Abstract Details

Poster 12: Design, Synthesis and Biological Testing of New Antimalarial Drug Candidates

Robert D. Clark1, Michael S. Lawless1, Walter S. Woltosz1
1Simulations Plus, Inc.
The World Health Organization has estimated that over 200 million people suffered from malaria in 2010 and that over 600,000 people died from it that year [1]. Growing problems with resistance to existing anti-malarial drugs makes identification of new drugs a high priority. We applied a series of state-of-the-art in silico tools to publicly available activity data from screens carried out on intact Plasmodium falciparum parasites to generate a handful of candidate molecules predicted to combine potency with good absorption, distribution, metabolism, excretion and toxicity (ADMET) properties.

Here we will describe the overall process used to design these molecules and provide a status report on the compounds’ activity against the parasite as well as how their experimentally determined ADMET properties compared to the corresponding predicted values.

We applied modeling tools in ADMET Predictor™ to literature data to train predictive artificial neural net ensemble (ANNE) models of the quantitative structure-activity relationships (QSARs) for two potential target enzymes: the cyclin-dependent kinase Pfmrk1 and dihydroorotate dehydrogenase (PfDHODH). Those models were then used to predict inhibitory potencies against both targets for all 13,533 compounds identified as active in a blood culture screen recently made public by GlaxoSmithKline (Pubchem biossay AID 2306) [2]. Applying these models and the substructure-based class generation technology in MedChem Studio™ to data from AID 2306, we identified one particular class of chemistry as likely to inhibit PfDHODH. Combinatorial elaboration of that class followed by in silico screening based on a wide variety of ADMET liabilities produced a set of candidate molecules that were further elaborated to include substituents not reported by GSK. Subsequent evaluation of likely dosage requirements by pharmacokinetic simulation using GastroPlus™ identified several novel molecular structures with attractive predicted efficacy and ADMET profiles. Seven candidates were synthesized and tested along with three precursors for biological activity as well as for some of their ADMET and pharmacokinetic properties.

All seven candidates exhibited half-maximal growth inhibition at concentrations at or below 10?M when assayed against wild-type parasites in blood culture; and five had XC50s below 1?M. Two had XC50s below 100nM, which was better than any example of the class reported by GSK. Although all were less potent when measured against chloroquine-resistant strains of the parasite, the two most active analogs still exhibited sub-micromolar XC50s.

pKa, log D, solubility and cytochrome P450 (CYP) susceptibilities for the candidates were well-predicted, and although the predicted rank-order of in silico estimates of half-life were confirmed by measurements in human liver microsomes. Nonetheless, half-lives were too short for any of them to be clinical candidates. The in silico design methods were remarkably successful, however, in identifying and advancing a promising area of antimalarial chemistry with a very limited expenditure of time and money.

This proof-of-concept demonstration showed that these in silico tools could be applied to phenotypic screening data to efficiently design novel molecules active against a selected target and with overall good ADMET properties. The seven candidates all exhibited good-to-excellent potencies in culture and hit all tested ADMET targets other than CYP 3A4 stability. More work remains to be done before compounds from this structural class might make it to the clinic, but preliminary analysis suggests a number of next-generation analogs whose synthesis may address the shortcomings of the compounds prepared to date.

1. World Malaria Report 2011 (see http://www.who.int/malaria/world_malaria_report_2011/wmr2011_ summary_keypoints.pdf)
2. F-J Gamo, LM Sanz, J Vidal, C de Cozar, E Alvarez, J-L Lavandera, DE Vanderwall, DVS Green, V Kumar, S Hasan, JR Brown, CE Peishoff, LR Cardon, JF Garcia-Bustos. Nature, 2010, 465, 305-310.

Return to Programme