Computational methods grow in importance in the field of medicinal chemistry. In the last decade, virtual screening has become a routine procedure in the process of hit identification on various therapeutic targets. Generally, a compound library represents a starting point of each virtual screening, while the quality of the former is crucial for the success of the latter. Our work was focused on a design of a hierarchical protocol for the construction of high-quality, filtered and enriched compound libraries, which served as starting points for subsequent structure- and ligand-based virtual screening campaigns. Virtual screening methods were applied in the discovery of novel hit compounds on various therapeutically important targets, such as receptor TLR4 and enzymes proteasome, immunoproteasome and butyrylcholinesterase.
Receptor TLR4 in complex with its accessory protein MD-2 specifically binds a cell-wall component of gram-negative bacteria (lipopolysaccharide), resulting in a cytokine release and immune cell activation, which is of crucial importance for the defence against infections. Given the fact that the dysregulation of TLR4 signaling has been directly implicated in an array of acute and chronic human diseases, receptor TLR4 represents an attractive target for the treatment of severe sepsis and neuropathic pain. In the process of the discovery of novel TLR4 antagonists, ligand-based virtual screening was performed which considered the structural features of known beta-amino alcohol derivative. Simultaneously, structure-based virtual screening of enriched compound library was carried out aiming to disturb protein-protein interactions between TLR4 and MD-2. Parallel virtual screening approach allowed us to identify three novel compounds with promising TLR4 antagonistic activities with IC50 values in micromolar range. These compounds exhibited immunomodulatory properities with the ability to suppress the cytokine secretion by human peripheral blood mononuclear cells. The specific affinity of the most potent hit (IC50 = 16.6 μM) was confirmed by surface plasmon resonance direct-binding experiments. The results of our study represent a very promising starting point for the development of novel small-molecule antagonists of TLR4. Proteasome is a multi-protein complex, which plays an important role in nonlysosomal protein degradation. Based on the structure and site of expression, three major classes of proteasomes can be identified: constitutive, immuno- and thymoproteasome. Proteasome inhibitors display important anti-inflamatory and antitumor activities via nuclear factor Nf-κB inhibition and induction of apoptosis of rapidly dividing cells. 3D ligand-based virtual screening was accomplished using the structure of aminopiperidine ligand with previously described selective cytotoxicity for Burkitt’s lymphoma cells through proteasome inhibition. Our aim was to discover ligands with improved potency and different scaffold compared to the query ligand. Virtual screening led to the identification of the chemically different compound exhibiting similar spatial conformation and improved selectivity and potency (IC50 = 6.7 μM) towards Burkitt’s lymphoma cells. Mechanism-of-action studies revealed different target profile in comparison with the previous query ligand, which does not involve the inhibition of the proteasome or the NFκB pathway. Recent studies revealed that the specific immunoproteasome inhibition represents an emerging strategy for the treatment of autoimmune disorders, such as rheumatoid arthritis, inflammatory bowel disease and lupus. Given the fact that non-peptide and reversible immunoproteasome inhibitors are rare, structure-based virtual screening was performed targeting β5i subunit. A psoralen derivative with marked specifity and selectivity towards β5i subunit of immunoproteasome (Ki = 15 μM) was identified, whereas no inhibition of constitutive proteasome was observed. Structure-activity relationship studies led us to identify compound 37 with improved potency (Ki = 1 μM) and lower molecular weight. Docking simulations of the psoralen derivatives proposed the incorporation of electrophilic moiety in proximity of the terminal –COOH functional group, which allowed us to recognize specific covalent inhibitors of immunoproteasome with activity in nanomolar range. Alzheimer’s disease is a progressive neurodegenerative disorder of central nervous system, which lead to neuron death and loss of cognitive functions. Enzyme BChE is regarded as a promising drug target as its levels and activity significantly increase in the late stages of Alzheimer’s disease. To discover novel BChE inhibitors, we used a hierarchical virtual screening protocol, which allowed us to identify three compounds with significant inhibitory activities against BChE. The most potent compound 1 (IC50 = 21.3 nM) was resynthesized through orthogonally protected piperidin-3-ylmethanamine and resolved into its pure enantiomers exposing significant stereoselective activity and a dissociation constant of 2.7 nM for the most potent stereoisomer (+)-1. The crystal structure of human BChE in complex with compound (+)-1 was solved, revealing the binding mode and providing clues for structure-based optimization of piperidine derivatives. Sulfonamide analogue 4 diplayed improved potency (IC50 = 4.95 nM) and lower degree of stereoselectivity, while the solved complex crystal structure BChE-inhibitor exposed distinctive mode of binding. Compound 4 exhibited positive effects in scopolamine model of cognitive deficit in mice in vivo. Through methoxyetyl chain modification of compound 1, one additional basic centre was incorporated in the structure, which enabled cation-π interactions in choline-binding pocket. The resulting compound GUK942 is the most potent inhibitor in this seris with dissociation constant in low picomolar concentration range. In this work, Koelle's histochemical staining method for cholinesterase activity detection on brain slices was optimized. Optimized Koelle's method allowed us to identify specific brain regions with highest BChE activity, which were further used to prove the ability of piperidine derivatives (compounds 1 and 4) to inhibit brain BChE on rat brain slices ex vivo. Compound 1 was further used to validate the usefulness of a novel software LiSiCA, which examines 2D and 3D similarity on the basis of clique algorithm. LiSiCA was demonstrated to be an effective tool in drug discovery, as it enabled the discovery of five new inhibitors of BChE with potency in nanomolar range. All hits possess different scaffolds and lower molecular weight, if compared to the compound 1, thus proving the ability of LiSiCA for scaffold hopping. Amyloid plaques are one of the hallmarks of Alzheimer's disease, while amyloid beta species are linked to the neurotoxicity. Compounds capable of inhibiting amiloid beta self-induced aggregation might exhibit neuroprotective effect and influence the progress of Alzheimer's disease. Studies revealed, that our novel cholinesterase inhibitors affected the amyloid beta self-induced aggregation into fibrils. Compound 1, which is a selective nanomolar inhibitor of BChE, displayed 61.7% inhibition of aggregation at 10 μM, while compound 13b as selective nanomolar inhibitor of AChE exhibited 65.96% inhibition at the same concentration. Both compounds possessed neuroprotective effect and protected SH-SY5Y cell-line against amlyoid beta species toxicity. Considering the important role of metal ions in progression of Alzheimer's disease, functional groups with the ability to chelate the divalent metal ions (Cu, Zn and Fe) were incorporated in the structures of the piperidine inhibitors, whereas the BChE activity and effect on amyloid beta self-induced aggregation was retained. Compound 8g showed selective nanomolar inhibition of BChE (IC50 = 215 nM), Cu2+ selective complexation, and the ability to inhibit amyloid beta self-induced aggregation at 10 μM. The resolved crystal structure of 8g in complex with BChE revealed the binding of metal chelating nitroxoline moiety in acyl-binding pocket of the enzyme. Nitroxoline derivatives thus represent an encouraging starting point for further development of multifunctional candidates for the treatment of AD.
|
