp70 machines act in concert with yet other molecular chaperone machines; for example, they often act upstream of chaperonins³⁴  and the Hsp90 chaperone machine.³⁵  Hsp70s are especially important under stress by preventing the aggregation of unfolded proteins and by refolding aggregated proteins.³⁶  This and different features described in Table 1.3 confer the capacity to Hsp70 to act as a survival factor, which is particularly relevant to provide resistance to apoptosis^37,38  and autophagy³⁹  in cancer cells.

Chaperonins are ring-shaped multi-subunit chaperones that encapsulate non-native proteins in an ATP-dependent manner. In bacteria, the GroE machinery consists of a multi-subunit structure of the two proteins GroEL and GroES. The closely related proteins of eukaryotic mitochondria are called Hsp60 and Hsp10, respectively. The non-native protein is trapped in the cavity of GroEL, which becomes a highly hydrophilic environment with a net negative charge where the protein is free to fold after the open structure is capped by GroES binding.²  The eukaryotic cytosolic chaperonin TRiC is independent of Hsp10. Instead, it contains finger-like projections in its apical domain, which act as a lid and replace the Hsp10/GroES functions.² 

These are ATP-independent molecular chaperones that interact with large numbers of partially folded target proteins to prevent their aggregation upon stress. They act as a depository for unfolded proteins, which will later be refolded by other chaperone machines like Hsp70 and Hsp100. In their native state, they form ring-like oligomers of 12–32 subunits with internal spaces in symmetrically blocked dimeric subunits.⁴⁰ 

The following are molecular chaperones that are not part of the major chaperone families, but are nevertheless worth thinking about as potential drug targets.

• ADCK3 (Chaperone activity of bc1 complex-like, mitochondrial): this chaperone-like protein kinase is essential for the proper conformation and functioning of protein complexes in the respiratory chain. Its absence produces a decrease of the Coenzyme Q10 (CoQ10), and increased ROS production and oxidation of lipids and proteins.⁴²  ADCK3 mutations were detected in patients with cerebellar ataxia.^43,44 

• AHSP (Alpha-hemoglobin-stabilizing protein): a molecular chaperone that is important to prevent the harmful aggregation of free α-hemoglobin during normal erythroid cell development and in β-thalassemic erythroid precursor cells.⁴⁵  Gene knockout studies in mice confirmed that AHSP is required for normal erythropoiesis. AHSP knockout mice exhibit anemia, decreased hematocrit and high levels of ROS, consistent with the presence of unstable α-globin.⁴⁶ 

• ANKRD13 (Ankyrin repeat domain-containing protein 13): acts as a molecular chaperone for G protein-coupled receptors, controlling their biogenesis and exit from the endoplasmic reticulum.⁴⁷  It also regulates the rapid internalization of ligand-activated EGFR.⁴⁸ 

• Histones chaperones: a group of proteins interacting with histones from their synthesis, during import into the nucleus, and for association with target DNA throughout DNA replication, repair or transcription.⁴⁹ 

• CLU (clusterin): functions as extracellular chaperone that prevents aggregation of non-native proteins. Maintains partially unfolded proteins in a state appropriate for subsequent refolding by other chaperones. In Alzheimer’s disease, CLU contributes to limit amyloidogenic Aβ species misfolding and facilitates their clearance from the extracellular space.⁵⁰ 

• TOR1A (Dystonia 1 protein): TorsinA is a member of the AAA-ATPase family of molecular chaperones, assisting in the proper folding of secreted and/or membrane proteins. Defects in TOR1A are the cause of dystonia type 1.⁵¹ 

• HYPK (Huntingtin-interacting protein K): it has a molecular chaperone activity that prevents polyglutamine (polyQ) aggregation of the huntingtin protein.⁵² 

Molecular chaperones are sensitive hubs of the proteostasis network. As a consequence, genetic alterations of the expression or sequence of members of this network may cause disease. Table 1.1 summarizes the disorders associated with mutations of genes encoding members of the molecular chaperones families.

In addition, the phenotypes of the genetic ablation of different members of the molecular chaperone families in mouse models further emphasize the importance of these genes at the organismic level. Table 1.2 displays a complete compilation of the annotated data concerning the knockout of several molecular chaperones and their co-chaperones present in the PhenomicDB.⁴¹ 

Tables 1.1 and 1.2 clearly show that genetic polymorphisms, mutations or the complete ablation of a member of any molecular chaperone machine have an impact on organisms, often with catastrophic consequences. The more or less severely perturbed proteostasis can affect multiple organs and physiological processes, including aging. In accordance with their key hub function, knockout mouse models are often embryonically lethal or have severe complications in development.

It is noteworthy that the presence (or disproportionate presence) of molecular chaperones is not always beneficial. Even when their participation in the protection against apoptosis could be favorable in the context of some disorders, it can unfortunately also promote the initiation and progression of cancer. In addition, several “clients” of these chaperone machines are oncoproteins. These relationships are summarized in Table 1.3.

This overview provides valuable information in view of the use of inhibitors of selected molecular chaperones for therapeutic interventions, for example against cancer or neurodegenerative diseases. It highlights the huge therapeutic potential, but it also gives a flavor of the extremes of the adverse effects that may have to be expected. To emphasize this point further, we have attempted to model the effect of removing, i.e. inhibiting, Hsp90 as one of the key hubs of proteostasis. Figure 1.1 illustrates the dramatic impact that such a treatment can have on part of the proteome.

The next chapters present various strategies aimed at modulating chaperone activities. In the interest of space, we decided to focus on a few major chaperone machines. This is not a serious limitation since a limited number of molecular chaperones have received almost all of the attention of developers. Other molecular chaperones have yet to be explored as drug targets altogether. Most of the “strategies” reported to date are based on small organic molecules that inhibit molecular chaperones, but more diversity in terms of types of molecules, approaches and impact on molecular chaperones can be expected from future efforts. And to the best of our knowledge and despite very high hopes, “molecular chaperone drugs” have yet to make it into the clinic. Nevertheless, the emerging diversity attests to the huge interest that this class of proteins has attracted over the last few years.

The functionality of Hsp90 requires highly complex conformational rearrangements regulated by a large spectrum of co-chaperones in which ATP hydrolysis plays a fundamental role in driving the chaperone machinery. Hsp90 features a conserved GHKL-type (gyrase, Hsp90, Histidine Kinase, MutL) ATPase domain at the N-terminus. Therefore, designing competitive inhibitors that target the N-terminal ATP binding pocket of Hsp90 has been the main strategy to block Hsp90.⁵³  The following paragraphs first present these inhibitors before giving an overview of inhibitors that target other Hsp90 domains and Hsp90 co-chaperones (see Table 1.4).

These types of inhibitors arrest the chaperone cycle by trapping Hsp90 in a conformation reminiscent of the ADP-bound conformation. This leads to an early release of immature client proteins followed by their degradation through the ubiquitin-proteasome pathway. Among the clients of Hsp90, many are oncogenic such as B-Raf, v-Src, HER2, Akt, mutant p53, HIF-1α and Bcr-Abl, which exemplifies the central role of Hsp90 in maintaining the homeostasis of cancer cells. As a result of Hsp90 inhibition and degradation of these and other clients, cancer cells stop proliferating and even undergo apoptosis. A corollary of the inhibition of Hsp90 is the induction of Hsp70 expression through the proteasomal stress response and by derepression of the heat-shock factor 1 (HSF-1).⁵⁴ 

Several classes of N-terminally targeted Hsp90 inhibitors have emerged from both the pharmaceutical industry and the academic world. Based on the natural and prototypical inhibitor geldanamycin, tanespimycin (17-AAG) and 17-DMAG were identified as improved derivatives with less toxicity and better water solubility. However, the reduction of the benzoquinone core by NQO1/DT-diaphorase is required for these inhibitors to exert full efficacy. This prompted the development of IPI-504 (Infinity Pharmaceuticals), a tanespimycin derivative bearing a stabilized hydroquinone ring as an attempt to reduce or to abolish the dependence on the NQO1-mediated reduction. This inhibitor is currently in clinical trials and demonstrates promising potency against non-small-cell lung cancer. However, despite an excellent specificity for Hsp90, many benzoquinone ansamycin-based molecules have intrinsic hepatotoxicity.⁵⁴ 

The discovery of the purine-based inhibitor PU3 brought a new impetus for the development of synthetic inhibitors. PU3 stabilizes the N-terminal ATPase pocket of Hsp90 in a conformation characterized by the formation of an α-helix between Leu107 and Gly114. In comparison with the ADP-bound form, this structural rearrangement creates a secondary hydrophobic binding pocket where appropriate hydrophobic decorations of ligands can be accommodated. Apart from the ansamycin family of inhibitors, most of the subsequently developed inhibitors take advantage of this secondary site.^55,56 

The purine-based inhibitors mimic the adenosine moiety of the natural ligand ATP, which ensures recognition by Hsp90. The more recently reported aminopyri(mi)dines, thienopyrimidines (from Astex Pharmaceuticals, Evotec, Abbott or Vernalis), the azaindole (from Sanofi-Aventis) and the benzamide (SNX-5422, SNX-2112, XL888) derivatives actually share the same or very similar but more cryptic bioisosteres that were inspired by the adenosine moiety.⁵⁶ 

Resorcinol derivatives constitute another important group of the N-terminally targeted Hsp90 inhibitors. Their dihydroxyl benzene group is present in radicicol, a natural antibiotic that binds tightly to Hsp90. The binding mode of the resorcinol motif appears to be very stable and conserved in Hsp90s and therefore provides an excellent “starting point” for further drug design by both structure-based and fragment-based approaches.⁵⁶ 

Gambogic acid

Gambogic acid binds to the N-terminal domain of Hsp90 and impedes the association with Hsp70 and Cdc37. This leads to the degradation of client proteins and up-regulation of Hsp70 and Hsp90.⁵⁷  It induces apoptosis through inactivation of the TNF-α/NFκB pathway. The binding of gambogic acid to the N-terminal domain of Hsp90 is not affected by the presence of geldanamycin, suggesting that the drug may bind to a site distinct from the ATP binding pocket.⁵⁸ 

Shepherdin (Peptide)

Shepherdin is a fragment of a peptide from the Hsp90 client survivin. Initially designed to block its interaction with Hsp90, the peptide was shown to bind to the N-terminal domain of Hsp90 and to disrupt the Hsp90 ATPase activity.⁵⁹  It induces apoptosis in tumor cells following the degradation of Hsp90 clients such as Akt, CDK-4, CDK-6 and survivin without affecting the levels of Hsp70.⁶⁰ 

Sansalvamide A Derivatives (Peptide)

Sansalvamide A is a cyclic pentapeptide that is extracted from a marine fungus of the genus Fusarium sp. The peptides were demonstrated to bind to the junction between the N-terminal and middle domains of Hsp90 and to inhibit the Hsp90 cycle. The mechanism is most likely an allosteric regulation, which is supported by the fact that the peptides preferentially bind to the closed conformation of yeast Hsc82 stabilized by the non-hydrolyzable ATP derivative AMPPNP.⁶¹  Further biological characterization showed that Sansalvamide A derivatives⁶²  induce caspase-dependent apoptosis in cells. Some of them impair the recruitment of clients or co-chaperones including IP6K2, FKBP38, FKBP52 and HOP. Like 17-AAG, these compounds also elicit the up-regulation of Hsp70.⁶³  In contrast, one of the compounds causes also cytotoxic effects and caspase-mediated apoptosis while it does not affect the interaction between the co-chaperone/clients and Hsp90, suggesting that it may be used as a unique chemical tool that inhibits Hsp90 without altering its binding to clients or co-chaperones.⁶² 

CP9

The 2-[6-(trifluoromethyl)pyrimidin-2-yl]thio-acetamide-based compound C9 was discovered as a molecule that disrupts the Hsp90-p23 interaction. It targets the N-terminus of Hsp90 and competes with 17-AAG for binding. The compound induces Hsp90 client degradation in cancer cell lines and selectively impairs the interaction of p23 with the Hsp90α isoform. Further chemical optimization is expected to deliver candidates with higher potency in live animals.⁶⁴ 

Derivatives of Coumarin Antibiotics (Novobiocin)

The coumarin antibiotic novobiocin binds to the C-terminal domain of Hsp90, but nevertheless inhibits the ATPase activity in an allosteric manner. A key feature of this family of molecules is that they seem to induce much less Hsp70 expression. Since the anti-apoptotic effects of Hsp70 are unwanted side effects of N-terminally targeted ATPase inhibitors, there is an interest in further developing those as alternatives. Indeed, certain derivatives of novobiocin such as KU-135 even have improved drug effectiveness and binding to Hsp90 compared to 17-AAG.⁶⁵ 

Celastrol

The proteasomal inhibitor Celastrol disrupts the association of Hsp90-Cdc37. Even though it does not occupy the ATPase pocket of Hsp90, its binding is exclusive to the ADP bound or nucleotide-free form of Hsp90,⁶⁶  suggesting it is an allosteric inhibitor acting through a different site. Celastrol inhibits the ATPase activity of Hsp90 and induces the degradation of client proteins such as Akt and Cdk4. The action is accompanied by the simultaneous increase of Hsp70 expression through the activation of the HSF-1 pathway.⁶⁷  In addition, the quinone methide moiety of Celastrol confers a reactivity towards thiols, which brings about the direct modification of cysteine residues in many proteins including Hsp90, p23 and Cdc37. Presumably, thiol oxidation of cysteines accounts at least in part for the ability of Celastrol to inhibit the Hsp90 activity.⁶⁸ 

Withaferin A

Withaferin A is a withanolide from the plant Withania somnifera. This steroidal lactone binds to the C-terminal domain of Hsp90 and inhibits its ATPase activity. Like Celastrol it disrupts the formation of the Hsp90-Cdc37 complex, and induces the degradation of Hsp90 clients (Akt, Cdk4, GR) and up-regulation of Hsp70 through the proteasomal stress response pathway.⁶⁹  Withaferin A shares the quinone methide moiety with Celastrol and therefore also induces direct thiol oxidation of cysteine residues on Hsp90, which causes the aggregation of Hsp90. Tubocapsenolide A inhibits Hsp90 using the same mechanism.⁷⁰ 

Cisplatin and LA-12

In addition to binding to DNA, cisplatin has been reported to target the C-terminal domain of Hsp90 and to disrupt selectively the Hsp90 chaperoning function for transcriptional factor clients such as the androgen or glucocorticoid receptors. Similarly to the coumarin antibiotics, cisplatin does not induce Hsp70 expression.⁷¹ 

An optimized derivative of the platinum-based complex, LA-12, exhibits enhanced binding affinity for Hsp90, presumably due to the increased hydrophobicity, which favors the binding to the C-terminal domain. Compared to cisplatin, this compound also stimulates the degradation of more Hsp90 clients, such as cyclin D1 and the estrogen receptor.⁷² 

Molybdate

Molybdate stabilizes a closed conformation of Hsp90 that resembles the ATP-bound state. It may do this by replacing the cleaved γ-phosphate of ATP. It has been used extensively to stabilize Hsp90-client interactions. However, its poor cell permeability essentially limits its use to biochemical experiments.⁷³ 

EGCG

(−)-Epigallocatechin-3-gallate (EGCG) is a polyphenolic cathechin found in green tea. Biochemical assays showed that the compound binds to the C-terminal domain of Hsp90 and induces the degradation of various client proteins such as Akt, Cdk4, Raf-1, HER2 and pERK.⁷⁴  EGCG also weakens the formation of Hsp90-p23 and Hsp90-Hsc70 complexes and suppresses Hsp70 expression induced by proteasomal stress.⁷⁵ 

Cdc37 is a co-chaperone that regulates the ATPase cycle of Hsp90. It appears to serve as an adapter that brings a wide spectrum of client kinases to Hsp90 and therefore acts as an important modulator of many signaling pathways. In comparison with Aha1 that promotes the Hsp90 ATPase activity, Cdc37 slows down the ATP turnover through an interaction with the lid segment of the N-terminal domain of Hsp90.^26,76 

Drugs (Celastrol, Withaferin A)

In addition to affecting Hsp90 directly, Celastrol was shown to target Cdc37 by covalently modifying the cysteine residues on the kinase-binding domain of Cdc37, which alters the global conformation of Cdc37 and leads to a disruption of the Hsp90-Cdc37 interaction.⁷⁷  The same appears to be true for Withaferin A.^68,69  Additional inhibitors that target Cdc37 directly and selectively would clearly constitute formidable tools and potential therapeutics since they might only affect kinases that are Hsp90 clients.

Cdc37 siRNA

Silencing Cdc37 by RNA interference results in a significant degradation of Hsp90 client kinases through the ubiquitin-proteasome pathway. Cdc37 silencing also sensitizes cancer cells against Hsp90 inhibitors. In addition, Cdc37 silencing does not derepress HSF-1 and therefore does not induce up-regulation of Hsp70.⁷⁶ 

Phosphorylation/Dephosphorylation of Cdc37

As the phosphorylation of specific sites is required for Cdc37 function, inhibition of casein kinase II (with TBB), which decreases the phosphorylation of Cdc37, can be envisaged as an alternative to repress Cdc37.⁷⁸  Over-expression of the phosphatase PP5 results in the same outcome.⁷⁹ 

By bridging Hsp70 and Hsp90, the Hsp70-Hsp90 Organizing Protein (HOP; also known as Sti1) helps to pass clients from Hsp70 to Hsp90. It contains several tetratricopeptide (TPR) domains. The TPR1 domain interacts with Hsp70 whereas the TPR2A domain enables the recognition of the C-terminal MEEVD motif of Hsp90 and the TPR2B domain interacts with both the N-terminal and middle domains of Hsp90. HOP prevents the N-terminal dimerization of Hsp90 and thereby inhibits the ATPase activity.⁸⁰ 

C9

Using an in vitro high-throughput screen, a family of 7-azapteridine compounds was first identified as lead compounds that bind to the TPR2A domain of HOP and block its interaction with the MEEVD motif of Hsp90.⁸¹  The derivative C9 was shown to be effective without inducing the HSF-1-mediated up-regulation of Hsp70. Interestingly, the cell death induced by C9 does not result from the activation of the caspase 3/7-dependent apoptotic pathway, which indicates a distinct mechanism in comparison with that of 17-AAG. C9 also suppresses the up-regulation of Hsp70 if applied in combination with 17-AAG or NVP-AUY922, although not with PU-H71, possibly due to the overwhelming inhibitory effect of PU-H71.⁸² 

Hybrid Antp-TPR Peptide

A hybrid TPR peptide was designed based on a structurally engineered TPR2A derivative that was rendered cell-permeable by fusion to a cell targeting peptide from the Antennapedia homeodomain protein. This hybrid peptide selectively induces apoptosis in cancer cell-lines and provokes Hsp90 client degradation without activating HSF-1 mediated Hsp70 expression.⁸³ 

HOP siRNA

The application of siRNA against HOP expression in pancreatic cancer cells has been shown to decrease invasiveness through MMP-2 down-regulation. Decreased expression levels of Hsp90 clients such as HER2, Bcr-Abl, c-MET and v-Src were also observed.⁸⁴ 

Celastrol

It was reported that Celastrol can modify the cysteine residues of p23 through direct thiol oxidation. This alters the three-dimensional structure of p23, which drives p23 into amyloid-like fibrils.⁸⁵ 

Genetic Interference

Deletion or silencing of p23 (Sba1 in yeast) results in a hypersensitivity to Hsp90 inhibitors (17-AAG) in the context of budding yeast and mammalian cells.⁸⁶ 

Aha1 is the only Hsp90 co-chaperone known to activate the Hsp90 ATPase. It does so by interacting with the middle and N-terminal domains of the Hsp90 dimer in trans and favors the dimerization of Hsp90.⁸⁷ 

Genetic Interference

The knockdown of Aha1 expression by RNA interference leads to a higher sensitivity to Hsp90 inhibitors despite not affecting the levels of Hsp90 clients. However, a decreased phosphorylation of MEK1/2 and ERK1/2 was observed, suggesting that Aha1 may be more involved in regulating the activation of these client proteins rather than their stabilization.⁸⁸  Surprisingly, over-expression of Aha1 results in a lower rate of Hsp90-dependent refolding of denatured luciferase.⁸⁹  Moreover, mutations in the Hsp90 ATPase lid domain that result in an increased affinity for Aha1 are associated with an increased resistance to Hsp90 inhibitors, indicating a protective role of Aha1 through regulation of the ATPase activity of Hsp90.⁹⁰ 

The structure of Hsp70 can be functionally divided into two domains: an N-terminal nucleotide-binding domain and a C-terminal substrate-binding domain. As for Hsp90, the functionality of Hsp70 relies on the binding and hydrolysis of ATP, which drives conformational changes. The ATPase activity is stimulated by Hsp70 co-chaperones of the Hsp40/DnaJ family through an interaction that involves the J-domain of Hsp40. Other co-chaperones such as the nucleotide exchange factors Hsp110, Bag-1 or HspBP1 facilitate the release of ADP.⁹¹ 

Although both Hsp90 and Hsp70 are ATPase-dependent molecular chaperones, intrinsic tight ATP binding, more hydrophilic ligand interaction and flexible ligand-pocket accommodation present obstacles to the design of competitive small-molecule inhibitors targeting the ATP-binding pocket of Hsp70 compared with the situation for Hsp90.⁵³ 

Nevertheless, given the significance of Hsp70 as a potential target for cancer therapy, there are many continued efforts for the development of Hsp70 targeting pharmacological agents in the form of small molecules, peptides or peptide aptamers. Among them, some directly compete with ATP for binding to the Hsp70 ATP binding pocket, some bind to the substrate-binding domain and inhibit the recruitment of client proteins, whereas some affect the association of Hsp40 with Hsp70, which impairs the activation of the Hsp70 ATPase. In the following paragraphs, we mention some examples of molecules that target different regions of Hsp70 (Table 1.5).

VER-155008

A small synthetic molecule mimicking adenosine, VER-155008, binds to the Hsp70 ATPase site. The compound also displays affinity for Hsc70 and the isoform GRP78 (Bip), which resides in the endoplasmic reticulum (ER). It induces the degradation of some Hsp90 client proteins such as Raf-1 and HER2, inhibits the proliferation and induces caspase 3/7-mediated apoptosis of several tumor cell lines.⁹² 

Apoptozole

Apoptozole (Az) was originally identified in an imidazole-based library as an agent that induces apoptosis in P19 and A549 lung cancer cells. Biochemical studies demonstrate that Az is able to bind to and inhibit the Hsp70/Hsc70 ATPase domain. Competitive binding between Az and ATP suggests that the compound may directly fit into the ATP-binding pocket. Moreover, Az has been shown to disrupt the association between the CFTR mutant ΔF508 and Hsp70, which promotes its degradation through the ubiquitin-proteasome pathway.⁹³ 

MKT-007

MKT-007 (1-ethyl-2-[[3-ethyl-5-(3-methylbenzothiazolin-2-yliden)]-4-oxothiazoli-din-2-ylidenemethyl] pyridinium chloride) is known to bind in the vicinity of the ATP-binding pocket of Hsc70. It favors the ADP-bound conformation of Hsc70, suggesting that the compound is an allosteric inhibitor of Hsc70. It has also been shown that MKT-007 promotes the ubiquitination and degradation of over-expressed tau.⁹⁴ 

AdaSGC

The water-soluble glycosphingolipid mimic adamantly-sulfogalactosylceramide (AdaSGC) binds to the Hsc70 ATPase domain and disrupts the Hsp40-stimulated ATPase activity, causing a decrease in client protein loading. It has been shown to promote the clearance of over-expressed CFTR ∆F508 through the ER-associated-degradation pathway.⁹⁵ 

Peptide Aptamer A17

Yeast two-hybrid experiments led to the discovery of the peptide aptamers A8 and A17, which bind to the Hsp70 peptide-binding and nucleotide-binding domains (but not to those of the Hsc70s), respectively. Both aptamers, when expressed in mouse melanoma B16F10 cells, suppress tumor progression and synergize with the effects of cisplatin. The synthetic peptide P17, which contains only the variable region of aptamer A17, sensitizes tumor cells to cisplatin and induces apoptosis. Although it binds the nucleotide-binding domain, the exact binding site still remains to be elucidated.⁹⁶ 

PES/Pifithrin-µ

2-phenylethyenesulfonamide (PES) is a small molecule that binds to both Hsc70 and Hsp70. It was first discovered as a molecule that impairs the mitochondrial localization of p53.⁹⁷  It was shown to induce apoptosis via activation of caspase-3 in leukemic cells, which also involves the degradation of client proteins such as Akt and ERK1/2. It thereby overcomes the protective effect afforded by the up-regulation of Hsp70 because of the proteasomal stress. This also contributes to enhanced anti-proliferative effects in combinatorial treatments with the Hsp90 inhibitor 17-AAG or with the histone deacetylase inhibitor SAHA.⁹⁸ 

ADD70

ADD70 is an artificially engineered protein derived from the flavoprotein apoptosis-inducing factor (AIF). It only contains the Hsp70-interacting region and binds to the substrate-binding domain of the inducible Hsp70.⁹⁹  Tumor cells expressing ADD70 demonstrate increased sensitivity to a treatment with cisplatin or the Hsp90 inhibitor 17-AAG. Extended studies indicated that the induction of antitumorigenic immune responses may account for the cytotoxicity of ADD70. This is supported by the observation that the synergy of ADD70 and cisplatin disappears in immunodeficient animals.¹⁰⁰ 

D-peptides

D-peptides are derived from the N-terminal fragment of rhodanese, which was known to bind to the bacterial DnaJ. The peptides RI1-17 and RI1-10 bind to DnaJ and disrupt the formation of the bacterial DnaJ/DnaK/GrpE chaperone complex.¹⁰¹ 

Dihydropyrimidine Family (NSC6340668-R/1, MAL3-101, 115-7c, 116-9e)

Based on the prototype molecule 15-deoxyspergualin (DSG) that demonstrates the ability to modulate the Hsc70 ATPase activity, NSC6340668-R/1 and MAL3-101 were identified as small molecules that inhibit Hsp70 by preventing the Hsp40 stimulation of the ATPase activity.¹⁰²  MAL3-101 was shown to be effective against multiple myeloma cell lines by inducing caspase-3-mediated apoptosis. It displays strong synergy with the Hsp90 inhibitor 17-AAG or with the proteasomal inhibitor MG-132.¹⁰³  In addition, other molecules in the family, such as MAL3-38, MAL3-90 or 115-7c, exert the opposite effect by stimulating the Hsp70 ATPase activity. Among them, 115-7c was demonstrated to compensate partially for the loss-of-function phenotype of yeast lacking the Hsp40 protein Ydj1. It preferentially binds to the bacterial DnaJ-DnaK complex rather than DnaK alone, suggesting that the compound regulates the DnaK ATPase activity through an allosteric mechanism. Compared to 115-7c, the derivative 116-9e toggles the effect back to a disruption of the DnaJ-DnaK interaction and inhibition of the DnaK ATPase activity. This illustrates the flexibility of the dihydropyrimidine family of compounds as allosteric regulators of the DnaK-DnaJ-GrpE chaperone complex.¹⁰⁴ 

CHIP Over-expression

The constitutive heat-shock cognate 70 (Hsc70)-interacting protein CHIP binds to Hsp70/Hsc70 through its TPR domain. It also possesses a U-box domain, which carries its E3 ubiquitin ligase activity. CHIP is responsible for the ubiquitin-mediated degradation of many Hsp70/Hsp90 client proteins. Over-expression of CHIP enhances client protein degradation in vivo.¹⁰⁵  Furthermore, it targets the molecular chaperones Hsp70 and Hsp90 themselves.¹⁰⁶ 

Hsp27 increases cell survival and enhances tumor migration and invasiveness. It is highly expressed in various cancers such as breast, ovarian or prostate cancer. There has been an increasing awareness that Hsp27 may be an interesting anticancer target.¹⁰⁷ 

KRIBB3

KRIBB3 was identified from a pool of compounds that inhibit tumor cell migration. The compound binds to Hsp27 with high affinity and affects the protein kinase C-dependent phosphorylation of Hsp27, which is necessary for its activation.¹⁰⁸  Moreover, KRIBB3 was demonstrated to induce cancer cell apoptosis by activating the mitotic spindle checkpoint because of the inhibition of tubulin polymerization.¹⁰⁹ 

Peptide Aptamers PA11 and PA50

Identified with a yeast two-hybrid screen, aptamers PA11 and PA50 have anti-proliferative and apoptosis-inducing activities in HeLa cells, while the overall expression levels of Hsp60, Hsp70 and Hsp90 were not affected. In parallel, biochemical studies showed that PA50 affects Hsp27 dimerization while PA11 impairs the oligomerization of Hsp27, and both aptamers modified the phosphorylation state of Hsp27. Expressing PA50 and PA11 in cells appears to be more effective than an shRNA-mediated knockdown of Hsp27.¹¹⁰ 

OGX-042

OGX-042 (OncoGeneX Pharmaceuticals Inc.) is an antisense oligonucleotide (ASO) that blocks the translation initiation site of the Hsp27 mRNA. It has a 2′-O-methoxy-ethyl (MOE) backbone, which enhances its protection against nucleases and increases its half-life in vivo. OGX-042 inhibits cell growth, induces apoptosis and sensitizes cells to the chemotherapeutic agent gemcitabine.¹¹¹ 

The expression of clusterin is known to be associated with higher grade, post-treatment stress and poor outcome in many cancers.¹⁰⁷ 

OGX-011 (Custirsen)

Like OGX-042, OGX-011 is an antisense oligonucleotide that is designed to block the expression of human clusterin. The depletion of clusterin leads to an enhanced sensitivity to chemo- and radiation therapy. OGX-011 has demonstrated promising effects to postpone the recurrence of castration-resistant prostate cancer.¹¹²  The combination of OGX-011 with the Hsp90 inhibitors 17-AAG or PF-04929113 potentiates the antitumor potency in xenografts by limiting the protective heat-shock response induced by the Hsp90 inhibitors.¹¹³ 

In response to proteotoxic stress, heat-shock or pharmacological agents such as Hsp90 or proteasome inhibitors, HSF-1 initiates the transcription of several major protective proteins including Hsp70, Hsp90 and Hsp27. Since HSF-1 is required by cancer cells to maintain proteostasis, targeting the HSF-1 pathway represents a promising strategy to suppress the transformation and malignant progression of cancer cells and to potentiate other drugs that elicit stress responses. The mechanisms of activation of HSF-1 are incompletely characterized. Moreover, there is no druggable HSF-1 domain that is well defined. This makes the development of potent inhibitors very challenging.¹¹⁴ 

QC

Some known antimalarial agents such as aminoacridine quinacrine (QC) are known to inhibit the HSF-1-mediated Hsp70 expression. The combination of QC with 17-AAG has shown enhanced potency in inducing caspase-mediated apoptosis. The pharmacological effects of QC on HSF-1 are fairly broad.¹¹⁴  Despite its low potency and poor specificity, which prevents its clinical development, it serves as an encouraging model.¹¹⁵ 

KNK437

KNK437 (N-formyl-3,4-methylenedioxy-benzylidene-gamma-butyrolactam) suppresses the HSF-1-dependent transcription of Hsp70 and Hsp27.¹¹⁶  The exact mechanism still remains to be established. As a result of lowering Hsp70 and Hsp27 levels and thus the proteasomal response, KNK437 potentiates the inhibitory effects of Hsp90 inhibitors in various cell lines.^117,118 

Triptolide

The diterpene triepoxide Triptolide is derived from the Chinese medicinal plant Triptergium wilfordii. It can effectively suppress the heat-shock response initiated by HSF-1 at nanomolar doses through direct interference with the transactivation domain of HSF-1, possibly at a post-transcriptional level.¹¹⁹ 

NZ28 and Emunin

NZ28 and Emunin were identified from 20,000 compounds by a screen for inhibitors of the induction of heat-shock proteins. Despite the analogy to emetine, which blocks translation, the compounds do not affect the overall level of transcription, translation or protein degradation. Instead they specifically suppress the up-regulation of Hsp70 and Hsp27 in response to Hsp90 or proteasome inhibitors. The inhibitory effect may be at the post-transcriptional level since synthesis and degradation of Hsp70 mRNA were unaffected by the treatment. Moreover, the molecules sensitize cancer cells to Hsp90 and proteasome inhibitors and induce Hsp90 client degradation.¹²⁰ 

This avenue is pursued because HSF-1 activation may be beneficial to treat neurodegenerative diseases that are associated with the accumulation of aberrant proteins. An appropriate induction of the heat-shock response by the activation of HSF-1 enhances cytoprotection against detrimental consequences of misfolded or aggregated stressors.¹²¹ 

Hsp90 Inhibitors

The induction of Hsp70 through HSF-1 activation in response to Hsp90 inhibitors is well established. It is thought that the release of the inhibitory Hsp90 complex from HSF-1 contributes to its activation. In keeping with the beneficial effects of Hsp70 induction, geldanamycin is effective against the formation of α-synuclein oligomers in cells. Several benzamide derivatives (e.g. SNX-0723), which have the added advantage of being orally available and passing through the blood-brain barrier, have been reported to prevent α-synuclein oligomerization through HSF-1-stimulated expression of Hsp70.¹²² 

Celastrol

As described above in the section on Hsp90 inhibitors, Celastrol inhibits Hsp90 through various mechanisms and induces the HSF-1 pathway.^85,123  However, its intrinsic toxicity, in part to due to pleiotropic thiol oxidation of proteins, makes it difficult to develop it further for clinical applications.

Geranylgeranylacetone

Geranylgeranylacetone (GGA) has been widely used as an anti-ulcer drug. It is also known to induce the expression of Hsp70. Biochemical studies demonstrate that GGA competes with HSF-1 for the binding to the substrate-binding domain of Hsp70. Upon dissociating from Hsp70, HSF-1 is activated.¹²⁴  Moreover, GGA seems to trigger ER stress, which leads to the up-regulation of GRP78. GGA elicits both pro-apoptotic and anti-apoptotic unfolded protein responses (UPR) through modulating the ATF6-CHOP, ATF6-GRP78 and IRE1-XBP1 pathways.¹²⁵  The cytoprotective role of GGA may result from its induction of thioredoxin (TRX) expression, which enhances the ability to reduce the oxidized thiol groups of proteins and to protect cells against oxidative stresses.¹²⁶ 

HSF1A

HSF1A is a benzyl pyrazole identified with a yeast-based high-throughput screen. The compound activates human HSF-1; it increases the expression of Hsp70 by facilitating the nuclear import and phosphorylation of HSF-1. HSF1A confers protection against cytotoxicity caused by polyQ aggregation. Interestingly, the TRiC/CCT chaperone complex was found to co-purify with biotinylated HSF1A, which is intriguing considering that TRiC/CCT acts as a negative regulator of polyQ aggregation.¹²⁷ 

Bimoclomol/Arimoclomol

Bimoclomol alone does not activate HSF-1. Instead, it potentiates the expression of Hsps under a pre-existing stress condition. The mechanism seems to involve direct binding of bimoclomol to HSF-1, which leads to a prolonged HSF-1/DNA interaction accompanied by a mild increase of HSF-1 phosphorylation.¹²⁸  As co-inducers of HSF-1-mediated chaperone expression, the drugs may be well adapted for the treatment of neurodegenerative diseases in which a strengthened molecular chaperone system increases the tolerance to misfolded stressors. The derivative arimoclomol (BRX-220) is in a Phase II/III clinical trial for the treatment of patients with superoxide dismutase (SOD1) positive amyotrophic lateral sclerosis (ALS).¹²⁹ 

Riluzole

Riluzole is an FDA-approved drug for the treatment of ALS. It has recently been shown that the drug inhibits chaperone-mediated autophagy of HSF-1. The increased availability of HSF-1 boosts heat-shock responses and ultimately improves cytoprotection against stresses.¹³⁰ 

NSAIDs

Non-steroidal anti-inflammatory drugs (NSAIDs) have been known for a long time to promote HSF-1 activity.¹³¹  However, only some (sulindac, cyclopentenone derivatives) are robust enough to elicit the HSF-1-induced Hsp70 expression by themselves.^132,133  Others (sodium salicylate, indomethacin) act as co-inducers, which facilitate Hsp70 expression when an additional stress pre-exists.^131,134  This differential ability may stem from the fact that while NSAIDs generally promote the binding of HSF-1 to the promoter of Hsp70, only some of the agents are capable of triggering its transactivation.

Small-molecule Proteostasis Regulators

With a cell-based high-throughput screen of over 900,000 compounds, the group of Morimoto¹³⁵  discovered several families of small-molecule proteostasis regulators that activate the heat-shock response. Some were demonstrated to promote the recruitment of HSF-1 to target promoters. Furthermore, they were shown to induce a whole panel of responses that are hallmarks of an activated stress response aimed at reestablishing proteostasis. One compound A1 was found to induce the degradation of Hsp90 client kinases without affecting Hsp90 activity in general, indicating that it may specifically affect the Cdc37-Hsp90 interaction.¹³⁵