Current Topics in Medicinal Chemistry - Volume 9, Issue 18, 2009

In the last decade fragment-based methods have become rapidly established in drug discovery and although it is too early for these approaches to have yielded a marketed drug, they have resulted in a significant number of drug candidates entering clinical trials and many more in pre-clinical development. Fragment-based methods were first successfully applied, further explored and rapidly adopted in the biotechnology and pharmaceutical industry, but due to the moderate costs of screening small libraries of simple compounds, fragment-based screening has also been enthusiastically embraced in academia. This has further enriched the thriving fragment-based drug discovery community and will foster the education of the next generations of scientists in modern drug discovery approaches. In this special issue of Current Topics in Medicinal Chemistry on Case Histories in Fragment-based Drug Discovery we have tried to bring together articles that, as well as putting fragment-based drug discovery in its historical context and providing a current overview of the field, offer an exciting outlook of how the field could further evolve in terms of technology and its application to target classes that were until recently deemed to be difficult or unsuitable for structure-based or fragment-based approaches. In their introduction Christophe Verlinde, Wim Hol and coworkers present an historical overview of fragment-based drug discovery (FBDD) including their own attempts on cocktail crystallography in the early 1990s, which clearly highlights the evolution of fragment-based drug discovery from the initial ideas of the 1980s to the mature field it is now. It certainly appears that a commonly used workflow for fragment screening is emerging, which involves initial hit detection using biophysical methods or high concentration biochemical assays, followed by structural characterization typically by X-ray crystallography. However, their article also shows that cocktail crystallography remains a powerful direct screening method and that it can be efficiently used in an academic structural genomics environment. Vicki Nienaber describes how the concepts underlying FBDD can be applied to drug targets in the central nervous system which require inhibitors to cross the blood-brain barrier and thus place additional restraints on their physicochemical properties. At a time when many fragment libraries have tended to increase in average molecular weight due to the use of high concentration bioassay as a pre-filter, or to allow for including certain desirable heavy functional groups in the compounds, she and her coworkers designed a fragment library with a lower average molecular weight than a typical fragment library to help keep the physicochemical properties of their inhibitors under tight control during lead optimization. Early examples indicate that the application of FBDD to neuroscience targets shows considerable promise. Tom Davies, Steven Woodhead and Ian Collins describe collaborative research by Astex Therapeutics and The Institute of Cancer Research aimed at designing potent and selective protein kinase B (PKB) inhibitors as antitumour agents. Their contribution shows how low affinity fragment hits were rapidly optimized to potent PKB inhibitors using a PKA-PKB chimeric protein as a surrogate system for high-throughput X-ray crystallography, which was crucial in structural characterization of the fragment hits and in guiding the medicinal chemistry in the hits-to-leads and lead optimization phases of the project. In addition, they show that selectivity of one of the inhibitor scaffolds for PKB.. over PKA could be structurally explained using protein-ligand structures of PKA, the PKA-PKB chimera and PKBβ. Moreover, inhibitors from two different chemical series showed biomarker modulation and a reduction in tumour growth in mouse xenografts, illustrating the ability to discover efficacious lead compounds with properties suitable for further development using FBDD methods. One of the ideas adopted quickly by scientists involved in FBDD is the concept of ligand efficiency, which can be thought of as a metric for normalizing potency while accounting for the number of heavy atoms in a fragment. Such normalization shows that although fragments may bind with low affinity, for their size they are actually efficient binders. Allen Reitz and coworkers detail the use of a range of efficiency indices in assessing fragment hit quality, such as the percentage efficiency index and Fit Quality. They show that ligand efficiencies are systematically lower for larger ligands and argue that the Fit Quality index is a more consistent metric for comparing small and large compounds. In addition, they describe how FBDD ideas can be transferred to more traditional screening campaigns by applying these indices appropriately in the analysis of the screening hits. Although fragment screening using a high concentration bioassay is becoming more popular, it is often the case that such assays are not sensitive enough to detect weakly binding fragments. In these cases one has to resort to sensitive biophysical methods for hit detection, typically NMR or X-ray crystallography. However, another biophysical technique increasingly used in fragment screening is Surface Plasmon Resonance (SPR). Helena Danielson thoroughly reviews the application of this relatively young technique in fragment screening and illustrates the experimental setup, successes and potential pitfalls with several examples. Similar to NMR and X-ray crystallography, SPR can be used in screening and may also provide valuable information in the more advanced hit-to-lead and lead optimization stages of a drug discovery project. Moreover, when used in combination with one or both of these techniques SPR could be very powerful in thoroughly characterizing the available hit matter, and more advanced inhibitors, thus accelerating the design of high quality lead compounds. Gregg Siegal and Johan Hollander describe one of the most exciting new technologies in fragment-based drug discovery, Target Immobilised NMR Screening (TINS). This is a ligand-based NMR method which, similar to SPR, is based on the immobilization of the protein target during screening using a variety of methods. However, in the TINS method the target and a reference protein are immobilized on a resin in a dual flow-cell placed in the NMR spectrometer. Binding is detected as a reduction in peak height in a ligand NMR spectrum in the presence of target protein as compared to the reference protein. Amongst the advantages TINS has in common with SPR are the low amounts of protein required, the ability to monitor the health of the protein during the screen, and the use of reference proteins to discriminate between real hits and non-specific binders. TINS can easily be used to obtain experimental insight into the drugability of a target and can be followed up by more advanced NMR methods to further characterize the hits and determine their binding constants, as exemplified by the authors for a protein-protein interaction target. In addition they address one of the big current challenges in FBDD, which is how to progress fragment hits when protein ligand-structures are not available or are very difficult to obtain within the timeframe of a drug discovery project. The authors sketch how TINS, in combination with additional NMR approaches such as the incorporation of paramagnetic centers to obtain distance and orientation information for bound ligands, could be used to obtain low-resolution structural information to aid the elaboration of fragments into more lead-like molecules. Finally, they describe the application of TINS to membrane proteins, targets that to date have not been amenable to structure-based or fragment-based drug discovery approaches due to the difficulties in obtaining high resolution structural data. Preliminary results from a TINS screen on the membrane protein DsbB, demonstrate that the technology could in principle be applied to certain classes of membrane proteins. It will therefore be exciting to see how TINS will be applied to pharmaceutically more relevant membrane proteins such as the G protein-coupled receptors. We hope that our selection of articles will convince you that FBDD continues to be an exciting approach in drug discovery with established utility and the promise for significant further advances. We would like to thank all the authors for their contributions. Finally we would like to thank Allen Reitz for the opportunity to put this issue together.

The history of fragment-based drug discovery, with an emphasis on crystallographic methods, is sketched, illuminating various contributions, including our own, which preceded the industrial development of the method. Subsequently, the creation of the BMSC fragment cocktails library is described. The BMSC collection currently comprises 68 cocktails of 10 compounds that are shape-wise diverse. The utility of these cocktails for initiating lead discovery in structure-based drug design has been explored by soaking numerous protein crystals obtained by our MSGPP (Medical Structural Genomics of Pathogenic Protozoa) consortium. Details of the fragment selection and cocktail design procedures, as well as examples of the successes obtained are given. The BMSC Fragment Cocktail recipes are available free of charge and are in use in over 20 academic labs.

Diseases of the central nervous system are among the most devastating to patients and their families. Despite this, treatments for these diseases have lagged behind other therapeutic areas. Although social and economic factors may be partly responsible for the paucity of therapeutic agents, a particularly daunting challenge for CNS drug discovery is the need for compounds to cross the blood brain barrier. Recent analyses of successful drugs have shown that their chemical properties have not changed substantially over the past 40 years while the properties of compounds entering the clinic have become inflated. This property inflation has only exacerbated the challenges of CNS drug discovery as the requirements for delivery to the brain are even more stringent than those for other tissues. New approaches are needed to meet these challenges. In this review, we discuss the merits of fragment based lead discovery and how it may be used to address the challenges of CNS drug discovery. We also summarize compounds discovered by high-throughput screening, substrate evolution and fragment-based lead discovery for well known CNS targets. The results indicate that FBLD may be a key method for discovery of brain penetrable CNS therapeutics.

Multiple ligand efficient fragment inhibitors of protein kinase B were identified through a combined in silico compound screen and high-throughput crystallographic analysis of protein-ligand structures. A well-validated apo-PKBPKA chimeric protein provided a convenient platform for high-throughput crystallography by soaking of inhibitors, and a method for the determination of PKB-ligand structures was developed to support inhibitor development. Pyrazole and azaindole fragment hits with micromolar activity were rapidly progressed to nanomolar inhibitors of PKB with activity in cells using crystallographic analysis of inhibitor binding modes to guide medicinal chemistry. Compounds with selectivity for PKB over PKA and other kinases were identified by this approach, resulting in potent inhibitors with in vivo activity through both oral and systemic administration, and suitable for further development.

Fragment-based drug discovery (FBDD) is an important new tool to understand the molecule basis of ligandbiological target interactions. By combining optimal fragments, it is often possible to construct larger molecular weight compounds that have greater potency in a shorter period of time than can been achieved by the initial screening of larger molecular weight compound libraries. Alternatively, if screening of more traditional larger libraries has occurred, then it may be possible to analyze the data during the process of hit triage in such as way as to essentially adopt a fragment-based approach in reverse. In this review, we highlight general principles associated with the efficiency indices such as Ligand Efficiency (LE) in which screening data is normalized for biophysical properties such as molecular size. We further focus on the concept of Fit Quality (FQ), which standardizes LE values across molecular weight for more realistic, direct comparison. Using these simple concepts, one can apply FBDD routinely in the stage of hit triage when evaluating the data obtained after screening of compound libraries in drug discovery.

The transition from high throughput screening of collections of drug-like compounds to screening of fragment libraries via lower throughput methods with high sensitivity has revolutionized early drug discovery. It is highlighting the need for sensitive biophysical techniques for interaction analysis rather than high throughput methods. Biosensors with SPR detection are well suited for this novel scenario. In less than 20 years the technique has been launched, established and become a highly informative method for a variety of applications in drug discovery. It is no longer limited to the detection of proteins or other high molecular weight analytes, but the detection of weakly interacting fragments is now feasible. This paper discusses the theoretical and experimental limitations for such applications and reviews a number of successful studies in the area of fragment-based lead discovery that have recently been published. It can be anticipated that the evolution of this young technique will be significantly influenced by the requirements for efficient fragment-based lead discovery.

Using localized NMR spectroscopy on immobilized targets provides us with a method to simultaneously assess binding of small molecules to two different samples. This Target Immobilized NMR Screening (TINS) has a number of advantages, not least is the requirement for minimal quantities of non-isotopically labeled protein and the applicability to insoluble or unstable targets. The technique is sensitive to binding with KD values in the range of 100 nM to 20 mM, while careful selection of the reference protein reduces the number of false positive hits. This combination ensures a maximal number of valid hits from which to select starting points for hit elaboration projects. Hits can be prioritized using biological assays when appropriate, as well as an array of biophysical techniques. So far a variety of soluble proteins, including kinases, GTPases, viral targets and proteases, as well as a membrane protein, have been successfully screened against our fragment library. Here we illustrate our experiences with a number of examples which emphasize the usefulness of the method in selecting and prioritizing fragment hits for elaboration towards leads.