Heat Shock Protein (HSP) Inhibitors
Proteins are involved in nearly every aspect of cellular function, and producing proteins with the correct amino acid sequence is only the first step for a cell. After this, most proteins, which are very large and complex molecules, need to fold into exactly the right three-dimensional shape before they can function properly. Proteins that are not folded properly are identified and tagged for degradation by the proteasome. The molecules that facilitate and carry out the folding process are proteins themselves and are known as chaperone molecules or heat shock proteins (HSPs). One protein of this type, HSP90, is important because it is a master regulator that controls a series of other HSPs.
FIGURE 6.109 Diagram showing the heat shock protein 90 (HSP90) in vivo cycle. The client protein binds to the co-chaperones HSP70 and HSP40, and then is loaded onto the HSP90 homodimer. Binding of ATP then causes a conformational change that stabilizes the client protein. Furthermore, the hydrolysis of ATP to ADP causes the release of the client protein which can undergo degradation through the proteasome (Taken from Lackie RE et al. (2017) The Hsp70/Hsp90 Chaperone Machinery in Neurodegenerative Diseases. Front. Neurosci. 11:254. doi: 10.3389/fnins.2017.00254).
It is crucial for the folding of several client proteins, many of which play a role in the proliferation of tumor cells (Figure 6.109). For example, client proteins include mutated p53 (important in most cancers), BCR-ABL (important in chronic lymphoblastic leukemia), human epidermal growth factor receptor 2 (HER2)/neu (important in breast cancer), other kinases such as Raf-1, ErbB2, Cdk4, c-Met, Polo-1 and Akt, telomerase hTERT, and steroid hormone receptors.
As cancer cells, due to their genetic mutations, typically contain a large number of proteins that do not fold properly, they often compensate by over-expressing HSP90. The result is that even mutated proteins usually fold sufficiently well to avoid disposal by the proteasomes, thus allowing cancer cells to survive. Therefore, inhibiting the activity of HSP90 should lead to a reduction in folding activity in all cells, with more client proteins being destroyed by the proteasomes. However, because cancer cells have a higher dependence on HSP90 than normal cells ((.Chapter 1), and so has the potential for broad-spectrum clinical activity across multiple cancer types. However, there is a risk of serious adverse effects with inhibitors of this type as HSP90 is also involved in normal cell functioning.
In the late 1960s, the Upjohn Company discovered geldanamycin (Figure 6.110), a natural product isolated from Streptomyces hygroscopicus, that has a suitable shape to interact in a “pocket” of the N-terminal domain of HSP90, thus blocking its ability to assist in the folding of proteins. Geldanamycin is a benzoquinone ansamycin, a member of a family of natural products originally attributed with weak antibiotic activity but later discovered to have potent antitumor activity. A natural product with similar activity, radicicol, was isolated from the mycoparasite Humicola fuscoatra (Figure 6.110).
FIGURE 6.110 Structures of geldanamycin, 17-AAG (Tanespimycin), 17-DMAG, and radicicol.
Eventually, X-ray crystallographic studies provided the structure of the geldanamycin-binding domain of HSP90, revealing a pronounced 15 А-deep pocket formed by residues 9-232 that is highly conserved across species. Geldanamycin binds into this pocket, adopting a compact structure similar to that of a polypeptide chain in a turn conformation. This binding mode, along with the pocket’s similarity to other substrate-binding sites, suggests that its probable function is to bind a portion of the client protein substrate and participate in the conformational maturation and re-folding reaction.
Unfortunately, although geldanamycin was active in preclinical studies, it was a poor candidate for clinical trials due to its poor water solubility, instability, and in vivo toxicity in animal models. Therefore, during the 1980s, the National Cancer Institute (NCI) worked on several geldanamycin analogs to develop more water-soluble and less-toxic versions of the drug. Eventually, in 1992, 17-allylamino-demethoxygeldanamycin (tanespimycin, 17-AAG) (Figure 6.110) was produced in collaboration with the company Kosan. This analogue was shown to bind specifically to HSP90 in a similar manner to geldanamycin but with a lower affinity. Therefore, it is surprising that tanespimycin and geldanamycin have similar cytotoxicities toward tumor cells in vitro. However, tanespimycin had a better toxicity profile than geldanamycin and was a preferred clinical candidate, reaching Phase I/II clinical trials with Bristol-Myers Squibb (BMS). Preliminary results indicated that a target dose of tanespimycin could be achieved without dose-limiting toxicity, and it was notably less hepatotoxic than geldanamycin and could be administered at higher doses. The most common side effects were anorexia, nausea, and diarrhea, all of which could be dose limiting. Less-common side effects included hepatotoxicity, fatigue and blood dyscrasias (e.g., thrombocytopenia, anemia). Despite being in late-stage clinical trials for the treatment of multiple myeloma, in 2010 BMS halted development of tanespimycin without explanation, although some observers suggested that the relatively short patent life and cost of manufacture may have played a role.
In the late 1990s, a further analog, 17-dimethylaminoethylamino-demethoxygeldanamycin hydrochloride (17-DMAG) (Figure 6.110), was identified by the NCI in collaboration with Kosan. This analogue had excellent bioavailability, was widely distributed to tissues, and was metabolized at a lower rate compared with 17-AAG. Although it reached the clinic and was evaluated for activity in both lymphomas and solid tumors, these studies were also terminated.
Finally, it is worth noting that results from preclinical studies indicate that geldanamycin and its analogs have synergistic activity with traditional cytotoxic agents and inhibitors of signal transduction pathways. Therefore, if a clinically viable HSP inhibitor could be developed in the future, it may have potential to be used in combination with other anticancer agents.
Conclusions
The molecularly targeted agents now represent one of the largest and most successful families of anticancer agents. They were originally designed to improve on the antimetabolites, antitubulin agents and DNA-interactive drugs which were the mainstay of cancer treatment in the second-half of the 1900s, working mainly through a generalized effect on cell division with little discrimination between tumor and healthy cells, and leading to serious side effects (e.g., bone marrow suppression, gastrointestinal, cardiac, hepatic and renal toxicities, nausea, vomiting and hair loss). The molecularly targeted agents were pioneered in the 1980s based on the concept of targeting biochemical pathways or proteins that are uniquely deregulated or mutated, respectively, in tumor cells which should produce greater efficacy and significantly less toxicity as the mutated proteins or deregulated pathways should not be present in healthy cells. A second important design concept was that the inhibitors should be orally available in contrast to the older families of agents that are mostly administered intravenously and require hospitalization. Crucially, these two design concepts of selective toxicity to tumor cells and oral availability raised the possibility of treating cancer as a chronic rather than acute disease in a similar manner to conditions such as diabetes and heart disorders. This design concept was validated with the discovery of imatinib (Gleevec™) in the 1990s for the treatment of chronic myeloid leukemia which achieved all the above objectives. This success led to the development of many other subfamilies of molecularly targeted agents which are described in this chapter. Companies are continuing to discover and develop new agents in this area, driven by their potential in a “Precision Medicine” scenario (CHAPTER 11) in which patients can be selected for treatment with a specific molecularly targeted agent by checking for the presence of the relevant mutated protein or deregulated pathway in their tumor cells before treatment begins, thus maximizing the likelihood of a clinical response. It is likely that research and development will continue unabated in this area in the coming decades with many more agents of this type reaching the approval stage.
