Conference participants heard about recent innovations in drug discovery research in a symposium on combinatorial chemistry and high throughput screening on September 14
The majority of pharmaceutical companies are now taking high throughput drug discovery very seriously.
|
Ian Matthews: Change of pace in drug discovery |
The emphasis in drug discovery was now on fast drug candidate selection, commented Dr Matthews, describing the application of combinatorial chemistry and high throughput screening (see Panel) in the pharmaceutical industry. Traditionally, the discovery and development of a new drug would take 12 years and cost £300m. This generally involved the identification of "hits" from screens followed by the selection of lead compounds and initiation of drug refinement and development. There was, he said, a driving force to shorten the pipeline of drug development.
|
CC and HTS
Combinatorial chemistry (CC) and high throughput screening (HTS) are drug development tools that have been developed within the past decade. As described by Dr Morphy, CC is "a technique by which large numbers (libraries) of compounds are synthesised simultaneously in the time usually taken to prepare only a few compounds by conventional methodology". Most commonly, CC is based on solid-phase chemistry whereby sequential reactions are carried out on initial building blocks that are linked to a resin until synthesis is complete. Solution-phase chemistry is increasingly favoured, producing greater yields and requiring less methodology development than solid-phase chemistry.
|
Describing the technology of combinatorial chemistry, Dr Matthews explained that it involved the generation of compounds by random or controlled assembly of predefined building blocks. The number of compounds that could be created from the permutations of relatively few such blocks was substantial, he said. For example, in a three component system, one million (100 x 100 x 100) compounds could be generated from a set of 100 building blocks. Describing the methodology, he commented that in the early days a "tea bag" method had been used whereby a resin, held in a solvent-resistant bag, was sequentially exposed to activated amino acids. [At its inception, combinatorial chemistry was predominantly applied to peptides and nucleic acids.] Other methods had now been developed, including those that encoded the identity of every sample by electronic tagging. In general, there had been a move away from the preparation and screening of compound mixes to the rapid and robotised production, analysis and testing of single molecules.
The diversity of the compounds synthesised was always a consideration, commented Dr Matthews. Structural variation could be achieved through variations of compound templates, backbones (in the case of peptides or peptide analogues) and side-chain functional groups. However, indiscriminate synthesis just increased the cost of drug discovery. Lower numbers of high diversity compounds were cheaper to screen, and perhaps just as likely to produce a "hit" as a larger set of compounds of lower diversity. In addition, building knowledge into library molecules to increase, for example, oral absorption and decrease undesirable effects, could further increase the efficiency of drug discovery.
Looking to the future, Dr Matthews commented that the new techniques "would shorten the gap to phase I trials". Further advances in high throughput screening, target validation and, in particular, in database mining, would enable drug discovery teams to keep up with the vast number of molecules synthesised. Drug development would also be accelerated by early "DME" studies (ie, studies of distribution, metabolism and elimination), high throughput toxicology testing and process miniaturisation.
Combinatorial chemistry at work
The application of combinatorial chemistry in drug discovery was further discussed by Dr Richard Morphy (section head of medicinal chemistry, Organon Laboratories, Lanarkshire). At Organon, he said, combinatorial chemistry was used both for "lead discovery" and "lead optimisation". For lead generation, large and structurally diverse libraries were preferred, whereas for optimisation, smaller, less diverse libraries were used to focus on existing leads. The current trend was away from sheer magnitude and towards developing libraries of better quality composed of more drug-like molecules. |
Richard Morphy: Libraries of compounds synthesised |
As an example, Dr Morphy described a lead discovery programme for drugs with central nervous system indications, such as anaesthesia, analgesia, psychosis and depression. It was no accident, he said, that 75 per cent of CNS active drugs contained a basic nitrogen. Dependent on pH, this functionality presented as a water-soluble salt, or a lipid-soluble base. Thus, a solid-phase combinatorial chemistry system, based on Hoffman elimination, was designed for the synthesis of tertiary amine containing compounds. The approach had paid off with a number of interesting hits found after screening. Moreover, the majority of the 3,000 compounds that now formed the library had appropriate partition coefficients for CNS penetration.
In a second example, Dr Morphy illustrated how combinatorial chemistry could be used in lead optimisation. He said that selective d-opioid agonists were under development as postoperative analgesics as they would avoid the unwanted side effects seen with m-opioid receptor activity. Using a cyclic-imide lead compound which had both good and bad qualities (ie, high selectivity and activity, 98nM, with poor aqueous solubility), a programme of solid-phase synthesis was undertaken to make the compound and a series of analogues to form a library. Within the 50-compound library produced, two analogues were found to have increased activity (6nM range). However, as solubility was not improved, a second library was generated to concentrate on this feature. Of 312 compounds, one lead had both good activity (23nM) and formed a water soluble salt. It was encouraging to find that the compound was active in the formalin test for analgesia, said Dr Morphy.
Three advances had led to considerable improvements in the productivity and efficiency of high throughput screening (HTS), said Dr Gary Allenby (senior scientist, lead discovery, Glaxo Wellcome). These advances - automation, miniaturisation and bead screening - had led to an increase in the quantity and, more importantly, in the quality of the data generated, he said. The drive to improve productivity had been motivated partly by the increasing cost of screening over the past 10 years, rising from between 5 and 8 per cent to 15 per cent of the research and development budget. This had been coupled with an overall reduction in market growth. |
Gary Allenby: Pros and cons of automation |
HTS systems needed to be rapid, reliable and relatively inexpensive: technology needed to be robust, and easy to automate and minimise. HTS assays had to be sensitive enough to detect active compounds at very low concentrations.
With regard to automation, Dr Allenby commented that the advent of robotic systems enabled large numbers of compounds to be screened in short periods of time. It was not that the machines always worked more quickly than people, but they could be in constant operation 24 hours per day. HTS platforms ranged in size from small, partly automated, desktop machines costing between £10k and £50k, to fully integrated, robotic systems. The latter, costing around £200k, tended to be operated continuously in dedicated laboratories. There were pros and cons to automation, said Dr Allenby. Undoubtedly, productivity and data quality were significantly improved, but there were cost considerations. In particular, staff had to be trained, technical experts employed, and it was often necessary to involve external contractors to maintain maximum system efficiency. In addition, when less stable reagents were used, regular, even hourly, manual intervention could still be necessary.
Miniaturisation of HTS systems had several advantages, said Dr Allenby. These included the ability to screen even greater numbers of compounds, reduced labour costs, savings in laboratory space, and reductions in the amount of raw materials and in subsequent waste. Glaxo Wellcome had introduced the use of miniaturised plates, each containing 1,536 wells. With these plates, 50,000 compounds could be screened in three hours using assays based on total reaction volumes of 1-10mL. Different technologies, based on syringe and piston systems, had been developed that could handle such small volumes accurately and rapidly. Advanced detection systems for reading the assay plates had also been developed and included charge-coupled devices (CCD) and photomultiplier-based (PMT) devices.
The third advance, bead-based screening, involved testing compounds following synthesis in bead-based combinatorial chemistry systems, explained Dr Allenby. This method involved solid-phase synthesis of a library of compounds on beads in which compound molecules were each linked to a bead by either light- or acid-labile linkers. Once synthesis was complete, the beads were pooled into different assay wells and half of the compound per bead was cleaved with acid. Beads in wells showing biological activity were then isolated into individual wells and, following exposure to light, the cleaved compound retested. The advantage of this technique, suggested Dr Allenby, was that large compound stores and electronic tracking of compounds were less necessary.
The pressure on the pharmaceutical industry to accelerate the production of new lead compounds had created a new problem of how to deal with all the data created, said Dr Stefan Güssregen (head of molecular design, Tripos Receptor Research, Bude).
|
Stefan Güssregen: How to deal with the data |
Dr Güssregen described the approach taken by his company to achieve such system integration. Tripos had developed a software medium, Chemspace, for handling the design of compound libraries. This software facilitated the application of chemical and project knowledge to "virtual libraries" of synthetically accessible compounds (generally 108-1012 in number) and enabled the selection of sets of compounds for synthesis (typically 103-104 compounds) that had been filtered for certain physicochemical or toxicological attributes. The diversity of compounds within the set could be varied dependent on whether the project required a general or a more focused screening approach. If a "hit" was identified, the original library could be searched again to look for related molecules. Such "library hopping" was facilitated in the Tripos system by "shape similarity searching", whereby a shape descriptor employing structural knowledge (eg, that of a side chain) was used to identify near structural neighbours.
It was essential that the information generated at the library design stage was integrated with that produced during the rest of the drug discovery process. With this in mind, Tripos had developed the Cheminfo system. This was inventory software that integrated all drug discovery aspects, including reagent ordering. Every compound synthesised was assigned a barcode identity and this information, together with a full synthetic history, was stored with analytical data and HTS assay results. Concluding, Dr Güssregen emphasised that an information technology backbone linking HTS, data analysis, and compound design and handling, was the key prerequisite for successful drug discovery.