Skip to main content

A Chemical Genetics Approach towards Cancer Therapy Targeting Histone Demethylases

Final Report Summary - HISTONE DEMETHYLASES (A Chemical Genetics Approach towards Cancer Therapy Targeting Histone Demethylases)

Selective inhibition of histone demethylases as a novel approach to cancer therapy

In the European Union (EU), cancer is responsible for 25 % of all deaths and is the biggest killer of people aged 45 - 64. Today cancer chemotherapy is mostly carried out with unselective compounds which have severe side effects. Consequently, there is a high need for more targeted cancer therapies.

In the last few years, key discoveries in epigenetics - the science processes that regulate gene expression - have been made. It became clear that chemical modification to DNA and to histones, which are DNA binding proteins, have the potential to activate or repress genes. As cancer is a disease that is characterised by aberrant gene activity, the enzymes that read, write and erase histone modifications are potential targets for focussed cancer therapies. This project was designed to validate specific enzymes, which remove methyl groups from histones (histone demethylases) as cancer therapy targets.

To this end, highly selective inhibitors of these enzymes should be generated in order to probe their role in cell culture models. The inhibitors have to inhibit specific histone demethylases potently and be selective over other histone demethylases. Furthermore, they must be able to cross both the cell membrane and the nuclear membrane in order to be useful compounds. Especially selectivity is a challenge as there are numerous histone methylation sites and the targeted histone demethylase belongs to the family of 2OG oxygenases, of which there are 80 in humans, which all share the same cofactor binding site.

In the original proposal, it was planned to generate specific enzyme-inhibitor pairs. Initial results, however, showed that this strategy was very difficult to pursue, as mutations of key amino acids in the protein might disrupt correct protein folding. We therefore turned our attention to finding inhibitors that bind the substrate binding site of the enzymes, because it offers the potential for generating highly specific inhibitors. Furthermore, this was previously unexplored for histone demethylases.

As targets we chose KDM4A (JMJD2A) because it is overexpressed in breast and prostate cancers and it is involved in disease onset and progression. Together with the very closely related KDM4B and KDM4C, it is the only enzyme that can remove the H3K36me3 histone mark which is associated with activation of genes. Interestingly, the enzymes are furthermore responsible for removal of the repressive mark H3K9me3, an activity that is also reported for the related KDM4D.

In collaboration with colleagues at the University of Tokyo we identified cyclic peptides, which highly selectively inhibit KDM4A-C, but not KDM4D or any other histone demethylase (KDM1, KDM2, KDM3, KDM5, KDM6, KDM7). This was possible by using a state-of-the-art mRNA-display-coupled selection method. Biophysical analyses revealed that the inhibitor is a substrate and not a cofactor competitor, as originally intended. Binding kinetics showed that the cyclic peptide is a tight binder with slow on and off rates. We were able to elucidate the origin of this unprecedented selectivity of one inhibitor by co-crystallising it with KDM4A. The structure revealed that the cyclic peptide binds to the active site of the enzyme, but adopts a previously unknown binding mode that differs significantly from the one observed for the substrate peptides. As a key feature an arginine side chain binds to the pocket usually occupied by the methylated lysine of the histone.

In conclusion, the key objective of the project, the generation of a highly selective inhibitor that exhibits subfamily selectivity has been reached. The inhibitor is the first compound known, which selectively inhibits all enzymes which remove the H3K36me3 mark, while not affecting other histone demethylases.

In a next step, we tried to investigate the effect of the compounds in cells. Initial experiments, however, showed that the peptides are not able to cross the cell membrane or get degraded inside cells. We tried to address this problem using three different techniques:

1) reduction of the peptide charge to allow efficient transport;
2) N-Methylation of the peptide backbone in order to make the peptide more stable and more lipohilic;
3) conjugation of the peptide to effector molecules which allow them to be transported inside the cell.

Specifically, the peptides were conjugated to folic acid, which is known to transport cargo inside cells. Unfortunately, the generated compounds did not show any effect, possible reflecting limited stability of the peptides in cells. Structure activity relationships showed that two arginines in the peptide were crucial to its potency. Therefore, modifications had to be introduced to compensate for the loss of the key side chains upon reducing charge.

We set out to regain activity by chelating the active site metal. To this end, several metal binding amino acids were synthesised and incorporated into different positions of the peptides. In case of the correct metal binder and peptide, position affinity for the target was nearly completely regained. Unfortunately, the peptides were not active in cells. As a next step different N-methylated peptides were synthesised which indeed showed the potential to affect H3K36me3 methylation in cancer cells as judged by immunofluorescence analyses using specific antibodies. This shows that increase in lipophilicity and stability can yield cyclic peptides that do cross both the cellular and nuclear membrane. Pleasingly, also N-methylated peptides with metal binding amino acids showed activity in cells.

In a final step, the effect of KDM4A-C inhibition in cells was investigated. To this end, the proliferation rate of different cancer cell lines was measured. All cells treated with the peptides showed reduced growth rate, reflecting the importance of histone H3K36 demethylation for normal cell growth. Staining of cell nuclei suggested that the cells are locked in one phase of the cell cycle, consistent with an important role of KDM4A-C in the regulation of gene expression.

In conclusion, the project resulted in potent and selective inhibitors for a specific subfamily of histone demethylases. Initial results in cell based assays suggest that KDM4A-C are valid targets for cancer therapy, as the growth of these cells is inhibited. The compounds generated will be of great value in the further design anti-cancer drugs. Altogether the study could lead to a novel cancer therapy, from which patients in the EU and the whole world would greatly benefit.

Contact details:
Prof. Christopher J. Schofield
Oxford University
Department of Chemistry
Mansfield Road
Oxford, OX13TA, United Kingdom

Related documents