1
Chemical Strategies for Evaluating New Drug Targets

Adrian J. Carter1, Raina Seupel2, Paul E. Brennan2, Michael Sundström3, Andrea Introini3, and Anke Mueller-Fahrnow4

1 Boehringer Ingelheim, Discovery Research Coordination, Binger Strasse 173, 55216 Ingelheim, Germany

2 University of Oxford, Structural Genomics Consortium and Target Discovery Institute, Nuffield Department of Medicine, NDMRB, Roosevelt Drive, Oxford, OX3 7FZ, UK

3 Karolinska Institutet, Structural Genomics Consortium, Karolinska Vägen 6, 17176, Sweden

4 Bayer AG, Target Discovery, Muellerstrasse 178, 13342 Berlin, Germany

1.1 Introduction

Discovering new drugs is difficult and expensive. Pharmaceutical companies typically spend at least $2.6 billion on average in research and development (R&D) for each drug before it reaches the market [1]. Interestingly, the high cost of R&D is not driven by the few programmes that succeed, but rather by the cost of pipeline projects that fail [2]. Only about 1 in 10 of drug candidates in phase I clinical trials actually makes it to become a new medicine [3,4], and about half of the projects that fail in phase II clinical trials do so because of clinical efficacy [5,6]. So why do so many drugs fail?

One answer is a lack of genetic evidence. An analysis of AstraZeneca's small molecule pipeline indicated that the success rate was over 70% for projects in phase II clinical trials with human genetic linkage of the target to the disease indication compared with 43% for projects without such a linkage [6]. Furthermore, another similar study concluded that selecting genetically supported targets can double the success rate in clinical development [7]. These observations have prompted some scientists to highlight the critical importance of the therapeutic hypothesis at the stage when a protein or gene is selected as a potential drug target [8]. However, it is often a long and difficult road between identifying a genetic link and understanding the underlying biological processes (see Chapter 6 for more details).

A major problem we are facing is that a large proportion of biomedical R&D focuses on only a small fraction of the genome despite the promised revolution in medicine following sequencing of the entire human genome [9]. Shortly after its announcement, scientists imagined that genome science would soon begin revealing the mysteries of hereditary factors in heart disease, cancer, diabetes, schizophrenia, and a host of other conditions and lead to new medicines [10]. Unfortunately, this has not happened. Indeed, more than 75% of protein research still focuses on the 10% of proteins that were known before the genome was mapped, even though many more have been genetically linked to disease [11]. A more recent analysis of drug targets highlights the continued dominance of a set of privileged target families across different disease areas, although there has also been a small growth of novel first-in-class mechanisms, particularly in oncology [12]. What can we do to help biomedical scientists worldwide to expand and prioritize the list of potential new drug targets?

One answer lies with high-quality chemical probes. We know that chemical tools can dramatically facilitate exploratory biomedical research. Let us take, for example, nuclear hormone receptors. When nuclear receptors were identified by sequence homology in the 1990s, all the family members were thought to have therapeutic potential. Scientists initially investigated those receptors that were found to have genetic links to disease or that had interesting knockout phenotypes. However, as time went on, research activity focused on a subset of eight of these receptors despite the fact that these eight were no more genetically interesting than the others. Indeed Edwards [11] postulated that the only connection among these eight receptors is that for each there exists a widely available, high-quality chemical probe that either enhances the receptor's activity or dampens it. In short, where high-quality tools are available, there is research activity; where there are no tools, there is none. What is a high-quality chemical probe and why are they so useful?

The Structural Genomics Consortium (SGC) (www.thesgc.org) is a large precompetitive public-private partnership between academia, private funders, and currently nine public pharmaceutical companies as well as patient advocacy and research organizations. The consortium has established a common set of principles for chemical probes, initially focused on epigenetic targets. A chemical probe is simply a small molecule that modulates the function of a protein in a specific and selective way. This allows a scientist to interrogate the biology and test hypotheses relating to the mechanism or role of the particular protein in a relevant cellular context [13]. The difference between specificity and selectivity is important to consider. Specificity is the capacity of a chemical probe to manifest only one kind of action. A chemical probe of perfect specificity of action might increase, or decrease, a specific function of a given cell type, but it would not do both, nor would it affect other receptors. In contrast, selectivity is the ability of a chemical probe to affect one cell population in preference to others, i.e. the ability of a chemical probe to affect one kind of cell, and produce effects, in doses lower than those required to affect other cells. This should not be confused with potency, i.e. the measure of the activity of a chemical probe, in terms of the concentration or amount required for producing a defined effect. Consequently, selectivity is actually a measure of the relative potency of a chemical probe in producing different effects.

The SGC has established a set of stringent criteria that a chemical tool compound must fulfil in order for its classification as a chemical probe (Figure 1.1). The compound must exhibit in vitro potency of less than 100 nmol/l for a single target or a small set (<5) of very similar targets and possess a minimum 30-fold selectivity relative to other sequence-related proteins of the same family. Furthermore, the probe must be profiled against a standard selection of other unrelated pharmacologically relevant targets and large protein families of relevance to drug discovery (specificity) and, finally, have demonstrated on-target effects in cells at less than 1 µmol/l (cellular activity) [13]. These criteria were developed jointly by academia and industry to help guide scientists to choose the best chemical probe for their needs because, unfortunately, there are numerous examples of less well-characterized tools that may generate misleading results. Although the chemical probe criteria could be stricter (e.g. 10 nmol/l potency, selectivity for a single target >100-fold, cellular activity <100 nmol/l), the chemical probe criteria outlined here are a pragmatic compromise between the cost of creating such a chemical probe and the potential value it brings to science.

Figure 1.1 Chemogenomic tools support the identification of potential target proteins, and chemical probes help interrogate biological systems selectively and specifically with the goal of ultimately developing safe and efficacious medicines.

Seven pharmaceutical companies have since worked together in a precompetitive collaboration to make available a large number of innovative, high-quality chemical probes from previous terminated pharmaceutical projects [14]. These probes have all been subject to a rigorous and independent scientific review by the SGC, are accompanied by comprehensive data packages, and include an appropriate structurally related yet inactive control compound. All of this is made available to the wide scientific community free of intellectual property restrictions via the portal https://openscienceprobes.sgc-frankfurt.de.

Regrettably, making a high-quality chemical probe entails a lot of effort especially for medicinal chemists. Indeed, the sort of efforts that are required to make such a tool are those that are only usually available for a full-fledged drug discovery project. How do we justify that level of investment if we do not know how important the target protein is a priori? Some scientists have likened it to a catch-22 situation. A scientist cannot justify the resources required for generating a chemical probe without knowing the importance of the target protein. On the other hand, the scientist cannot judge the importance of a protein without being able to interrogate its function with a chemical probe. But help is at hand. One way around this is to build chemogenomic sets of compounds for specific protein families known as chemogenomic libraries [15].

Chemogenomics was coined as a term to describe the use of target family-directed chemical libraries in target or cell-based assays as a means of accessing new areas of biology and accelerating drug discovery research based on the assumption that similar receptors bind similar ligands [16-18]. Such sets, although containing compounds that individually do not fulfil the stringent criteria of a chemical probe, nevertheless, can be used to interrogate multiple members of protein families to help...