In Silico and In Vitro Studies of Potential Novel Anti-Cancer Agents Targeting Angiogenin and Glycogen Phosphorylase

Abstract

Glioblastoma remains the most aggressive cancer of the central nervous system, characterised by its high proliferative rate, invasiveness, and angiogenesis. Increased drug resistance and side effects have greatly reduced the overall efficacy of current treatments, and in most cases, these tumours are recurring. Novel therapeutic strategies combined with conventional treatments are therefore imperative for enhancing the overall outcome and quality of life for cancer patients. Tumour growth is not only closely associated with angiogenesis, they are able to undergo metabolic rewiring which increases their glucose uptake and consumption to ensure their survival. Inhibition of key angiogenic factors as well as inactivating this metabolic switch could be used as novel therapeutic strategies. As such, the proteins human angiogenin, an angiogenic factor, and glycogen phosphorylase, a key regulator of glucose levels, are both attractive pharmaceutical targets for the design of novel anti-cancer agents.

An in silico screening of the ZINC-15 Biogenic database targeting the cell-binding and active site of human angiogenin (hAng) is presented. A docking consensus scoring approach was implemented and, using a Simple Sum rank combination of Glide-SP and -XP, GOLD, and AutoDock Vina, the compounds sorted and ranked. The protein-ligand interactions and MM-GBSA binding free energies of native ligand (4a) displaying potent anti-angiogenic activity, was studied and used as a benchmark for the selection of ten diverse compounds demonstrating dual-action inhibition.

Furthermore, this thesis investigated several binding sites of glycogen phosphorylase, namely the allosteric, quercetin, catalytic, and inhibitor sites, for the discovery of potential anti-glioblastoma compounds. For the allosteric site, a benchmarking study was carried out to assess different combinations of docking programs, as well as exploiting pharmacophore models to enrich the overall accuracy of the results. The Simple Sum Rank combination using Glide-SP and -XP docking scores, and filtering using a 4-point pharmacophore model was selected determined from the superior statistical metrics obtained including the EF, EF’, and BEDROC statistical performance metrics. Following the virtual screening, 29 compounds were selected based on their superior docking scores and interactions with key binding site residues.

For the GP quercetin binding site, a similarity search using the ZINC-12 database was performed, and together with the selection of several known flavonoid compounds previously identified as GP inhibitors, these compounds were screened using Glide-SP and -XP docking. Additionally, a screening of the ZINC-15 and Analyticon databases were performed and compounds filtered using the docking scores of the cognate ligand quercetin for comparison, but also considering Lipinski’s rule of five, Jorgensen’s rule of three, Veber’s rules, and E-pharmacophore/receptor cavity pose filters. Furthermore, the key interactions formed between the site residues and quercetin was used to guide the selection of the final 27 compounds.

For the GP inhibitor site, the docking predictions (Glide-SP and -XP docking scores and residue interactions) of baicalein in comparison to the previously studied flavonoid, chrysin, was assessed.

Towards validating predicted compounds as true inhibitors, baicalein, together with the selected allosteric (26 compounds) and quercetin (27 compound) candidates, their % inhibition against rmGPb was initially determined (using 100 μM compound concentration). Subsequently, nine compounds were taken forward for further enzyme kinetic assays revealing low to moderate activities. Additionally, all phase 2 compounds showed > 95 % inhibition against rmGPb. Through X-ray crystallographic studies, the binding sites/mode of these compounds were determined, with additionally a novel binding discovered on the surface of GP.

Subsequently, five of these compounds binding at the inhibitor, quercetin and allosteric sites, along with one new allosteric and three previously studied catalytic site inhibitors were taken forward for cell viability assays. Their effects on cell viability and growth in healthy human foetal glial cells (SVG p12), and three glioblastoma cell lines (U87 MG, T98G, and U251 MG), in comparison to the known inhibitor, flavopiridol, is reported. Compounds CP-91149, baicalein, and pelargonidin showed significant decrease in cell viabilities overall, which was reflected in their low μM IC50 values. This is an important outcome towards validating GP as a novel target for glioblastoma.

The future work entails combination treatment with TMZ, repeating cell viability assays in hypoxic conditions, as well as flow cytometry experiments to elucidate the effects on the cell cycle, proliferation rate and apoptosis induction. Additionally, ten potential anti-angiogenic compounds have been identified that can be further assessed at the cellular level.