The rise, fall and revival of combinatorial chemistry

Looking back over the past 25 years, we see that combinatorial chemistry significantly increased the number of screening compounds available for drug discovery research, contributed to the development of organic synthesis as well as chemoinformatics, and accelerated technical innovation in areas including automation, purification, analytics, and logistics. Last but not least, combinatorial thinking motivated the shift from the traditional single-compound design approach to a multiple-compound (library) design process.

Combinatorial chemistry emerged around 30 years ago. In the 1990s, drug discovery was facing a long-lasting crisis: Despite significant increases in R&D expenditures, the number of approved new chemical entities (NCEs) was declining. Combinatorial chemistry offered a remedy: Increasing the number of screening compounds would make it possible to cover a higher proportion of the molecular space.

Other factors also contributed to the birth of combinatorial chemistry: Studies of the DNA sequences of the human genome resulted in the identification of 5,000 to 10,000 novel proteins; among these, roughly 1,000 to 3,000 would serve as novel therapeutic targets. In order to identify small molecule modulators to those proteins, large numbers of novel compounds would be required.

In parallel, the widespread adoption of recombinant protein production and the development of high-throughput screening generated a high demand for large numbers of compound collections. Traditional preparative organic chemists could provide only 40 to 50 new compounds per year. New concepts and innovative technologies were required.

Split-mix approach

The roots of combinatorial chemistry can be traced back to solid-phase peptide synthesis. Árpád Furka (Eötvös-Loránd-University, Budapest) invented the peptide mixture synthesis or split-mix approach. In the split-mix approach, polymer beads to which starting compounds are attached are mixed after each coupling step and then redistributed to carry out the following step with new amino acids. In this revolutionary method, peptides were synthesised with strikingly low numbers of chemical steps in every combination of the amino acids within the sequence, forming a combinatorial peptide library. The mixtures are screened, and the active peptides are sequenced. The method was adapted to non-peptide small molecules. However, identification of the active compounds still faced serious difficulties and required complicated additional activities.

While many innovative chemistries were described for the split-mix approach (see Figure), the concept was most successfully adapted to the field of solution-phase organic chemistry, where the use of isolated, parallel reaction vessels resulted in pure compounds. The direct use of many thousands of known classical organic reactions represented another key advantage, and it was suitable for preparing compounds where diverse building blocks were connected in each combination. This approach also allowed high-throughput operation and automation.

By the year 2000, solution-phase parallel synthesis methods dominated the generation of combinatorial libraries. Since then, combinatorial chemistry has reached maturity, and the earlier preference for quantity in libraries shifted towards high quality.

From in vitro to in silico

In the early years, the focus was on technological innovations and the rapid generation of huge libraries, which led to low hit rates in high-throughput screening (HTS) campaigns. Thus, the majority of such libraries provided inactive compounds. Combinatorial chemistry did not help to increase the number of new chemical entities as originally expected, in part due to the high attrition rate (80 %) observed during the discovery process.

This high failure rate was partly attributed to the in vitro nature of HTS technologies that did not take into consideration the compounds‘ pharmacokinetic properties. A new trend has emerged to remove the potentially unsuitable candidates as early as possible in the pipeline – under the motto „fail fast, fail cheap“ – saving costs.

These facts spurred the development of in silico filtering and screening tools, which reduce the number of the synthesisable libraries to only those compounds that meet the requirement. The filtering cascade also involved removing structural redundancies through so-called diversity selection. Diversity is the key term in combinatorial chemistry; it characterises the breadth of coverage and the dissimilarity of the chemical space. There are various approaches: appendage (substituent), functional group, stereochemical, and skeletal diversity. Oral absorption could be estimated by meeting empirical physicochemical parameter ranges defined by Lipinski’s Rule of 5 and the Veber Rules. Furthermore, predictive expert systems were developed to select compounds likely to have favourable „ADME-Tox“ properties (where A = absorption, D = distribution, M = metabolism, E = excretion, and Tox = toxicity).

In order to achieve large diversity with a minimum of synthetic steps, combinatorial chemists devised new strategies or rediscovered old techniques such as multicomponent reactions (Ugi, Passerini, and Biginelli reactions, for example).

Mimicking the large skeletal and stereochemical diversity of the natural products, Schreiber and co-workers introduced diversity-oriented synthesis (DOS), which provided complex structures by applying a general BCP (build, couple, pair) synthetic algorithm.

Hungary‘s role

Over the years Hungarian scientists, universities, pharmaceutical companies and enterprises have contributed significantly to the development of combinatorial chemistry or have provided services.

Among the universities, Eötvös-Loránd-University, Szeged, and Debrecen integrated combinatorial chemistry into their research programs. The pharma companies Chinoin/Sanofi and Richter Gedeon also developed libraries that feed their HTS systems. ComGenex/AMRI/ComInnex has been playing an important role in developing solution-phase parallel synthesis of diverse libraries. By the year 2000, ComGenex had become one of the leading library providers worldwide. Its former subsidiary TargetEx has focused on in silico library selection and in vitro screening, while Vichem has grown to be an international specialist in combinatorial kinase inhibitor discovery.

The European Society for Combinatorial Sciences was established in 2000 with the active participation of Hungarian scientists. Its first official symposium was held in Budapest in 2001, when Árpád Furka was elected as Honorary President. At that time, a bright future was predicted for combinatorial chemistry, which unfortunately has not come true. By the end of the decade it became clear that combinatorial chemistry had not increased the success rate of drug discovery, that costs were continuing to increase, and that the profitability of the pharma companies continued to decrease. The generation of huge libraries has stopped; well-designed small libraries have become predominant, and in practice chemists have partly returned to the classical laboratory techniques. The publication peak of combinatorial chemistry was between 2002 and 2003.

Failure and rebranding

Many analyses appeared on the failure of combinatorial chemistry. Combinatorial chemistry provided huge numbers of compounds that were often redundant and useless. Most suppliers applied the same, relatively easy synthetic steps, leading to similar libraries. While the compounds were largely ineffective, huge investments were devoted to innovative synthesis and screening technologies. The early-phase combinatorial libraries had a number of weak points in common: structural redundancy, low coverage of the chemical space, poor novelty, the fact that the compounds mostly had planar rather than 3D shapes, and the predominance of compounds that were lipophilic and unable to cross the cell membrane.

After all these negative features were recognised, the word “combinatorial” became taboo, and the technology was rebranded as “high-throughput synthesis”.

In the middle of the 2010s, we began witnessing a revival of the technology by way of a return to the original concept – mixture synthesis. The key hurdle of this approach was solved by applying DNA tags and PCR techniques to the amplification of sequences for decoding. Thus, with this innovative combination, it has become possible to synthesise very large compound libraries in very small quantities and then screen them. Hits can be rapidly identified. Though chemistry imposes some limitations, this approach has become popular for primary screening. In addition, flow chemistry has opened up new opportunities for other old concepts, such as dynamic combinatorial chemistry, and has made it possible to integrate drug discovery, synthesis, and screening in one flow. Novel library categories, such as covalent-bond forming and fragment-based libraries, have also been introduced and are being successfully exploited.

The author

György Dormán is a Honorary Professor at University of Szeged at the Faculty of Pharmacy and Budapest University of Technology, at the Faculty of Chemistry. Currently he is a scientific consultant of Targetex and Innostudio. He obtained his Ph. D. in organic and medicinal chemistry from the Technical University of Budapest, Hungary in 1986. He worked at Sanofi – Chinoin, spent a post-doctoral year in the UK, was a Visiting Scientist at the State University of New York, Stony Brook, served ComGenex/AMRI Hungary as a Chief Scientific Officer and was Director of Scientific Innovation at ThalesNano from 2008 until 2015.

Wissenschaft + ForschungSchlaglichtthema: Chemie in Europa