EZH1-dependent H3K27me1 is an adaptive chromatin barrier that limits DNMT inhibitor response in colorectal cancer

This study identifies EZH1-dependent H3K27me1 as a key adaptive barrier limiting DNMT inhibitor efficacy in colorectal cancer and demonstrates that combining DNMT inhibitors with dual EZH1/2 inhibitors overcomes this resistance by eliminating H3K27 methylation, resolving induced bivalent chromatin states, and suppressing oncogenic transcription to enhance therapeutic response.

The Gist

The Big Picture: Unlocking the Cancer Cell's "Locked Door"

Imagine a cancer cell as a house where the lights are turned off and the doors are locked tight. The "lights" are the genes that tell the body to stop growing or to fight back (tumor suppressors), and the "doors" are the switches that control them.

In many cancers, the cell has put a heavy, permanent padlock on these doors. This padlock is made of DNA methylation (a chemical tag that silences genes). Doctors have a key called a DNMT inhibitor (like Decitabine) that tries to pick this padlock. When it works, the door opens, and the "lights" turn on, potentially killing the cancer.

The Problem: The cancer cell is smart. When it feels the padlock being picked, it doesn't just sit there. It quickly installs a second, backup security system made of histone proteins. Specifically, it uses a machine called EZH2 to put a "Do Not Enter" sign (H3K27me3) on the door. This stops the lights from turning on, even if the padlock is picked.

The New Discovery: The "Shadow Guard" (EZH1)

Scientists thought that if they stopped the main security machine (EZH2) with a drug, the backup system would fail, and the door would open. They were half-right.

However, this paper discovered a sneaky Shadow Guard called EZH1.

• The Main Guard (EZH2): Usually puts the big "Do Not Enter" sign (H3K27me3) on the door.
• The Shadow Guard (EZH1): When the Main Guard is stopped, the Shadow Guard steps in. It doesn't put the big sign; instead, it puts a smaller, sticky note (H3K27me1) on the door. It's a subtle barrier, but it's enough to keep the door locked and the lights off.

The Analogy: Imagine trying to open a door.

1. DNMT Inhibitor: Picks the main lock.
2. EZH2 Inhibitor: Stops the main guard from putting up a "Do Not Enter" sign.
3. The Catch: The Shadow Guard (EZH1) is still there, putting up a sticky note that says "Stay Out." The door is still stuck.

The Solution: Double-Teaming the Guards

The researchers found that to truly open the door, you can't just stop the Main Guard; you have to stop both guards at the same time.

They tested a combination therapy:

1. Pick the Padlock: Use the DNMT inhibitor.
2. Stop Both Guards: Use a drug that inhibits both EZH1 and EZH2 (like Valemetostat).

What Happens When You Do This?
When you stop both guards, the "sticky notes" disappear. Suddenly, the door is wide open. But here is the magic part:

• The Bivalent State: Once the guards are gone, the cell tries to put a "Welcome" sign (H3K27ac) on the door. However, because the original padlock (DNA methylation) is still partially there, the door is in a weird "bivalent" state—it's half-open, half-locked. It's ready to go, but not quite there yet.
• The Final Push: When you combine the "Stop Both Guards" drug with the "Pick the Padlock" drug, the padlock is fully removed. The "Welcome" sign takes over completely. The lights turn on, and the cancer cell starts to die.

The "Good" vs. "Bad" Lights

The paper also found something fascinating about which lights turn on.

• The Good Lights (Tumor Suppressors): The combination therapy turns on the genes that tell the cancer to stop growing and die.
• The Bad Lights (Oncogenes): Cancer cells usually have a "super-bright" light on for genes that make them grow fast (like the MYC and E2F genes). The combination therapy actually turns these bad lights off.

The Metaphor: Think of the cancer cell as a car with a stuck accelerator (bad genes) and a broken brake (good genes).

• The old drugs tried to fix the brake but the car kept speeding because the accelerator was jammed.
• This new combination therapy fixes the brake AND jams the accelerator. The car stops immediately.

Why This Matters

1. It Explains Why Some Drugs Fail: If you only stop the Main Guard (EZH2), the Shadow Guard (EZH1) saves the day, and the cancer survives.
2. It Offers a New Strategy: To beat colorectal cancer, we need drugs that target both EZH1 and EZH2 simultaneously, not just one.
3. Personalized Medicine: The paper suggests that patients with high levels of the "Shadow Guard" (EZH1) might need this specific double-drug approach to get better.

Summary in One Sentence

Cancer cells use a "Shadow Guard" (EZH1) to hide behind when doctors try to stop the "Main Guard" (EZH2); by stopping both guards and picking the original lock at the same time, we can finally unlock the cancer's defenses and turn off its growth engine.

Technical Summary

1. Problem Statement

DNA methyltransferase inhibitors (DNMTi), such as decitabine (DAC) and azacitidine, are effective in hematological malignancies but show limited efficacy in solid tumors like colorectal cancer (CRC). A primary barrier to their success is adaptive epigenetic reprogramming: when DNMTi induces DNA hypomethylation, cancer cells compensate by reinforcing gene silencing through the acquisition of repressive histone post-translational modifications (PTMs).

Previous work established that the Polycomb Repressive Complex 2 (PRC2) catalytic subunit EZH2 drives this resistance by depositing H3K27me3. While EZH2 inhibitors (e.g., tazemetostat) can synergize with DNMTi, they often fail to fully overcome resistance. The specific role of the PRC2 paralog EZH1 and its catalytic product H3K27 mono-methylation (H3K27me1) in mediating this adaptive resistance remained uncharacterized.

2. Methodology

The study employed a multi-omics approach combining pharmacological screening, genetic manipulation, and high-resolution epigenomic profiling:

• Pharmacological Screening: A diverse panel of PRC2 inhibitors (including SAM-competitive inhibitors, PPI disruptors, and PROTAC degraders) was screened in an SFRP1-Nanoluciferase (NLuc) reporter line (HCT116 background) to identify compounds that synergize with low-dose DAC to reactivate silenced tumor suppressor genes.
• Biochemical Assays: In vitro methyltransferase activity assays using recombinant PRC2 complexes (EZH1-PRC2 vs. EZH2-PRC2) and nucleosome substrates were performed to determine inhibitor specificity and potency (IC50).
• Genetic Models: CRISPR/Cas9 was used to generate isogenic EZH1-KO and EZH2-KO RKO cell lines. Inducible shRNA knockdown was used to deplete EZH1 in an EZH2-KO background to assess compensatory mechanisms.
• Epigenomic Profiling:
  • siQ-ChIP-seq: Quantitative chromatin immunoprecipitation sequencing for H3K27me1, H3K27me3, and H3K27ac to map genome-wide distribution changes.
  • Mass Spectrometry: Quantitative histone PTM mass spectrometry to validate global changes in methylation and acetylation states.
  • EM-seq (Enzymatic Methyl-sequencing): Performed on ChIP-enriched DNA (me1EM-seq and acEM-seq) to directly correlate histone marks with local DNA methylation status.
  • RNA-seq: Transcriptomic profiling to assess gene expression changes and pathway enrichment (GSEA).
• Functional Assays: Cell viability and longitudinal proliferation assays were conducted to measure therapeutic efficacy.

3. Key Contributions & Findings

A. Identification of EZH1-Dependent H3K27me1 as an Adaptive Barrier

• Screening Results: Dual inhibitors of EZH1 and EZH2 (Valemetostat/VAL, Tulmimetostat/TUL, Mevrometostat/MEV) showed superior synergy with DAC compared to EZH2-selective inhibitors (Tazemetostat/TAZ).
• Mechanism of Action: The most effective compounds (VAL, TUL, MEV) potently inhibited both EZH1 and EZH2, leading to the ablation of all three H3K27 methylation states (me1, me2, me3). In contrast, EZH2-selective inhibitors reduced me2/me3 but preserved or increased H3K27me1 at specific genomic loci.
• Genetic Validation:
  • EZH2-KO: Resulted in reduced H3K27me2/me3 but a compensatory increase in H3K27me1.
  • EZH1-KO: Had minimal effect on global methylation in wild-type cells.
  • Dual Depletion (EZH2-KO + EZH1-KD): Completely eliminated H3K27me1 and increased H3K27ac.
  • Conclusion: EZH1 acts as a "backup" methyltransferase that maintains H3K27me1 when EZH2 is inhibited, creating a barrier to full chromatin de-repression.

B. Context-Dependent Genomic Redistribution

• H3K27me1 Dynamics: H3K27me1 is not uniform; it marks both repressive Polycomb domains and actively transcribed gene bodies.
  • EZH2 Inhibition (TAZ): Causes a redistribution where EZH1 deposits H3K27me1 onto Polycomb-enriched regions (ReprPC, Bivalent promoters) that lost H3K27me3.
  • Dual Inhibition (VAL): Eliminates H3K27me1 across all contexts, including gene bodies and enhancers.
• DNA Methylation Correlation: H3K27me1-enriched regions are often highly DNA-methylated. Dual inhibition removes the repressive H3K27me1 mark without immediately altering DNA methylation, creating a unique chromatin state.

C. Induction of a Therapy-Associated Bivalent Chromatin State

• The "Bivalent" State: Dual EZH1/2 inhibition induces H3K27ac (an activating mark) at enhancers and promoters that remain DNA-methylated. This creates a "bivalent" state (H3K27ac + DNA methylation) where chromatin is poised for activation but silenced by DNA methylation.
• Resolution by DNMTi: The addition of DNMTi resolves this bivalency by removing DNA methylation, allowing the pre-loaded H3K27ac to drive robust transcriptional activation of tumor suppressors and immune pathways.
• Superiority of Dual Inhibition: EZH2-selective inhibition fails to fully resolve this state because it retains H3K27me1 at certain loci, which competes with or blocks the full activation potential even when DNA is demethylated.

D. Mechanism of Growth Suppression: Oncogenic Repression

• Transcriptional Programs: While combination therapy activates immune/inflammatory pathways (viral mimicry), the cancer cell-intrinsic growth suppression is primarily driven by the repression of oncogenic transcription programs (MYC, E2F, G2/M checkpoint).
• Coordinated Erosion: The suppression of MYC/E2F networks requires a coordinated loss of three epigenetic features:
  1. H3K27me1 loss across gene bodies (disrupting transcriptional fidelity/elongation).
  2. H3K27ac depletion at active promoters.
  3. Gene-body DNA hypomethylation (induced by DAC).
• Dependency on EZH1: High EZH1 expression correlates with poor patient survival in CRC and predicts sensitivity to dual EZH1/2 inhibition over EZH2-selective inhibition.

4. Significance

• Therapeutic Strategy: The study provides a mechanistic rationale for combining DNMT inhibitors with dual EZH1/2 inhibitors (rather than EZH2-selective inhibitors) to treat solid tumors. This combination overcomes the adaptive barrier of H3K27me1 maintenance.
• Biomarker Discovery: EZH1 expression levels are identified as a potential biomarker to stratify patients for PRC2-targeted therapies. High EZH1 expression suggests a reliance on the adaptive H3K27me1 barrier, necessitating dual inhibition.
• Epigenetic Plasticity: The work redefines the understanding of PRC2 function in cancer, highlighting that EZH1 is not merely a redundant paralog but a critical mediator of adaptive resistance that must be targeted to achieve durable epigenetic reprogramming.
• Clinical Translation: Given that dual inhibitors like Valemetostat are already in clinical trials, these findings support their evaluation in combination with DNMTi for colorectal and potentially other solid tumors.

Conclusion

The paper establishes that EZH1-dependent H3K27me1 is a critical adaptive barrier limiting DNMT inhibitor efficacy in colorectal cancer. Dual inhibition of EZH1 and EZH2 eliminates this barrier, induces a bivalent chromatin state that is resolved by DNMTi, and synergistically suppresses oncogenic transcription networks while activating tumor-suppressive pathways. This defines a new precision medicine strategy for epigenetic therapy in solid tumors.