Hypothesis

Progressive aging alters the H3K4me3/H3K27me3 ratio at bivalent promoters in a tissue‑specific manner, thereby switching the functional output of KDM5 and KDM6 histone demethylase isoforms from tumor‑suppressive to tumor‑promoting activities. This switch explains the paradoxical, context‑dependent roles of these enzymes in cancer and provides a mechanistic link between age‑associated epigenetic drift and chemotherapy resistance.

Rationale

• et al. show that KDM5A/B drive chemoresistance via DNA‑repair and mitochondrial pathways, whereas KDM5D can be tumor‑suppressive or resistance‑promoting depending on cancer type[1][2].
• KDM6A/B act as tumor suppressors in bladder cancer but promote colorectal cancer stemness through H3K27me3 demethylation[1][3].
• Bivalent promoters with a high H3K27me3:H3K4me3 ratio are more prone to DNA hypermethylation in cancer, and the ratio itself predicts epigenetic vulnerability[4].
• Chemotherapy‑induced loss of H3K27me3 at bivalent regions activates persister programs; timely KDM6 inhibition sustains the mark and reduces persisters[5].

Together, these observations suggest that the balance of activating (H3K4me3) and repressive (H3K27me3) marks determines whether KDM5/6 isoforms act as enzymes that remove inhibitory marks (promoting transcription) or as scaffolds that recruit repressive complexes. Aging is known to cause a global decline in H3K27me3 and a relative increase in H3K4me3 at developmental promoters, shifting the ratio toward a more "open" chromatin state.

We propose that this age‑driven ratio shift reorients the functional output of KDM5/6 isoforms: in young cells, a high H3K27me3:H3K4me3 ratio favors demethylase‑dependent removal of H3K27me3, activating tumor‑suppressive pathways; in aged cells, a lowered ratio makes the same demethylase activity insufficient to overcome prevailing H3K4me3, causing the isoforms to act primarily through non‑catalytic scaffolding functions that recruit oncogenic co‑factors (e.g., EZH2, MDR1).

Testable Predictions

1. Isoform‑specific ChIP‑seq in isogenic young vs. senescent human fibroblasts will reveal a global decrease in H3K27me3 enrichment and a concomitant increase in H3K4me3 at bivalent promoters with age, quantified as a reduced H3K27me3/H3K4me3 ratio.
2. In aged cells, KDM5B and KDM6A occupancy at these promoters will shift from correlating with loss of H3K27me3 (young) to correlating with gain of H3K4me3 and recruitment of MDR1/EZH2 complexes (aged), detectable by re‑ChIP for demethylase and co‑factor.
3. Pharmacological inhibition of KDM5/6 demethylase activity will reduce chemoresistance in young cancer cells but increase resistance in aged‑mimicked cells, whereas disruption of the protein‑protein interaction domains (scaffolding mutants) will have the opposite effect.
4. Exposing young tumor models to an accelerated aging regimen (e.g., chronic low‑dose ROS) will recapitulate the isoform functional switch and render them resistant to temozolomide or gemcitabine, an effect reversible by restoring the H3K27me3/H3K4me3 ratio via EZH2 activation or KDM5 inhibition.

Experimental Design

• Cell models: IMR‑90 fibroblasts (young, irradiated‑induced senescent) and matched cancer lines (e.g., U87 glioblastoma, PANC‑1 pancreatic) treated with senescence inducers.
• Assays:
  • ChIP‑seq for H3K4me3, H3K27me3, KDM5A/B/D, KDM6A/B.
  • CUT&RUN for EZH2, MDR1, and demethylase interaction partners.
  • RNA‑seq to correlate promoter changes with expression of drug‑resistance and stemness genes.
  • Viability assays (temozolomide, gemcitabine) with demethylase inhibitors (CPI-455, GSK‑J4) and scaffolding‑disrupting peptides.
• Controls: Catalytically dead mutants (H‑domain mutants) and scaffolding‑deficient mutants (deletion of ARID or tetratricopeptide repeat domains).

Potential Outcomes & Falsification

• Support: Observation of an age‑related ratio shift coupled with a demethylase‑to‑scaffolding functional switch, and reversal of chemoresistance phenotypes by restoring the ratio or blocking scaffolding interactions.
• Refutation: No significant change in H3K27me3/H3K4me3 ratio with aging, or KDM5/6 isoform activity remains strictly demethylase‑dependent across ages, indicating that other mechanisms drive the context‑dependent phenotypes.

This hypothesis directly links chromatin‑state dynamics in aging to the enigmatic duality of KDM5/6 enzymes in cancer, offering a clear, falsifiable framework for future mechanistic and therapeutic studies.