The Pharmacoinformatics Research Group

Originally established as Emerging Field Pharmacoinformatics at the Faculty of Life Sciences at the University of Vienna, the Emerging Field developed towards a full professorship of Pharmacoinformatics. The international team, headed by Prof. Dr. Gerhard Ecker, follows a holistic approach, combining multi-dimensional annotation, structural modeling of biomolecular systems, structure-based drug design, chemometric and in silico chemogenomic methods, statistical modeling and machine learning approaches, to develop predictive computational models. The experimental validation and optimisation of the obtained in silico models by strong links to experimental groups is an integral part of these activities. Currently the targets under investigation include several ABC-transporter (ABCB₁, ABCC₂, ABCB₁₁, ABCG₂), the serotonin and dopamine transporter, the GABA transporters, the hERG potassium channel, the TRPV₁ channel, the GABA_A receptor, and the insulin receptor. On a methodological basis, the group works on development of new, similarity-based descriptors and the application of unsupervised neural networks for hit-identification.

Ligand-based Studies on ABC-transporters

The main line of research is along the molecular basis of drug-transporter interaction. This started in 1993 with the discovery that the class 1C antiarrhythmic agent propafenone is able to inhibit the multidrug transporter P-glycoprotein (P-gp, ABCB₁). P-gp transports a wide variety of functionally and structurally diverse cytostatic agents out of tumor cells, thus preventing accumulation of these drugs within the cells. Inhibition of the transport leads to resensitisation of multidrug resistant cells and thus presents a useful concept for therapy of resistant tumors. Analogous transport systems were also identified in bacteria and fungi. In a strong collaboration with Peter Chiba from the Medical University of Vienna we started a lead optimization program using mainly ₂D- and ₃D-QSAR studies, which resulted in several tool compounds with potency in the nanomolar range. Besides standard techniques such as Hansch-analysis, Free-Wilson analysis, CoMFA, and GRIND, the group also explores the potential of artificial neural networks, i.e. feedforward backpropagation networks and self organizing maps. The latter have also been successfully applied for scaffold hopping and virtual screening, both for identification of new P-gp inhibitors and insulin receptor activators, as well as for classification of hERG channel blockers.

P-glycoprotein and related transporters are also strongly influencing ADME properties, such as intestinal absorption, blood-brain barrier permeation, and canalicular excretion. Thus, substrates of P-gp and related transporters might face the risk of poor bioavailability, while inhibitors may lead to drug/drug interactions and drug induced cholestasis. The group is thus also actively engaged in developing classification models for prediction of P-gp substrates and inhibitors. In order to mimic real life scenarios, where the number of inactives by far exceeds the number of actives (100:1 to 1000:1), machine learning methods such as random forest and support vector machines, both combined with cost sensitive bagging, are applied for classification of large data sets.

Experimental Data Guided Ligand Docking

One major focus of the Pharmacoinformatics Research Group is to combine ligand- and structure-based methods. Briefly, the extensive ligand-based in silico studies reveal a clear picture on pharmacophoric sub structures and guide the selection of compounds to be docked into protein homology models of the target under consideration. Subsequent unrestrained docking of these ligands followed by clustering of common scaffolds reveals clusters of binding hypotheses, which are prioritized and validated on basis of the SAR known. As a final prove of concept, top prioritized poses serve as basis for the generation of structure-based pharmacophore models, which are used for screening vendor libraries. Top ranked compounds are ordered and biologically tested. This approach has been successfully applied for identification of new ligands of the benzodiazepine binding site at the GABA-A receptor, inhibitors of P-gp, and inhibitors of the serotonin transporter. The latter two projects are pursued under the framework of the national research cluster SFB₃₅ – “Transmembrane Transporters in Health and Disease” (www.sfb35.at).

Current and Future Challenges

With the increased understanding of physiological and pathophysiological processes driven by systems biology, pharmacoinformatics will move towards pathway driven drug discovery and integrated approaches. First attempts made in our group focus on prediction of cns related side effects of antidepressant agents based on their interaction profiles with a set of 30 transporters and receptors expressed in the cns. Another challenge, the prediction of the outcome of a 28 day rat toxicity study, is pursued under the framework of eTOX, a European project funded by the Innovative Medicines Initative (IMI; www.imi.europa.eu). Finally, by coordinating the Open PHACTS project (www.openphacts.org), the group made a strong move towards semantic integration of public and private data sources relevant for drug discovery. Open PHACTS is also funded by the IMI and aims at developing an Open Pharmacological Space utilising sematic web technology based on rdf triples and nanopublications. This will allow to answer questions such as “give me all compounds which have been annotated with liver toxicity and the interaction profile of these compounds with all transporters expressed in the liver” within a few minutes.

About Gerhard Ecker

Gerhard Ecker is Professor for Pharmacoinformatics and Head of the Pharmacoinformatics Research Group at the Department of Medicinal Chemistry, University of Vienna. He also coordinates the research focus “Computational Life Sciences” of the Faculty of Life Sciences. Gerhard received his doctorate in natural sciences from the University of Vienna and performed his post-doctoral training at the group of J. Seydel in Borstel (Germany). He has published more than 100 articles mainly related to SAR and QSAR studies on P-glycoprotein (P-gp), edited 3 books and gave more than 100 invited lectures. Gerhard is Editor of Molecular Informatics and member of Editorial Advisory Boards and Editorial Boards of several journals. In the years 2009 – 2011 he served as President of the European Federation for Medicinal Chemistry (EFMC), and from 2012 on he chairs the Strategic Advisory Board of EFMC. Finally, he is also strongly engaged in national and international educational activities, such as the national PhD network “Molecular drug targets” (moltag.univie.ac.at) and the EUROPIN PhD Programme in Pharmacoinformatics (www.europin.at), which comprises seven top groups in the field of pharmacoinformatics from 6 European countries.

_____________________________________________________

Contact

Prof. Gerhard F. Ecker
Dept. Medicinal Chemistry, Univ. Vienna
Althanstrasse 14, 1090 wien, Austria
Phone: +43-1-4277-55110
Fax: +43-1-4277-9551
Email: gerhard.f.ecker@univie.ac.at
Lab's website: pharminfo.univie.ac.at

_____________________________________________________

Key publications