Towards Inhibition of Oncogenic RhoA Signaling using Stapled Peptides

Abstract

Given that cancer is the second leading cause of death worldwide with its major treatments, surgery and chemotherapy, causing significant side-effects, diminished quality of life and leaving a risk of recurrence, more efficient and less toxic treatments are needed. A challenge in targeted therapy is the "undruggability" or difficulty in developing traditional small molecule inhibitors to many of the identified oncogenic targets, including the Ras superfamily of small GTPases. Peptide therapeutics are emerging as promising new modalities to more efficiently inhibit these "undruggable" targets. RhoA is a member of the Rho family within the Ras superfamily of small GTPases and is a major regulator of cytoskeletal remodelling, cell migration and proliferation. It is found overexpressed, overactivated or mutated (eg. RhoA G17V) in many cancers including gastric, testicular, hepatocellular carcinoma and lymphomas, and is associated with invasion, mestastasis, therapy resistance and poor prognosis. Therefore, RhoA is a potential therapeutic target in cancers with high RhoA activity. Given the relatively large and shallow surface of RhoA that interacts with its downstream effectors, deep, hydrophobic pockets suitable for small molecule inhibitors are nonexistent. The larger size of peptide inhibitors may, therefore, lead to more specific and tighter-binding candidates for further drug development. This work presents the investigation of the HR1a domain from the RhoA effector, PRK1/PKN1 kinase, as a starting point for peptide inhibitor design. HR1a binds to RhoA with nanomolar affinity and forms an anti-parallel coiled coil. This dissertation starts with mutagenesis studies on the RhoA-binding residues of HR1a to identify the minimal region of hotspots, which was found to encompass a 17-mer, α-helical sequence (residues 48-63) within the N-terminal helix of HR1a. A starting peptide scaffold made of this minimal sequence (called P1) was synthesised and assessed in Circular Dichroism and in a competition-based binding assay, Scintillation Proximity Assays. P1 was mostly disordered in solution and fully competed against HR1a for RhoA binding with an IC50 of 2.1 mM. All-hydrocarbon stapling and lactamisation of the peptide at the N-terminal 50-54 residues increased α-helicity to around 50%, but did not improve RhoA binding affinity. A series of mutations on the linear peptide were screened to identify those with improved binding affinity, out of which A56_cycloleucine emerged with a 26% lower IC50 than P1. A second stage of peptide optimisation was performed. The peptide was lengthened to 23-residues (numbered as 46-68) to include fraying ends and an extra RhoA-interacting residue. Molecular Dynamics simulations on a series of all-hydrocarbon stapled peptides identified two staples - a C-terminal i,i+4 on 60-64 residues (L4) and a central i,i+7 staple on 53-60 positions (L7) - that were the most likely to stabilise α-helical conformations conducive to RhoA binding. Biophysical assays revealed that both stapled peptides promoted α-helicity to 41-57% and lowered IC50s to the micromolar range. Furthermore, the A56_cycloleucine mutation was introduced into both L4 and L7, yielding the final, most improved peptide - L7_cyL with the 53-60 staple and A56_cyL mutation - exhibiting a 130 μM IC50 and 62.5 μM KD, which is an overall 16-fold enhanced IC50 from the first, original peptide tested, P1. This dissertation demonstrates how using a native binding partner as template and the use of molecular dynamics simulations can assist in the rational design of stapled peptide inhibitors targeting an "undruggable" protein-protein interaction.