Identification of Novel Acid-Sensing Ion Channel 3 Modulators using in silico Modelling and Screening

Abstract

Nociception is a protective mechanism alerting an organism to noxious stimuli and potential harm. However, dysregulation of the nociceptive system can result in chronic pain, which has a prevalence of approximately 40 % in the adult population. Current therapeutics are often inefficacious, and the growing opioid crisis demonstrates a clear need to develop new analgesics and improved pain management strategies. A common feature of inflammation is tissue acidosis, and the raised extracellular proton concentration can activate acid-sensing ion channel 3 (ASIC3), which is most highly expressed in those sensory neurones tuned to detect noxious stimuli, i.e., nociceptors. To date, the most potent non-peptide ASIC3 inhibitors act at micromolar concentrations in vitro on rat-ASIC3 (rASIC3), but are non-selective, thus inhibiting other ASIC subunits. Intriguingly, certain non-steroidal anti-inflammatory drugs (NSAIDs) directly inhibit rASIC3 but are neither potent ligands nor is it understood precisely how they interact with ASIC3. Here, I aimed to use in silico modelling to not only predict the plausible binding mode of some known ASIC3 modulators to this channel, but to further identify new ligands that can modulate ASIC3. Homology models of rASIC3 were constructed based on published 3D structures of chicken ASIC1a solved at various states. These models were then used for blind docking with some known small molecule modulators of ASIC3 that notably included the NSAID diclofenac. The resultant poses of these ligands were then subjected to further refinement using a focused docking approach. Altogether, this led to a prediction of a potential binding site and mode of binding for the ASIC3 selective NSAID inhibitors near the acidic pocket domain of rASIC3. A 2D-ligand similarity approach was undertaken to identify scaffolds possessing key functional groups and physico-chemical properties that were similar to those known ASIC3 modulating NSAIDs, and subsequently docked to predict binding interactions. Using these criteria, three molecules (diflunisal, fenbufen and tolmetin) were chosen from a number of hits and were then tested for their ability to modulate the function of rASIC3 transiently transfected in Chinese hamster ovary cells using whole-cell patch-clamp electrophysiology. This in silico approach was also conducted for pro-inflammatory mediators known to activate/enhance ASIC3 activity, which identified potential physiological modulators. Upon activation, the ASIC3 current showed two characteristic phases: a rapid transient phase followed by a prolonged and smaller sustained phase in the presence of continued stimulation, and this was in complete agreement with existing literature. Diclofenac significantly inhibited the sustained, but not the transient phase of the current at pH 4, but no effect on either phase was observed at pH 5. Conversely, the three hits identified in silico showed a varying degree of inhibition on the sustained phase at pH 4 and 5. Finally, site-directed mutagenesis was conducted to validate those amino acids computationally predicted to be involved in NSAID modulation of ASIC3. This thesis outlines a method to predict binding regions of ASIC3 ligands and identifies a possible functional region of ASIC3 by which these ligands interact. These results provide a workflow for identifying novel modulators of ASIC3, which may be of analgesic application.