Poster 10: In Silico Structure-Based Drug Design (SBDD) on S100P: Identification of a lead compound towards a therapeutic intervention in pancreatic cancerRamatoulie Camara1, Sharon Rossiter1, Vishal Kholi1, Stewart B. Kirton1
|1Department of Pharmacy, School of Life & Medical Sciences, University of Hertfordshire, Hatfield, Herts, AL10 9AB, UK.|
|S100P is a 95-amino acid calcium-binding protein first identified and isolated from human placenta. S100P exerts its effect both intracellularly through calcium ion modulation and extracellularly through binding to RAGE, the receptor for advanced glycation end-products. The protein has been reported to be over expressed in many cancers but despite an abundance of structural information, not much has been done to elucidate its binding site and postulate how it interacts with cromolyn, a ligand shown to inhibit its interaction with RAGE .|
We report on the computational studies carried out on the NMR ensemble of S100P (PDB ID 1OZO) in which we identify potential binding sites, dock cromolyn to these sites and generate a pharmacophore for virtual screening of databases of lead-like compounds.
The S100P NMR ensemble contains sixteen conformers. These were clustered into conformationally-related subfamilies to determine the most representative model from each cluster using the program OLDERADO . Using four different pocket-detecting algorithms (Pocket-Finder, Q-SiteFinder, Fpocket and Site Finder), potential binding sites were identified. Focus was placed on those cavities found at the dimeric interface, which has been reported to be necessary for target recognition in S100 proteins . Using MOE, the known inhibitor of the S100P-RAGE interaction cromolyn, was docked to the putative binding sites and the resulting binding interactions between the ligand and protein were used to develop three-point and four-point pharmacophores. These pharmacophores were used to virtually screen the MOE and ZINC databases in an attempt to identify potential lead compounds. Hits were clustered and a representative of each cluster was bought or synthesised and tested against BxP-3 S100P pancreatic cancer expressing cells and non-S100P expressing cells.
Of the sixteen conformers investigated, only two models had pockets at the dimeric interface large enough to accommodate cromolyn. However, docking studies revealed only one conformer was a feasible starting point for SBDD experiments, as it was only in this conformer that the observed interactions between cromolyn and the protein were sufficient to develop three-point and four-point pharmacophores. Virtual screening of the MOE database of lead-like compounds using the three-point pharmacophore resulted in a hit rate of 0.008%. Clustering of the hits obtained gave 17 hits representative of each cluster. Virtual screening of the ZINC database (from which compounds identical to those in the MOE database were removed) using the same three-point pharmacophore gave a diverse set of leads from those obtained from the MOE database screening. Biological screening of the 17 representative hit compounds against BxP-3 S100P pancreatic cancer expressing cells and non-S100P expressing cells identified five compounds with an equal or higher potency against the former compared to cromolyn at the same concentration or less and no significant effect on the latter.
Hence, by using computational methods, we have been able to show that only one of the publicly available S100P NMR conformers is a credible candidate for SBDD. Docking studies showed key interactions between the ligand and the putative active site which led to the development of pharmacophores used in the virtual screening of lead-like databases. Early indications show five of the leads identified as being active against S100P expressing pancreatic cancer cells.
 Arumugam, T., V. Ramachandran and C.D. Logsdon (2006). Effect of cromolyn on S100P interactions with RAGE and pancreatic cancer growth and invasion in mouse models. J. Natl. Cancer Inst., 98(24) 1806-18.
 Kohli, V. (2011). Where to begin? Computational analysis of public domain S100P structures, a protein implicated in pancreatic cancer, to determine appropriate starting points for structure-based drug design . Unpublished Masters thesis, School of Pharmacy, University of Hertfordshire, Hatfield, AL10 9AB.
 Réty S., et al.,(2000). Structural basis of the Ca2+-dependent association between S100C (S100A11) and its target, the N-terminal part of annexin I. Structure, 8(2) 175-84.