Biochemical and cellular studies of PI3 kinase inhibition

Phosphatidylinositol 3-kinase (PI3K) is a lipid kinase that catalyzes the biosynthesis of PI(3)P, PI(3,4)P₂ and PI(3,4,5)P₃ – second messengers that trigger a wide range of downstream signaling cascades involved in cell survival, growth, adhesion and proliferation. The class I PI3K proteins have been the subject of much study due to their association with disease. They are heterodimeric, composed of a regulatory subunit (p85) complexed with either one of four different isoforms of the catalytic subunit (p110α, p110β, p110γ and p110δ). Cancer has been a particular focus of regarding the roles of each isoform. The PI3KCA gene encoding the p110α-isoform has been found to be frequently mutated in cancers such as breast, prostate, colon, liver and brain. Thus, PI3K inhibitors showing selectivity towards the p110α isoform have been identified as targets for treatment of cancer. Other isoforms have also been identified as participating in the progression of certain cancers. This project is divided into two parts; the first part of the study is aimed at elucidating molecular mechanisms governing isoform selective inhibition of PI3K while in the second part of the thesis, the concept of using prodrugs as a means of targeting tissue selectivity is examined. There has been a growing number of class I PI3K inhibitors described to date with some showing selectivity to different PI3K isoforms. However the basis of selectivity of these inhibitors is still ambiguous. Previous studies have shown that specific regions within the catalytic subunit contain non-conserved residues which are involved in isoform selectivity. In order to rationally develop isoform selective inhibitors, the nature of the binding interactions between the enzyme and the inhibitors need to be understood. The approach taken in this thesis has been to use site-directed mutagenesis, to interrogate the role of key residues postulated to be responsible for the observed isoform selectivity of certain inhibitors. We have produced a series of in vitro mutants at position 770 and 859 (p110α-numbering) using two baculovirus expression systems. It was found that the Bac-to-bac baculovirus expression system provides an efficient method to produce in vitro PI3Kα mutant enzymes with enzyme activity comparable to the wild type enzyme. These mutant enzymes were assessed against PI3Kα selective inhibitor PIK-75, pan-PI3K inhibitor ZSTK-474 and some analogues in order to identify the role of the non-conserved residues, Arginine 770 and Glutamine 859, in determining inhibitor potency and selectivity. It was shown that changes to Arginine 770 and Glutamine 859 of p110α isoform did not influence isoform selectivity of the inhibitors. These results were used to compare some conflicting models of PIK-75 binding and were found to contradict some of the theoretical models and support one specific model proposed. Subsequent studies involving mutagenesis at other residues further confirmed this model. In Chapter 4, a similar examination of non-conserved residues in the binding of well characterised pan-PI3K inhibitor, GDC-0941 and an analogue compound 19 (CNIO-19) has been presented. Despite crystallographic evidence of isoform-specific binding interactions, these mutations had little effect on ligand binding. These results suggest that GDC-0941 can bind to PI3K isoforms with high flexibility and therefore that GDC-0941 non-selectivity is a function of redundancy in multiple possible binding orientations in different PI3K isoforms. In vitro mutagenesis using a panel of p110α mutant enzymes showed that mutation of the non-conserved residues did not affect GDC-0941 binding due to another mode of binding being utilised. CNIO-19 has an isosteric structure to GDC-0941 but an altered selectivity profile, with a markedly reduced potency for p110α. The comparison could provide a reference point for validating hypotheses relating to the binding mechanism. Again with this compound, the influence of mutations in p110α had only modest effects upon potency, albeit in the case of R770A and R777A the changes in potency were in opposite directions. It is likely that an extended series of isoform mutants and analogues would be required to adequately assess the role of binding site residues in dictating selectivity. The second part of this thesis presents a different aspect of this project which was the development of a study model to assess drug candidates for cancer therapeutics. The archetypal PI3K inhibitor, LY294002 and a pyridinyl analogue were modified by introduction of a tetrazole substituent at the 6-position resulting in two potent PI3Kα inhibitors, showing a 50-fold increase in potency relative to LY294002 itself. However neither compound showed activity in inhibiting MCF-7 breast cancer cell proliferation, which was proposed to be due to poor cellular permeability. Therefore a prodrug strategy was adopted to enhance cellular uptake whereby a pivaloyl ester group was added to the molecule. Both inhibitors possessed anti-proliferative activity against the MCF-7 cancer cell line when delivered using the prodrug strategy. The susceptibility of the compounds to a model esterase and the MCF-7 cell lysate were examined, which suggested that intracellular cleavage to the active compounds could occur but was possibly incomplete in the in vitro experiments. The results in total suggest this is a potentially promising avenue to pursue in the development of anti-cancer agents with reduced side effects. Overall, this thesis presents our current understanding of the biochemistry and pharmacology of several known PI3K inhibitors. We found that in vitro site-directed mutagenesis of PI3Kα enzymes is a powerful tool to assess the roles of non-conserved amino acids within the enzyme catalytic pocket especially in identifying selective inhibitor:enzyme interactions. The outcome from this study will be a facilitation of future design and development of novel isoform selective PI3K inhibitors based on these characteristics: 1) improved potency, 2) isoform selectivity, 3) cellular permeability and 4) enhanced anti-proliferative activity.