Advancing Academic Drug Discovery
A Collaborative Model Involving Two Germany-based Organizations
- Fig. 1: Representation of the chemical space of the European ScreeningPort (green) and ChemBioNet (red) screening collections in comparison to the WDI (blue) . 32 P_VSA descriptors (Labute P (2000) A widely applicable set of descriptors. J Mol Graph Model 18:464-477) were calculated and then a principal component analysis was performed.
- Table 1: Comparison of the European ScreeningPort and ChemBioNet compound libraries. Searches for exact matches were performed with the default parameters of Instant JChem, whereas dissimilarity searches were based on binary fingerprints with an allowed maximum dissimilarity (distance) of 0.1.
- Jörn Lenz, Philip Gribbon and Mira Grättinger, European ScreeningPort
- Ronald Frank, Jens Peter von Kries, Ronald Kühne, Jörg Rademann, ChemBioNet
Alliances between academia and industry are important sources of innovation in Drug Discovery and this type of working model is expected to be of critical importance in efforts to replenish the depleted pipelines of the Pharmaceutical Industry. As part of these efforts, the German initiative for Chemical Biology, ChemBioNet and the European ScreeningPort (ESP), have established a Drug Discovery collaboration based on a common small molecule screening collection. This article discusses the properties of the common collection and how it will be used to help deliver value for the two organizations, their academic collaborators and the associated funding bodies.
The related fields of Chemical Biology and Academic Drug Discovery share many common aspects in terms of technologies applied, skills sets needed and procedures adopted. Chemical Biology is a quintessentially interdisciplinary discipline, focusing on identification and scientific exploitation of small molecule based chemical probes and their effects on the functions of genes, pathways and cells. Typically, these functions may be already linked to known disease processes, and there are many instances where novel therapeutic agents have been identified as an indirect, or even direct, output from academic Chemical Biology projects. In Academic Drug Discovery, activities span from target identification and validation, through to clinical proof of concept studies, activities which for reasons of high risk and cost have historically been conducted within the Pharmaceutical and Biotech industries.
Within Europe, Chemical Biology research infrastructures exist as part of a multi-layered framework. In Germany, at the institutional level there are organizations such as the FMP in Berlin partnering with Helmholtz-Research Centers in Berlin and Braunschweig which founded the ChemBioNet, a consortium offering expertise and technology platforms for open access to academic researchers and SME. At the national level, the ChemBioNet organization links well equipped Chemical Biology facilities within Germany. Other facilities are operated by the Max-Plank Chemical Genomics Centre in Dortmund and the EMBL in collaboration with the DKFZ in Heidelberg.
There are also well developed plans to initiate a European-wide Chemical Biology capability, EU-OPENSCREEN , coordinated by the ChemBioNet. In the area of Academic Drug Discovery, multiple initiatives are being initiated , although these are predominantly institutionally based with generally restricted access and are more likely to involve the traditional "heavyweight" research driven organizations with their long track records of collaborations with major Pharmaceutical companies, who are now looking to retain a greater proportion of the perceived value from basic research output of their organizations. Examples include the European ScreeningPort (open access), the Lead Discovery Centre of the Max Plank Society, the Dundee Drug Discovery Centre and the Drug Discovery Facility at Imperial College. In the UK, some of these institutions have "self-organized" into the UK Drug Discovery Consortium  with the aim of influencing external stakeholders such as funding bodies and reagent and service suppliers, as well as creating a common platform from which to engage the Pharmaceutical industry. On the Europe wide scale, preliminary attempts to address this problem are in place under the EATRIS program . Worldwide, the Society of Biomolecular Screening's Academic Screening Facilities Directory, lists over 60 Universities in 10 countries with advertised screening capabilities and a declared interest in Academic Drug Discovery .
Although the ultimate objectives of Chemical Biology and Academic Drug Discovery may remain distinct, there are still many opportunities for effective collaboration between the disciplines. In this article we report on one such collaboration which has been conceived to create tangible benefit in reducing redundancy in terms of basic science and capabilities and thus maximizing the eventual return on investment for funding stakeholders and ultimately patients. The two entities involved are the European ScreeningPort and the ChemBioNet.
Description of the Organizations and Their Stakeholders
The ChemBioNet consortium offers compound logistics; small molecule screening platforms; genome-wide RNAi screening for drug target identification and validation; and a broad range of chem- and bioinformatic expertise to academia and SME industry. Currently, the screening unit, the medicinal chemistry group and the drug design group of the Leibniz-Institut für Molekulare Pharmakologie (FMP) and the Department of Chemical Biology at the Helmholtz-Zentrum für Infektionsforschung (HZI) in collaboration with the Max-Delbrück Zentrum für Molekulare Medizin (MDC) provide coordinated support to the European research community in the area of Chemical Biology. Both institutions represent the core facilities of the ChemBioNet. Recently, the University of Oslo has participated in enlargement of the ChemBioNet compound collection. ChemBioNet projects are not exclusively restricted to addressing disease relevant targets.
European ScreeningPort is a Public Private Partnership which receives infrastructure funding from SME and regional government organizations. In conjunction with partners, it offers Pharma scale small molecule hit-finding capability as a fee-for-service. ESP provides expertise in assay development, access to high quality compound libraries, state-of-the art HTS facilities and bioinformatics solutions.
Both organizations prosecute a broad range of screening projects involving novel or under-exploited targets. Funding originates from individual sponsorship by institutions (e.g. Universities, medical research foundations etc.), grant calls at the national level (e.g. BMBF in Germany, Medical Research Council in the UK) and at the European level. The ESP and the ChemBioNet share many common stakeholders amongst this collection of funding bodies. Therefore, there is a strong incentive for ChemBioNet and ESP to actively collaborate and build on their distinct strengths to deliver optimal solutions for the academic community.
Properties of the Shared Library
The first step of the practical realization of the collaboration has been the creation of a commonly accessible shared small molecule screening library selected from each organization's pre-existing collections. We analyzed whether a combination of both chemical libraries would be beneficial for incoming screening projects and thus compared their chemical descriptors. The compound collection currently available at European ScreeningPort can be divided into two major categories. The first one contains approximately 2.500 small organic compounds (ESP_1), the second of about 2.500 natural analogues and natural compounds (ESP_2.1, ESP_2.2 and ESP_2.3). A subset of 17,000 diverse compounds designed from analysis of the World Drug Index (WDI)  which comprise part of the ChemBioNet screening library, was selected. Table 1 shows the results of the comparisons of the two libraries that were carried out using Instant JChem as well as the tools provided by ChemAxon . Only 9 identical structures were found and based on the average dissimilarity in the range of 0,55 to 0,74, the libraries appear to be complementary in make-up.
ChemAxon's LibMCS , which is based on a hierarchical maximum common substructure (MCS) search, was used for diversity analysis and clustering of compounds in each library. Using a minimal MCS size of six, a total of 578 clusters in the ChemBioNet library were identified, 96 in ESP_1, 68 in ESP_2.1, 117 in ESP_2.2 and 214 in ESP_2.3. After merging the ChemBioNet and ESP libraries, a total of about 550 clusters were identified using the same minimal MCS size. Though being fewer in overall number, the resulting structural density of the 550 clusters was increased. Additionally, both collections were compared in terms of their distribution within chemical space. The scatter plots of the first five principal components are shown in figure 1. The diagonal shows the Gaussian shape histograms of the principal components. The overlay demonstrates that the collections of both organizations are within the chemical space of WDI bioactive compounds, (blue), although they do not fully encompass the WDI distribution. Both collections have overlapping regions as well as unique regions. Thus the combination of both collections allows us to offer a broader chemical spectrum to our academic partners entering a screening campaign.
Academia based research expertise has the potential to address some of the gaps in the Pharmaceutical Industry pipeline by enabling broad identification and validation of novel drug targets, which are currently not addressed by industry for reasons of cost or perceived complexity. The broad scope of projects within the ChemBioNet, which are selected on scientific as opposed to commercial grounds, and the industry focused disease oriented approach adopted by the ESP are highly complementary. Both partners are at the beginning of the process of establishing the effective transference of academic and industrial know-how. It is anticipated that intelligent usage of the shared screening library described in this article will, as a minimum, assist in identifying novel "druggable" targets. For the ESP, the use of the library will be a valuable tool to assess the likelihood that novel targets are amenable to small molecule drug discovery, i.e. to answer the question: is this target "druggable"? This would then allow for more effective decision making when deciding if particular targets should progress through to Pharma scale hit and lead finding stages, which are both costly and resource intensive. Finally, as an additional key benefit, we expect the implementation of this new resource should help catalyze the setting of European standards in Academic Drug Discovery within a worldwide setting.
 Gribbon P.: Drug Discovery Today, in Press (2009)
 Derwent World Drug Index, Release 2005, Derwent Information Ltd., London, www.daylight.com/products/wdi.html
 Instant JChem, version 2.4.2, 2008, ChemAxon, 1037 Budapest, Hungary, www.chemaxon.com
 LibMCS, version 0.7, ChemAxon, 1037 Budapest, Hungary, www.chemaxon.com
Mira Grättinger, Vice President
Philip Gribbon, Chief Scientific Officer
Jörn Lenz, Head of Discovery Informatics
Ronald Frank, Helmholtz-Zentrum für Infektionsforschung mbH (HZI)
Jens Peter von Kries, Head of Screening Unit; Ronald Kühne, Head of Drug Design Group; Jörg Rademann, Head of Medicinal Chemistry
Leibniz-Institut für Molekulare Pharmakologie (FMP) im Forschungsverbund Berlin e.V.