αKG/Fe2+ Availability Switches KDM5 from Enzymatic Demethylase to Bivalent Domain Scaffold in Aging HSCs versus Cancer

Chloe Vargas (PhD)· Mar 26, 2026

Mechanism: KDM5 enzyme function switches based on αKG/Fe2+ levels, acting as a scaffold in aged HSCs (low nutrients) to preserve H3K4me3, but as an active demethylase in cancer (high nutrients) to remove H3K4me3. Readout: Readout: Supplementing aged HSCs with αKG/Fe2+ decreases H3K4me3 and self-renewal gene expression, while blocking αKG in cancer restores H3K4me3 and reverses therapy resistance.

Hypothesis

The activity state of KDM5 demethylases is governed by intracellular α‑ketoglutarate (αKG) and Fe2+ levels, determining whether KDM5 functions as an enzymatic H3K4me3 remover or as a non‑catalytic scaffold that stabilizes bivalent domains. In aged hematopoietic stem cells (HSCs), low αKG/Fe2+ favors a scaffolding role that preserves H3K4me3 at bivalent promoters despite global H3K27me3 loss, whereas in cancer cells elevated αKG/Fe2+ drives KDM5‑dependent H3K4me3 demethylation at tumor‑suppressor and immune promoters.

Mechanistic Reasoning

Cancer context – KDM5A/B/C/D are overexpressed and cooperate with high αKG/Fe2+ (produced by glutaminolysis and TCA cycle flux) to catalyze H3K4me3 removal at PTEN, MGMT, HLA‑A and STING, fostering therapeutic resistance and immune evasion[1].
Aging context – Global H3K4me3 remains stable in neurons and HSCs[2]; HSCs gain ~4‑fold more bivalent domains at self‑renewal genes during aging[4]. Simultaneously, H3K27me3 drifts downward at enhancers and developmental loci[3]. The observed H3K4me3 stability suggests KDM5 is not actively demethylating these marks.
Nutrient‑sensing switch – αKG is a required co‑factor for the JmjC domain of KDM5; Fe2+ is the catalytic iron. When αKG/Fe2+ are limiting, KDM5 can still bind chromatin via its ARID and PHD domains but lacks catalytic activity, effectively acting as a placeholder that prevents nucleosome remodeling and maintains the bivalent configuration.
Predicted outcome – In aged HSCs, supplementing αKG or Fe2+ should convert KDM5 from a scaffold to an active demethylase, reducing H3K4me3 at bivalent promoters and precipitating loss of self‑renewal gene expression. Conversely, inhibiting αKG production (e.g., with AGI‑5198) or chelating Fe2+ in cancer cells should suppress KDM5 enzymatic activity, preserving H3K4me3 at tumor‑suppressor loci and sensitizing cells to therapy.

Testable Predictions

Prediction 1: Aged HSCs exhibit lower intracellular αKG and Fe2+ levels than young HSCs or cancer cell lines. Measurement via LC‑MS metabolomics will show a significant decrease (p<0.01).
Prediction 2: Chromatin immunoprecipitation followed by sequencing (ChIP‑seq) for KDM5 in aged HSCs will reveal peak enrichment at bivalent promoters without concomitant loss of H3K4me3, whereas in cancer cells KDM5 peaks will correlate with H3K4me3 depletion.
Prediction 3: Exogenous αKG (2 mM) or Fe2+ (50 µM) treatment of aged HSCs for 48 h will increase KDM5‑dependent H3K4me3 demethylation (measured by CUT&RUN) and reduce expression of self‑renewal genes (e.g., Hoxa9, Mecom) by >30 % (qPCR).
Prediction 4: In a breast‑cancer model resistant to KDM5 inhibitors, combining the inhibitor with an αKG antagonist (e.g., AGI‑5198) will restore H3K4me3 at PTEN and HLA‑A promoters and decrease interferon‑stimulated gene signaling, reversing resistance.

Experimental Approach

Metabolite profiling – Isolate young (2 mo) and aged (20 mo) murine HSCs, plus a cancer line (MDA‑MB‑231). Quantify αKG, succinate, Fe2+ using targeted metabolomics.
Chromatin assays – Perform KDM5 ChIP‑seq and H3K4me3/H3K27me3 CUT&RUN on the same samples. Integrate peaks to assess occupancy‑mark relationships.
Nutrient manipulation – Culture aged HSCs ex vivo with αKG, Fe2+, or vehicle; assess KDM5 catalytic activity in vitro using a demethylase assay with histone peptides.
Functional readouts – Colony‑forming unit (CFU) assays for self‑renewal; flow cytometry for lineage markers. In cancer lines, measure proliferation, apoptosis, and response to radiation or immunotherapy after combined KDM5 inhibitor and αKG blockade.

Falsifiability

If aged HSCs display normal or elevated αKG/Fe2+ levels, or if KDM5 occupancy does not correlate with bivalent domains irrespective of nutrient status, the hypothesis would be refuted. Likewise, if αKG/Fe2+ supplementation fails to increase KDM5‑dependent H3K4me3 loss or alter self‑renewal gene expression, the proposed metabolic switch is invalid.

References

[1] https://elifesciences.org/articles/106249 [2] https://pmc.ncbi.nlm.nih.gov/articles/PMC11353134/ [3] https://elifesciences.org/articles/35368 [4] https://pmc.ncbi.nlm.nih.gov/articles/PMC7534803/ [5] https://aacrjournals.org/clincancerres/article/30/22/5166/749574/Epigenome-Reprogramming-Through-H3K27-and-H3K4 [6] https://www.frontiersin.org/journals/epigenetics-and-epigenomics/articles/10.3389/freae.2025.1594400/full }

Hypothesis

Mechanistic Reasoning

Cancer context – KDM5A/B/C/D are overexpressed and cooperate with high αKG/Fe2+ (produced by glutaminolysis and TCA cycle flux) to catalyze H3K4me3 removal at PTEN, MGMT, HLA‑A and STING, fostering therapeutic resistance and immune evasion[1].

Aging context – Global H3K4me3 remains stable in neurons and HSCs[2]; HSCs gain ~4‑fold more bivalent domains at self‑renewal genes during aging[4]. Simultaneously, H3K27me3 drifts downward at enhancers and developmental loci[3]. The observed H3K4me3 stability suggests KDM5 is not actively demethylating these marks.

Nutrient‑sensing switch – αKG is a required co‑factor for the JmjC domain of KDM5; Fe2+ is the catalytic iron. When αKG/Fe2+ are limiting, KDM5 can still bind chromatin via its ARID and PHD domains but lacks catalytic activity, effectively acting as a placeholder that prevents nucleosome remodeling and maintains the bivalent configuration.

Predicted outcome – In aged HSCs, supplementing αKG or Fe2+ should convert KDM5 from a scaffold to an active demethylase, reducing H3K4me3 at bivalent promoters and precipitating loss of self‑renewal gene expression. Conversely, inhibiting αKG production (e.g., with AGI‑5198) or chelating Fe2+ in cancer cells should suppress KDM5 enzymatic activity, preserving H3K4me3 at tumor‑suppressor loci and sensitizing cells to therapy.

Testable Predictions

Prediction 1: Aged HSCs exhibit lower intracellular αKG and Fe2+ levels than young HSCs or cancer cell lines. Measurement via LC‑MS metabolomics will show a significant decrease (p<0.01).

Prediction 2: Chromatin immunoprecipitation followed by sequencing (ChIP‑seq) for KDM5 in aged HSCs will reveal peak enrichment at bivalent promoters without concomitant loss of H3K4me3, whereas in cancer cells KDM5 peaks will correlate with H3K4me3 depletion.

Prediction 3: Exogenous αKG (2 mM) or Fe2+ (50 µM) treatment of aged HSCs for 48 h will increase KDM5‑dependent H3K4me3 demethylation (measured by CUT&RUN) and reduce expression of self‑renewal genes (e.g., Hoxa9, Mecom) by >30 % (qPCR).

Prediction 4: In a breast‑cancer model resistant to KDM5 inhibitors, combining the inhibitor with an αKG antagonist (e.g., AGI‑5198) will restore H3K4me3 at PTEN and HLA‑A promoters and decrease interferon‑stimulated gene signaling, reversing resistance.

Experimental Approach

Metabolite profiling – Isolate young (2 mo) and aged (20 mo) murine HSCs, plus a cancer line (MDA‑MB‑231). Quantify αKG, succinate, Fe2+ using targeted metabolomics.

Chromatin assays – Perform KDM5 ChIP‑seq and H3K4me3/H3K27me3 CUT&RUN on the same samples. Integrate peaks to assess occupancy‑mark relationships.

Nutrient manipulation – Culture aged HSCs ex vivo with αKG, Fe2+, or vehicle; assess KDM5 catalytic activity in vitro using a demethylase assay with histone peptides.

Functional readouts – Colony‑forming unit (CFU) assays for self‑renewal; flow cytometry for lineage markers. In cancer lines, measure proliferation, apoptosis, and response to radiation or immunotherapy after combined KDM5 inhibitor and αKG blockade.

Falsifiability