To erase or not to erase? A catalytically inactive isoform of the KDM5B histone demethylase regulates H3K4 tri-methylation and gene expression

To erase or not to erase? A catalytically inactive isoform of the KDM5B histone demethylase regulates H3K4 tri-methylation and gene expression

The KDM5B histone demethylase is a master regulator of the H3K4-methylome in stem cells and development and its expression is deregulated in many types of cancers, where it generally behaves as a very effective oncogene. It is therefore crucial to achieve a complete understanding of the mechanisms of its regulatory action in cancer. The scientific community has devoted extensive efforts towards this goal, which however has been frustrated by the complexity and heterogeneity of the biological effects that this enzyme can exert in different cellular contexts. This complexity and heterogeneity may be partly linked to the expression of different protein isoforms of the gene, displaying different and -in some cases- even opposite functional roles, depending on the cellular contexts.

We characterized one of these isoforms, namely KDM5B-NTT, which accumulates in breast cancer cell lines due to its increased protein stability compared to the canonical PLU-1 isoform, which shows a much faster turnover. This isoform is a truncated and catalytically-inactive form of the protein, produced by a transcription start site located downstream of that of the PLU-1 isoform, with the consequent use of an alternative ATG for translation initiation. KDM5B-NTT also differs from the PLU-1 isoform for the inclusion of an additional exon (exon-6), previously attributed to other putative isoforms. We showed that KDM5B-NTT has a prevalent nuclear localization as PLU-1. Overexpression of this isoform in luminal breast cancer cells leads to increased bulk levels of H3K4 methylation and induces de-repression of a large set of genes, including genes previously shown to be bound and repressed by the canonical PLU-1 isoform, including the tumor suppressor CAV1. In this context, KDM5B-NTT behaves as a dominant negative, inhibiting the action of PLU-1 as illustrated in the Schematic.

Currently, it is crucial to assess whether the truncated isoform NTT can directly bind to some of the PLU-1 genomic targets, exerting an inhibitory action on the H3K4-demethylating and repressing activity of PLU-1. A second crucial aspect to be addressed regards the relative abundance of PLU-1 and NTT isoforms across different breast cancer subtypes, both in pre-clinical and clinical models. Indeed, KDM5B behaves like an oncogene in several luminal breast cancer cell lines, supporting high proliferation rates, while in some triple-negative cell lines (i.e. MDA-MB-231) its expression has been shown to display a tumor suppressor role. We showed that the PLU-1/NTT protein ratio is higher in MCF7 than in MDA-MB-231, which suggests that a different expression of the two main KDM5B isoforms could be linked to their distinct biological effects; however the mechanistic details of this correlation needs more investigations.

Another interesting point concerns the mechanisms explaining the faster turnover of PLU-1 compared to NTT: we proved that the N-terminal portion of PLU-1, which is missing in the truncated version, is responsible for its faster degradation by the proteasome, unveiling a possible strategy to identify the protein domains involved in the proteasome targeting. As a future perspective, investigating how KDM5B-PLU-1 proteaosomal degradation is regulated could help design therapeutic strategies based on the combination of epigenetic drugs and proteasome inhibitors.

Implications for diagnostic and therapeutic perspectives

The implications of this study in terms of diagnostic and therapeutic approaches for breast cancer treatment are manifold. The KDM5B-NTT isoform expression level, or the PLU-1/NTT ratio, could be tested as a biomarker indicating a different set of regulatory activities in different breast cancer subtypes, possibly correlating a particular disease transcriptomic profile with a specific range of PLU1/NTT protein expression ratio. Furthermore, developing strategies that either boost PLU-1 turnover or stimulate the expression of the dominant negative NTT isoform would allow inhibiting KDM5B catalytic activity in a selective fashion, without affecting other demethylases which could have different regulatory targets.

Finally, it will be very relevant to understand if other non-catalytical isoforms of histone demethylases exist and what can be their relevance in cancer epigenomics.

Please sign in or register for FREE

If you are a registered user on Nature Portfolio Cancer Community, please sign in