Recent Advances in Cancer Therapeutics

Nicola Chessum; Keith Jones; Elisa Pasqua; Michael Tucker    Cancer Research UK Cancer Therapeutics Unit, The Institute of Cancer Research, London, United Kingdom

1 Introduction

In the past 20 years, cancer therapeutics has undergone a paradigm shift away from the traditional cytotoxic drugs towards the targeting of proteins intimately involved in driving the cancer phenotype [1]. The poster child for this alternative approach to the treatment of cancer is imatinib, a small-molecule kinase inhibitor designed to target chronic myeloid leukaemia driven by the BCR–ABL translocation in a defined patient population [2]. The improvement in survival achieved by treatment of this patient cohort with imatinib is impressive. Thus, the aim is to provide efficacy but with low toxicity. The role of the medicinal chemist in oncology drug discovery is now closely aligned with the role in most other therapeutic areas with high-throughput and/or fragment-based screening, structure-based design, selectivity, pharmacokinetic optimisation and pharmacodynamic biomarker modulation, all playing a familiar part in the process.

In a short review of this nature covering such a large and active field, we have had to be selective in our choices of examples. We have chosen four areas in which compounds are either approved drugs or in clinical trials. These are chaperone inhibitors, kinase inhibitors, histone deacetylase inhibitors and inhibitors of protein–protein interactions (PPIs). Even within these areas, we have had to be selective, particularly for kinase inhibitors, and we have tried to exemplify newer approaches and novel aspects of medicinal chemistry.

2 Chaperone Inhibitors

Although the targeting of oncogenic proteins such as kinases has led to significant clinical benefit over the last 15 years, expectations have outweighed the reality in terms of outcomes. There are many reasons for this including tumour heterogeneity, intrinsic and acquired drug resistance and the presence of multiple oncogenic drivers in any single cancer. Targeting the cellular machinery responsible for protein quality control provides a more wide-ranging, yet still targeted, approach to inhibiting oncogenic proteins. Proteins consist of an elaborate arrangement of folds and secondary structure and, although many aspects of the folding are inherent in the properties of the protein itself, the process is complex and errors occur [3]. Indeed, the final, stable structure is often characterised by a free energy gain of some 3–7 kcal/mol over a range of partially misfolded states [4]. In a crowded cellular environment, correct protein folding is made even more difficult because of collisions between protein molecules [5]. Cells have developed a number of mechanisms to cope with ensuring that correct protein conformations are maintained. In the nuclear and cytosolic compartments, the heat-shock response, involving a range of heat-shock proteins (HSPs), is a conserved mechanism for dealing with misfolded proteins. Originally thought to be an emergency response to sudden stress, it is now recognised to be a constant process enabling protein homeostasis. In the context of cancer cells, the heat-shock response is a vital method of maintaining protein function in the stressed oncogenic state, and targeting this response may provide a combinatorial blockade of multiple oncogenic proteins.

Heat-shock protein 90 (HSP90) is a member of the high molecular weight HSPs, along with HSP70. It accounts for some 1–2% of all cellular protein [6], and there are four closely homologous, important isoforms: HSP90α and HSP90β which are cytosolic, GRP94 which is found in the endoplasmic reticulum and TRAP1 found in mitochondria. HSP90 has a long list of client proteins that include a number of key oncogenes such as ERBB2, RAF and the androgen receptor [7]. Active HSP90 is a homodimer with each monomer consisting of an N-terminal ATP-binding domain, a middle domain and a C-terminal dimerisation domain. Seminal crystallographic studies by the Pearl group identified the ATP-binding pocket of HSP90 [8] and provided strong evidence for a catalytic cycle driven by ATP hydrolysis [9]. A variety of co-chaperones have been identified as playing roles in the catalytic cycle but HSP90 remains the key effector (Figure 1).

The first HSP90 inhibitors to be recognised were the natural products, geldanamycin (1) and radicicol (2). Geldanamycin is a benzoquinone ansamycin isolated from a Streptomyces species and was originally thought to be a tyrosine kinase inhibitor. In 1994, it was shown to bind to HSP90 [10], and this was followed in 1999 by a co-crystal structure of geldanamycin bound to the N-terminal domain of HSP90 [11]. Structurally, the ATP-binding pocket of HSP90 belongs to the unusual Bergerat fold class of ATP-binding sites that is shared by relatively few proteins [12]. The co-crystal structure of the HSP90 N-terminal domain and geldanamycin is both interesting and informative. The ligand adopts a folded conformation in its bound state in which the benzoquinone ring folds back over the macrocycle with the benzoquinone at the top (open) part of the pocket. There are a number of water-mediated hydrogen bond interactions between the ligand and the protein, a feature of all ligands that bind HSP90 in this pocket including the natural ligands ATP and ADP (Figure 2).

Geldanamycin shows not only potent in vitro and in vivo antitumour activity but also severe hepatotoxicity in preclinical animal models. It also has poor physicochemical properties including solubility, and owing to its benzoquinone moiety, it is a substrate for NQ01: a quinone reductase that converts the parent quinone to the more active hydroquinone. Semi-synthetic derivatives of geldanamycin have addressed the solubility issues. The two successful ones are 17-allyaminogeldanamycin (17-AAG, 3) and 17-dimethylaminoethylgeldanamycin (17-DMAG, 4) in which the 17-methoxy group of the quinone has been replaced by the appropriate amino group in a simple addition–elimination reaction. Following promising anticancer activity in preclinical models [13], 17-AAG (tanespimycin) entered clinical trials in 1999 (administered intravenously), and evidence of clinical activity was seen in a variety of cancers as a single agent and in combination [14] but its development was stopped after Phase II trials owing to formulation and patent life issues. The greater solubility of 17-DMAG (alvespimycin) allowed easier formulation, and a complete response was reported following i.v. administration as a single agent in castration-refractory prostate cancer in a Phase I/II trial [15]. Again, clinical investigations have been halted owing to toxicity issues [16]. The only significant alternative approach to address the issues with geldanamycin and its derivatives was by Infinity Pharmaceuticals who prepared the hydroquinone hydrochloride salt of 17-AAG, designated IPI-504 or retaspimycin hydrochloride. Following initial positive results in a Phase II trial, it has now been withdrawn owing to toxicity [17].

The second natural product, radicicol, is an extremely potent inhibitor of HSP90...