I started the project focused on the first approach to development of covalent inhibitors for lysine demethylases. I have initially performed structural studies on KDMs, and based on alignments of the JmjC-KDMs, a cysteine residue was identified as a non-catalytic residue close to the binding site present only in the KDM5 members. This specific cysteine was identified as a potential nucleophile for covalent modifications in the KDM5 family, and covalent docking studies were performed to narrow down the potential modifications based on a cyanopyrazole scaffold previously identified as a 2-OG competitive inhibitor. Key interactions of the main core of the cyanopyrazole scaffold with key residues were kept, and different structural modifications were performed. After optimization of the synthetic methodology, final compounds with different cysteine selective covalent warheads were successfully obtained. The synthesis of the designed compounds is modular, which allowed structural optimization of all parts of the proposed scaffolds. X-ray crystal structures of some compounds showed the reactive covalent warhead within proximity of the targeted cysteine.
Since covalent inhibitors possess time-dependent inhibition due to the kinetics of covalent binding to the protein, their activity was better assessed through determination of the kinetic parameter kinact/Ki, rather than a simple IC50 .These parameters were calculated by using an established method to derive Ki and kinact directly from time-dependant IC50 values. The IC50 values were determined after optimization of the biochemical assay, and we showed that compounds possessed time-dependent inhibition through covalent binding to KDM5B leading to potent nanomolar IC50 values. As well as improving their potency, I performed selectivity and 2-OG competition assays, which showed that the compounds were selective towards the KDM5 members and to our delight, the addition of a covalent warhead to the inhibitors also reduce competition with 2-OG. The covalent binding of the inhibitors with KDM5B was confirmed through MS-labelling experiments. I have also developed a NanoBRET tracer that was used as a tool to test cellular target engagement of the compounds, which was also confirmed. I have finally performed ChIP-seq on the most promising covalent compounds, which showed an increase in H3K4me3 and confirming the efficiency of the compounds (Vazquez-Rodriguez et al. Angew. Chem. Int. Ed., 2019).
After the successful development of the covalent KDM5 inhibitors, I moved into the next main objective, which was the development of KDM5 PROTACs. PROTACs are synthetic chimeric molecules composed of at least two distinct molecular moieties that have separate functional activities. For our purpose, one molecular component is a KDM5 inhibitor scaffold, and the second component, which binds to a protein involved in endogenous protein degradation, is either cereblon or VHL. By bringing the target protein into contact with protein degradation machinery, these chimeric molecules catalyze the degradation of the target protein using the natural machinery of the cells.
I synthesized a variety of PROTACs based on a non-covalent version of the most promising covalent inhibitors previously developed. A variety of linkers taking into account the hydrophobicity and hydrophilicity ratio, the length and the nature of the attachment to the KDM5 and E3 ligase ligands was obtain. A toolbox of E3 ligand-linkers allowed combinatorial synthesis for the final steps of the synthesis of the PROTACs. Biochemical assays showed that the compounds were in the nanomolar range of potency, as their covalent counterparts. Cellular target engagement with the NanoBRET technology confirmed the engagement with both targets, KDM5 and the corresponding E3 ligase. Cellular timecouse also showed that the PROTAC strategy was also effective, degrading the KDM5 in a time and dose-dependent manner. There results are currently under preparat
