Isoform selective inhibitors of phosphatidylinositol 3-kinase
2017-02-08T01:02:13Z (GMT) by
Inhibitors of the phosphatidylinositol 3-kinase (PI3K) pathway are at the forefront of the search for novel cancer treatments, as aberrations in the pathway are found in up to 50% of human cancers. Class I PI3K consists of four isoforms, PI3Kα, PI3Kβ, PI3Kγ and PI3Kδ, which phosphorylate phosphatidylinositol 4,5-bisphosphate to form the second messenger, phosphatidylinositol 3,4,5-trisphosphate. This second messenger signals downstream to regulate cell metabolism, growth and proliferation. Oncogenic mutations in PI3Kα occur in approximately 30% of colorectal cancers and 25-40% of breast cancers. The work in this thesis has sought to add to the understanding of isoform selectivity in PI3K inhibitors, which will hopefully contribute to the successful treatment of human cancers. This has been conducted in two parts: (i) the structural investigation of wild-type PI3Kα, examining the differences in regulation compared with the oncogenic mutant, p110αHis1047Arg, and the characterisation of the lipid-binding site; (ii) the design, synthesis and testing of novel PI3K inhibitors designed to explore the determinants of isoform selectivity. Chapters 3 and 4 have described the synthesis and testing of 38 analogues of the pan PI3K inhibitor, ZSTK474. In Chapter 3, an efficient, regioselective synthetic method was developed for the synthesis of 5- and 6-substituted benzimidazole derivatives of ZSTK474. Methoxy substitutions at these positions resulted in a decrease in potency at PI3Kα. A more dramatic impact on PI3Kγ potency was observed, suggesting alternative substitutions at this position may be a suitable method to dial out PI3Kγ inhibition. Two hydroxyphenyl substituents were investigated as alternatives to the 2-difluoromethylbenzimidazole substituent of ZSTK474. The 3-hydroxyphenyl substituted compounds (37-42) had a greater preference for the PI3Kα isoform than the equivalent 2-difluoromethylbenzimidazole compounds (25a-f), despite a decrease in potency. The 3-hydroxyphenyl substituted compounds also showed significantly greater selectivity against PI3Kγ, suggesting it may be a better affinity pocket binding moiety for the development of PI3Kα selective inhibitors. More work is required to optimise the other binding moiety, replacing a morpholine group of ZSTK474, to improve potency and PI3Kα selectivity. In Chapter 4, a series of inhibitors targeting a specific non-conserved residue in PI3Kα, p110αGln859, was designed, synthesised and tested. Compounds showed a general trend of improved selectivity towards PI3Kδ, due to the presence of a similar residue at the analogous position, p110δAsn836. Compound 46 was the most PI3Kδ selective compound, showing a selectivity of 6-35-fold. Compound 50 displayed the greatest selectivity for PI3Kδ against PI3Kβ of 21-fold. Site-directed mutagenesis assay results of compound 50 are consistent with interactions at the residue p110αGln859. The work in thesis has investigated the design of novel PI3Kα inhibitors. The identification of key differences in interaction and regulation between wild-type PI3Kα and the oncogenic mutant, p110αHis1047Arg has the potential to inform the design of novel, oncogenic mutant specific inhibitors. The characterisation of the PI3Kα lipid-binding site can aid in the structure-based drug design of lipid-competitive inhibitors, potentially providing increased selectivity over related protein kinases. We have demonstrated the applicability of targeting specific non-conserved residues in the PI3K binding site for the rational design of isoform selective inhibitors, and the importance of the affinity pocket binding portion of the inhibitor for dialling out PI3Kγ inhibition.