Research Papers:

Large scale integration of drug-target information reveals poly-pharmacological drug action mechanisms in tumor cell line growth inhibition assays

Richard Knight, Mikhail Gostev, Sergei Ilisavskii, Anne E Willis, Gerry Melino and Alexey V Antonov _

PDF  |  HTML  |  Supplementary Files  |  How to cite  |  Order a Reprint

Oncotarget. 2014; 5:659-666. https://doi.org/10.18632/oncotarget.1597

Metrics: PDF 1575 views  |   HTML 2007 views  |   ?  


Richard A. Knight1, Mikhail Gostev2, Sergei Ilisavskii2, Anne E. Willis1, Gerry Melino1,2,3, Alexey V. Antonov1

1 Medical Research Council Toxicology Unit, Leicester LE1 9HN, UK

2 Molecular Pharmacology Laboratory, Technological University, St-Petersburg, Russia.

3 Department of Experimental Medicine & Surgery, University of Rome “Tor Vergata”, 00133 Rome, Italy

Corresponding Author:

Alexey V. Antonov, e-mail: aa668@le.ac.uk

Received: November 12, 2013      Accepted: December 12, 2013      Published: January 21, 2014


Understanding therapeutic mechanisms of drug anticancer cytotoxicity represents a key challenge in preclinical testing. Here we have performed a meta-analysis of publicly available tumor cell line growth inhibition assays (~ 70 assays from 6 independent experimental groups covering ~ 500 000 molecules) with the primary goal of understanding molecular therapeutic mechanisms of cancer cytotoxicity. To implement this we have collected currently available information on protein targets for molecules that were tested in the assays. We used a statistical methodology to identify protein targets overrepresented among molecules exhibiting cancer cytotoxicity with the particular focus of identifying overrepresented patterns consisting of several proteins (i.e. proteins “A” and “B” and “C”). Our analysis demonstrates that targeting individual proteins can result in a significant increase (up to 50-fold) of the observed odds for a molecule to be an efficient inhibitor of tumour cell line growth. However, further insight into potential molecular mechanisms reveals a multi-target mode of action: targeting a pattern of several proteins drastically increases the observed odds (up to 500-fold) for a molecule to be tumour cytotoxic. In contrast, molecules targeting only one protein but not targeting an additional set of proteins tend to be nontoxic. Our findings support a poly-pharmacology drug discovery paradigm, demonstrating that anticancer cytotoxicity is a product, in most cases, of multi-target mode of drug action.

Creative Commons License All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 License.
PII: 1597