Figure 56-1 Hsp90 function is required for the establishment and maintenance of each of the eight hallmarks of cancer. Importantly, Hsp90 function is also critical for cancer cells to survive the genetic instability on which acquisition of the eight hallmarks depends, and the environmental stresses to which they are frequently subjected.
Figure 56-2 Hsp90 inhibitors in clinical development. See text for further details.
Figure 56-3 Co-chaperones and posttranslational modifications affect Hsp90 function. In eukaryotes, Hsp90 activity requires the contribution of numerous co-chaperones, each with a specific function. Acetylation, phosphorylation, and nitrosylation of specific residues on Hsp90 affect its interaction with client proteins and co-chaperones.
Several recent, excellently detailed reviews of the mechanics of Hsp90 function are in the scientific literature. 11,13,14,26–28 Hsp90 is a conformationally flexible ATPase that associates with a distinct set of co-chaperones depending on ATP binding to an amino-terminal purine binding pocket. Identification of this pocket as the GA binding site led Chiosis and colleagues to design a series of highly potent purine scaffold Hsp90 inhibitors with markedly improved drug-like properties. 29–32 Workman and colleagues used a high-throughput screen based on inhibition of Hsp90 ATPase activity to identify 3,4-diarylpyrazoles as a novel class of Hsp90 inhibitors. 33,34 Other groups have developed and validated a number of structurally distinct inhibitors of this purine pocket. 35 As of early 2012, more than 10 Hsp90 inhibitors had reached clinical trial (see Figure 56-2).
Hsp90 ATPase Activity and Chaperone Function
A model of Hsp90 function has emerged in which ATP binding to the amino-terminal pocket initiates a series of conformational changes that endow Hsp90 with ATPase activity. This process involves participation of a number of co-chaperones that interact with Hsp90 to form a “super-chaperone machine.” 36 Certain co-chaperones play specific roles in this dynamic process (Figure 56-3 ). ATP hydrolysis completes the chaperone cycle, at which point the process (which is frequently iterative) can begin again. Many Hsp90 client proteins first associate with an Hsp70/Hsp40 chaperone complex. 37 This assemblage associates with Hsp90 via p60Hop, an Hsp90/Hsp70 interacting protein. At this point, when the client protein is being loaded on Hsp90, the chaperone is thought to be in an apo (nucleotide-free) conformation. ATP binding alters Hsp90 conformation, causing it to release p60Hop and the Hsp70/Hsp40 complex, and to recruit another set of co-chaperones, including p23 and certain immunophilins. In the case of kinase client proteins, the co-chaperone p50Cdc37 usually delivers the client to Hsp90. Although ATP hydrolysis is essential for chaperone activity, the ATPase activity of Hsp90 is very weak. Association of the co-chaperone Aha1 is necessary to increase Hsp90 ATPase activity sufficiently to drive the chaperone cycle forward. 38–41 In higher eukaryotes, the ordered association/dissociation of client proteins and co-chaperones is also affected by a series of sequential phosphorylation events. 42 Other posttranslational modifications also affect Hsp90 chaperone activity or client binding (see Figure 56-3). 43–45
Hsp90 Inhibitors Target Client Proteins to the Proteasome
A highly orchestrated and tightly regulated process drives the ATP-dependent Hsp90 super-chaperone machine. Hsp90 inhibitors block the chaperone cycle by displacing ATP from the amino-terminal purine pocket, inducing client protein ubiquitination and proteasome-mediated degradation (Figure 56-4 ). Although the mechanics of this process are not well understood, they appear to involve a “handing-off” of the client from inhibitor-bound Hsp90 to an Hsp70-ubiquitin ligase complex. 46 E3 ubiquitin ligases that have been implicated in mediating Hsp90 inhibitor-induced client protein ubiquitination in mammals include CHIP and Cullin 5. 47,48 Importantly, even if the proteasome is inhibited, client proteins are not rescued from Hsp90 inhibition, but instead accumulate in a misfolded, inactive state in detergent-insoluble subcellular complexes. 49 A selection of several oncogenic Hsp90 clients (in addition to HER2 and EML4-ALK, mentioned earlier) is shown in Table 56-1 . The reader is directed to the references therein for more detailed information.
Figure 56-4 The Hsp90 chaperone machine (A) Client proteins associate weakly with an Hsp90 dimer in the absence of ATP. On ATP binding to an amino-terminal pocket in the chaperone, the N-lobes of each Hsp90 monomer transiently dimerize, resulting in tight binding of the client protein to Hsp90 and in acquisition of ATPase activity. On ATP hydrolysis, stimulated by various co-chaperones, the N-lobes of Hsp90 dissociate, releasing the now-folded client protein. (B) Geldanamycin (GA) and other N-terminal Hsp90 inhibitors block ATP binding to Hsp90, preventing dimerization and maintaining Hsp90 in a conformation that weakly associates with client protein. In the absence of ATP binding, the client protein dissociates from Hsp90, becomes polyubiquitinated by chaperone-dependent E3 enzymes, and is ultimately degraded by the proteasome.
Hsp90 Inhibitors May Prevent Oncogenic Switching
Recent studies have identified development of resistance to tyrosine kinase inhibition (TKI) as a significant roadblock to effective targeted therapy. One mechanism of resistance recently appreciated involves “oncogene switching,” or the reactivation of signaling pathways by one or more redundant upstream activators. In breast cancer models, ErbB TKIs such as gefitinib have been shown to lose the ability to modulate ErbB-driven signaling pathways over time, even though ErbB inhibition is maintained. 51 Similarly, ErbB kinase activation has been reported to confer resistance to MET TKIs in MET oncogene-addicted gastric cancer cells. 52 This model of “oncogene plasticity” does not rely on kinase mutation or on drastic changes in gene expression but merely on the inherent redundancy of biological systems, and it may explain some of the disappointing results seen in previous TKI clinical trials. Using several models, we and others have found that Hsp90 inhibition prevents oncogene switching and is itself a targeted therapy that is not prone to this phenomenon. 53–55 Although Hsp90 inhibitors may have single-agent activity in cancers that are strongly dependent on an Hsp90 client (e.g., HER2 or ALK), these drugs may find a broader role in combination with TKIs that target Hsp90 clients.
Table 56-1
List of Selected Hsp90 Clients Involved in Cancer
| Cancer | Hsp90 Client | References |
| Anaplastic large cell lymphomas | NMP-ALK | 103, 104 |
| Acute myeloid leukemia | FLT3 | 105, 106 |
| Chronic myelogenous leukemia | BCR-ABL | 107, 108 |
| Human mastocytosis Gastrointestinal stromal tumors |
KIT | 109 |
| Melanomas | B-RAF EGFR |
89, 110–113 |
| Prostate cancer | Androgen receptor | 114–116 |
| Clear cell renal cell carcinoma (ccRCC) Nonhereditary sporadic ccRCC |
HIF1-α | 117, 118 |
| Glioblastoma | VEGF | 119 |
| Leukemia | BCL2 APAF |
120 |
| Small-cell lung cancer | MET | 121–124 |
| Multiple endocrine neoplasia type 2 (MEN2A, MEN2B) Familial medullary thyroid carcinoma (FMTC) Papillary carcinoma of thyroid |
RET | 125–128 |
The Proteasome as an Anticancer Molecular Target
Regulated degradation of intracellular proteins is mediated by the proteasome, a 2.4-MDa molecular machine comprising approximately 60 subunits that together account for 2% of total cell protein. Not only do proteasomes regulate the half-lives of many signaling proteins in response to environmental stimuli, another primary function of the proteasome is to rapidly degrade hopelessly misfolded proteins that, if allowed to accumulate, can lead to apoptosis. Although the process leading from aggregation of misfolded proteins to cell death is not well understood and is likely to be multifactorial, one hypothesis is that accumulation of aggregated proteins results in the sequestering of numerous cellular chaperones and proteasome components in an insoluble and nonfunctional state, thus negatively affecting normal cell homeostasis and promoting cellular apoptosis. Deregulated protein aggregation has been shown to cause mitochondrial membrane depolarization, release of cytochrome c, and activation of caspase cascades. Indeed, deficiency in proteasome processing of misfolded proteins underlies a number of neurodegenerative diseases characterized by abnormal deposition of insoluble misfolded proteins that cause the apoptotic death of neuronal cells. The frequent acidosis and hypoxia to which cancer cells are subjected cause free radical–mediated damage to cellular proteins. If this damage cannot be repaired, such proteins are cleared from the cell via degradation in the proteasome. Thus, in order to maintain homeostasis, cancer cells are highly dependent on the ability of the proteasome machinery to operate with maximal efficiency.
The proteasome is composed of a 20S core particle containing three proteolytic activities that recognize hydrophobic, basic, and acidic amino acids, respectively. The 26S proteasome is composed of a 20S core particle capped on either end by a 19S ubiquitin chain recognition particle that also uses ATP to unwind substrate protein, allowing it to enter the 20S core where it is degraded into small peptides 2 to 25 residues in length. On exiting the proteasome, these small peptides are rapidly degraded to their amino acid components by cytosolic peptidases. The polyubiquitin chain is also removed and disassembled to monoubiquitin for reuse. The interested reader is referred to several excellent reviews on proteasome function and on validation of the proteasome and ubiquitination machinery as drug targets. 56–58
Figure 56-5 The ubiquitin-proteasome pathway Using ATP, a series of enzymes (E1, E2, and E3) attach multiple units of the small protein ubiquitin to a substrate protein destined for degradation. Substrate specificity is provided by the E3 complex. Once polyubiquitinated, the substrate protein is recognized by the 19S cap of the 26S proteasome and in an ATP-dependent process is unwound and fed into the proteasome for degradation. The 19S core of the proteasome contains chymotryptic, tryptic, and caspase-like proteolytic activities, thus ensuring efficient degradation of the substrate protein into small peptides that, on exiting from the proteasome, are cleaved to their constituent amino acids by cytosolic peptidases. Polyubiquitin chains are disassembled to resupply the cellular monoubiquitin pool. Proteasome inhibitors inhibit one or more enzymatic activities of the 19S core (see text).
As stated earlier, the 26S proteasome recognizes polyubiquitin chains, and most proteins destined for proteasomal degradation are first tagged by sequential covalent addition of four or more ubiquitin moieties (a 76-amino-acid, highly conserved protein present in the cytoplasm and nucleus of all eukaryotes). This process involves ATP-dependent charging of a ubiquitin-activating enzyme (E1), which then transfers ubiquitin to a ubiquitin-conjugating enzyme (E2), which in the presence of a ubiquitin-ligating enzyme (E3) transfers a single ubiquitin moiety to the protein to be degraded (Figure 56-5 ). Interestingly, there is only one E1 enzyme in mammalian cells, approximately 50 E2 enzymes, but perhaps 1000 E3 enzymes. Thus, substrate specificity is primarily regulated by the selectivity of the E3 enzyme. Interaction of the E3 ubiquitin ligase with its substrate is frequently dependent on posttranslational modification (e.g., phosphorylation) of the substrate protein or, in the case of misfolded proteins, on initial interaction of the substrate protein with one or more molecular chaperones, including Hsp90.
Anticancer Activity of the Proteasome Inhibitor Bortezomib
Bortezomib, a dipeptidyl boronic acid–based reversible inhibitor, was the first proteasome inhibitor to enter clinical trials in hematologic malignancies. It has shown significant activity toward multiple myeloma, and in 2005 the U.S. Food and Drug Administration (FDA) approved its use in patients with relapsed multiple myeloma. Multiple additional clinical trials are currently under way examining the efficacy of this agent in various hematologic and solid tumors. 59 Preliminary data suggest promising activity in mantle-cell lymphoma and follicular lymphoma. Several clinical trials are also testing bortezomib in combination with other therapeutic agents, including dexamethasone, doxorubicin, melphalan, and Hsp90 inhibitors. For an in-depth review of bortezomib’s possible mechanism(s) of action and clinical evaluation, the reader is referred to an excellent review by Roccaro and colleagues. 59 A comprehensive review of the current status (as of 2012) of proteasome inhibitors in multiple myeloma is also recommended. 60
Second-Generation Proteasome Inhibitors
In recent years, several second-generation proteasome inhibitors have been developed and are being evaluated in the clinic. MLN9708, like bortezomib, is a boronate derivative, but it has demonstrated greater tissue penetration in preclinical studies and is the first orally available proteasome inhibitor to be evaluated in multiple myeloma. 61 Preliminary clinical data reveal encouraging activity, and durable responses have been seen in heavily pretreated patients. 62 CEP-18770 is a third boronate derivative, also orally active, that is in development. 63
Carfilzomib, an epoxyketone, is an irreversible proteasome inhibitor that is structurally and mechanistically distinct from the boronate-based drugs. Carfilzomib has demonstrated reduced off-target activity compared to bortezomib, and, importantly, activity has been seen against bortezomib-resistant cell lines and primary multiple myeloma cells. 64,65 Carfilzomib has demonstrated durable antitumor activity in patients with relapsed/refractory multiple myeloma and, unlike bortezomib, has caused limited neurotoxicity. 66 An analog of carfilzomib, ONX-0912, is also under development. 67
Marizomib (NPI-0052), a natural product lactone isolated from a marine bacterium, unlike the boronate compounds and carfilzomib, is an irreversible proteasome inhibitor that abrogates both the chymotrypsin-like and trypsin-like protease activities while only minimally affecting the caspase-like activity of the proteasome. As a result, it does not exhibit cross resistance with other proteasome inhibitors, has a unique safety profile, and has demonstrated antitumor activity in preclinical models of multiple myeloma, other hematologic diseases, and solid tumors. 68–70 Marizomib also demonstrates oral bioavailability.
Combined Inhibition of Hsp90 and the Proteasome
Proteasome-mediated degradation is the common fate of Hsp90 client proteins in cells treated with Hsp90 inhibitors. 71,72 Thus, at first glance, combining a proteasome inhibitor with an Hsp90 inhibitor may seem counterintuitive. However, proteasome inhibition does not protect Hsp90 clients in the context of chaperone inhibition—instead, client proteins become insoluble. 49,73 Because the deposition of insoluble proteins is toxic to cells, 74,75 interest has arisen in combining proteasome and Hsp90 inhibitors, the goal being to promote enhanced accumulation of insoluble proteins and trigger apoptosis. This hypothesis is particularly appealing given the clinical efficacy of proteasome inhibitors alone. 59 Initial experimental support for such a hypothesis was provided by Mitsiades and colleagues, 76 who reported that Hsp90 inhibitors enhanced multiple myeloma cell sensitivity to proteasome inhibition. Clinically, a combination of tanespimycin and bortezomib has been associated with durable responses in heavily pretreated patients with multiple myeloma, including those with bortezomib-refractory disease. 77–80 Additional Hsp90 inhibitors are being evaluated in this setting. 81–83
Importantly, transformed cells are more sensitive to the cytotoxic effects of this drug combination than are nontransformed cells. Thus, 3T3 fibroblasts are fully resistant to the combined administration of 17-AAG and bortezomib at concentrations that prove cytotoxic to 3T3 cells transformed by HPV16 virus encoding viral proteins E6 and E7. 84 In the same study, Mimnaugh and co-workers demonstrated that the endoplasmic reticulum is one of the main targets of this drug combination. In the presence of combined doses of both agents that show synergistic cytotoxicity, these investigators noted a nearly complete disruption of the architecture of the endoplasmic reticulum. Because all secreted and transmembrane proteins must pass through this organelle on their route to the extracellular space, it is not surprising that a highly secretory cancer such as multiple myeloma would be particularly sensitive to combined inhibition of Hsp90 and the proteasome. One might speculate that other highly secretory cancers, including hepatocellular carcinoma and pancreatic carcinoma, would also respond favorably to this drug combination.
Additional Rationales for Inhibiting the Ubiquitin-Proteasome System in Cancer
As stated earlier, the proteasome inhibitor bortezomib has demonstrated single-agent activity in multiple myeloma and in other hematologic malignancies. Although general interference in the clearance of misfolded proteins is likely to be a major contributor to the efficacy of this agent, other more specific effects of proteasome inhibition should also be considered. 59 In these hematologic cells, the transcription factor NFκB plays a particularly important role. Not only does it inhibit apoptosis, but it actively upregulates transcription of growth factors such as interleukin 6 and angiogenic factors such as VEGF. As a transcription factor, the activity of NFκB requires nuclear entry. This in turn is regulated by targeted, proteasome-mediated degradation of IκB, a protein that interacts with NFκB and restricts it to the cytosol. Treatment of cells with bortezomib has been shown to prevent the degradation of IκB, thus resulting in retention of NFκB in the cytosol.
Other important tumor suppressor proteins degraded by the proteasome include p53 and p27. Thus, proteasome inhibition promotes the accumulation of these proteins. Recently, investigators have identified more specific approaches to prevent inappropriate p53 and p27 degradation by searching for inhibitors of the E3 ligases that recognize these proteins. The E3 interacting with p53 is termed MDM2 (double minute protein 2). Two small molecules that interfere with MDM2/p53 interaction have recently been identified. 85,86 Although mechanistically distinct, these agents both result in accumulation of wild-type p53 in tumor cells in vitro and shrink tumors growing in mice. The tumor suppressor p27 is degraded by the E3 ligase SKP2 (S phase kinase-associated protein 2). SKP2, which also targets other antiproliferative molecules in the cell—including the retinoblastoma family protein p130, the cyclin-dependent kinase inhibitors p21 and p57, and the inhibitory transcription factor FoxO (forkhead box protein O)—is overexpressed in many human cancers. 57 Molecular knockdown of SKP2 using RNA interference techniques or intracellular injection of SKP2-specific antibodies slows the proliferation of cancer cells in vitro. 57,87 Thus, SKP2 is a valid therapeutic target in its own right, and drug discovery efforts specifically targeting this ubiquitin ligase are under way. 88
Why Are Tumor Cells Uniquely Sensitive to Hsp90 and Proteasome Inhibition?
It is apparent, from both preclinical and clinical observations, that Hsp90 inhibitors can be administered in vivo at doses and schedules that significantly affect tumor growth but display acceptable target-related toxicity in normal tissues or in the whole organism. This is the case for several small-molecule inhibitors, including 17-AAG and 17-DMAG, the synthetic purine mimetic PU24FCl, and is even true for a novel peptidomimetic inhibitor of the N-terminal Hsp90 nucleotide binding site, shepherdin. 89–93 When murine model systems are examined in vivo, Hsp90 inhibitors are found to concentrate in tumor tissue, while being rapidly cleared from normal tissue with a half-life similar to that of drug in plasma. 89,90,92,93 The Hsp90 inhibitor 17-AAG also has been reported to actively concentrate in tumor cells in vitro. 94
Because preferential accumulation of Hsp90 inhibitors in tumor versus normal tissue may provide the observed therapeutic (or at least biologic) index, it is important to understand the reason for this phenomenon. A possible explanation put forth by Kamal and colleagues suggests that enhanced drug binding to tumor cell Hsp90 reflects the activity state of the Hsp90 chaperone machine in tumor versus normal cells. 95 They proposed that the enhanced ATPase activity of the chaperone in tumor cells, which is dependent on preferential recruitment of Hsp90 to a multicomponent chaperone complex, is responsible for the increased affinity of Hsp90 inhibitors in tumor cells. More recently, the suggestion that tumor cell Hsp90 more avidly binds Hsp90 inhibitors compared to the Hsp90 in nontransformed cells has received experimental support, 96 and posttranslational modifications of Hsp90 that affect drug binding have been described. 43,44
The expression of NAD(P)H:Quinone Oxidoreductase I (NQO1), also known as DT-diaphorase, dramatically enhances cellular sensitivity to 17-AAG. 89,97 NQO1 generates the hydroquinone form of 17-AAG, which has been shown to bind more avidly to Hsp90 when compared to 17-AAG itself. 98 Further, the presence of NQO1 in a cell seems also to lead to increased total accumulation of intracellular ansamycin molecules, presumably reflecting the increased Hsp90 affinity of 17-AAG hydroquinone. The presence of NQO1 in tumor cells dramatically affects cellular sensitivity to 17-AAG. 89,97,98 Because high levels of NQO1 have been observed in diverse tumor types (e.g., liver, lung, colon, breast) compared to normal tissues of the same origin, 99 these data suggest an explanation for the disparate sensitivity of tumor and normal tissue to 17-AAG and to retaspimycin (IPI-504), a stabilized hydroquinone form of 17-AAG that is in clinical trial. 100,101 A similar preference of non-ansamycin Hsp90 inhibitors for tumor cell Hsp90 supports the hypothesis that unique modifications of Hsp90 in tumor versus normal cells also contribute to this phenomenon.
Proteasome inhibitors also display selective cytotoxicity toward tumor cells, both in vitro and in vivo. Why are they not more toxic to normal cells? Kisselev and colleagues 102 have recently shown that, at therapeutic doses of bortezomib in vivo, and following the intermittent schedule of administration approved for patients, only the chymotryptic activity of the proteasome is significantly inhibited and the overall rate of protein degradation is reduced by less than 40%. Because cancer cells may require their proteasomes to be at full capacity in order to handle the load of continually generated misfolded proteins, moderate reduction in proteasome activity, even for a brief period, may prove deleterious. In contrast, nontransformed cells may be able to tolerate reduced proteasome function for an extended period of time. Some of our data described earlier in this review support this hypothesis. In the preclinical studies in which we observed dramatically enhanced toxicity (and dramatically enhanced insoluble ubiquitinated protein deposition) by combining low-dose Hsp90 inhibitors with bortezomib, we found that maximal benefit of the combination required only a 50% reduction in proteasome activity. 84 Importantly, the satisfactory safety profiles of second-generation irreversible inhibitors—including marizomib, which inhibits two of three proteasome-associated protease activities—suggest that nontransformed cells may be able to tolerate prolonged, significantly reduced proteasome activity, further amplifying the therapeutic index of these drugs.
Conclusion
The proteasome and Hsp90 together comprise approximately 4% of total cellular protein. Separately and together they regulate critical mechanisms responsible for maintaining cellular homeostasis in the face of severe environmental stress and constitutive genetic instability. Although Hsp90 and the proteasome certainly can be labeled as “housekeeping” proteins, both proteins have proven to be exciting and clinically relevant anticancer molecular targets. Because it is now clear that cancer cells, by their very nature, are more dependent on optimal function of homeostatic mechanisms compared to normal cells, further exploitation of these pathways as drug targets will likely provide additional therapeutic strategies to attack this disease.
References
1. The onset and extent of genomic instability in sporadic colorectal tumor progression . Proc Natl Acad Sci U S A . 1999 ; 96 : 15121 – 15126 .
2. Modelling the molecular circuitry of cancer . Nat Rev Cancer . 2002 ; 2 : 331 – 341 .
3. Insights from pre-clinical studies for new combination treatment regimens with the Bcr-Abl kinase inhibitor imatinib mesylate (Gleevec/Glivec) in chronic myelogenous leukemia: a translational perspective . Leukemia . 2002 ; 16 : 1213 – 1219 .
4. Cancer robustness: tumour tactics . Nature . 2003 ; 426 : 125 .
5. Protein misfolding and disease: the case of prion disorders . Cell Mol Life Sci . 2003 ; 60 : 133 – 143 .
6. Chaperone-related immune dysfunction: an emergent property of distorted chaperone networks . Trends Immunol . 2006 ; 27 : 74 – 79 .
7. Navigating the chaperone network: an integrative map of physical and genetic interactions mediated by the hsp90 chaperone . Cell . 2005 ; 120 : 715 – 727 .
8. Inhibition of Hsp90: a new strategy for inhibiting protein kinases . Biochim Biophys Acta . 2004 ; 1697 : 233 – 242 .
9. Hallmarks of cancer: the next generation . Cell . 2011 ; 144 : 646 – 674 .
10. Heat-shock protein 90 inhibitors in antineoplastic therapy: is it all wrapped up? Expert Opin Investig Drugs . 2005 ; 14 : 569 – 589 .
11. Hsp90: the vulnerable chaperone . Drug Discov Today . 2004 ; 9 : 881 – 888 .
12. Combinatorial attack on multistep oncogenesis by inhibiting the Hsp90 molecular chaperone . Cancer Lett . 2004 ; 206 : 149 – 157 .
13. Altered Hsp90 function in cancer: a unique therapeutic opportunity . Mol Cancer Ther . 2004 ; 3 : 1021 – 1030 .
14. Targeting multiple signal transduction pathways through inhibition of Hsp90 . J Mol Med . 2004 ; 82 : 488 – 499 .
15. The Hsp90 chaperone complex as a novel target for cancer therapy . Ann Oncol . 2003 ; 14 : 1169 – 1176 .
16. Heat shock protein 90 as a molecular target for cancer therapeutics . Cancer Cell . 2003 ; 3 : 213 – 217 .
17. Hsp90 as a capacitor for morphological evolution . Nature . 1998 ; 396 : 336 – 342 .
18. Hsp90 as a capacitor of phenotypic variation . Nature . 2002 ; 417 : 618 – 624 .
19. HSP90 and the chaperoning of cancer . Nat Rev Cancer . 2005 ; 5 : 761 – 772 .
20. Identification of aneuploidy-selective antiproliferation compounds . Cell . 2011 ; 144 : 499 – 512 .
21. Geldanamycin, a new antibiotic . J Antibiot (Tokyo) . 1970 ; 23 : 442 – 447 .
22. The benzoquinone ansamycin 17-allylamino-17-demethoxygeldanamycin binds to HSP90 and shares important biologic activities with geldanamycin . Cancer Chemother Pharmacol . 1998 ; 42 : 273 – 279
Buy Membership for Hematology, Oncology and Palliative Medicine Category to continue reading. Learn more here
