Disrupting the ASH2L–DPY30 PPI in cancer: structure, function, and therapeutic opportunities in H3K4 methylation

Abstract

The ASH2L–DPY30 interaction is a structurally conserved and functionally essential component of the COMPASS family of histone methyltransferases responsible for H3K4 trimethylation. This minimalist helix–groove interface plays a critical allosteric role in stabilizing ASH2L, aligning the catalytic SET domain on nucleosomes, and enabling efficient methylation of chromatin targets. Recent structural, biochemical, and genetic studies have demonstrated that disrupting this contact—whether by point mutation, domain deletion, or competitive peptides—leads to widespread collapse of H3K4me3, transcriptional silencing of oncogenic programs, and suppression of cell proliferation, particularly in MLL-rearranged and MYC-driven cancers. In parallel, chemical-biology tools and fragment-based screening efforts have begun to yield the first ligandable scaffolds, setting the stage for drug discovery targeting this axis. This review synthesizes the current knowledge surrounding the ASH2L–DPY30 interface, covering its molecular architecture, catalytic importance, disease relevance, and therapeutic tractability. We also discuss resistance mechanisms, assay platforms, and the challenges and opportunities for translating this target into a first-in-class epigenetic therapy.

Introduction

Histone H3 lysine-4 (H3K4) methylation is a hallmark of active or poised chromatin and a key gate-keeper of transcriptional plasticity [1]. Because many tumors re-wire the H3K4 methyl-landscape, the writers, erasers and readers of this mark have become attractive drug targets in “epigenetic oncology” [1]. In mammals, H3K4 mono-, di- and trimethylation is catalyzed by six COMPASS/SET1 family holo-enzymes (SET1A/B, MLL1-4) [2–4]. Each complex is built around a catalytic SET protein and four obligate core subunits—WDR5, RBBP5, ASH2L and DPY30—that act together as an allosteric engine to achieve processive trimethylation [6]. Mechanistically, H3K4me3 is established in a transcription-coupled, stepwise manner in which COMPASS complexes are recruited to promoter regions and convert promoter-proximal H3K4me1/2 into H3K4me3 on nucleosomes flanking transcription start sites (TSSs) [7]. This promoter-centered H3K4me3 peak reinforces productive initiation/early elongation and provides a chromatin “landing pad” for reader proteins that couple chromatin state to transcriptional output. Functionally, different COMPASS paralogs preferentially support distinct gene programs (e.g., broad promoter marking at housekeeping genes versus developmental/lineage-specific transcriptional circuits), which helps explain why cancers driven by aberrant transcriptional states show heightened dependence on this pathway [7, 8]. Disrupting any one of these core proteins compromises global H3K4me3 and derails lineage-specific gene expression programs [5].

Among the core constituents, the ASH2L–DPY30 pair forms a discrete helical bundle in which the 30-residue SDI (SDC–DPY30-interacting) helix of ASH2L docks into the hydrophobic groove of the DPY30 homodimer (Fig. 1) [9, 10]. Deleting the SDI helix, mutating the groove, or out-competing it with cell-penetrant SDI mimetics such as MM-401 or MM-589 abolishes DPY30-driven stimulation of COMPASS on nucleosomal—but not free-histone—substrates; orthogonal WDR5 antagonists (WDR5-0103, piribedil, OICR-9429) compound this effect by further destabilizing the WRAD platform (Fig. 1) [11]. At a structural level, DPY30 acts as an ASH2L “molecular clamp”, tightening otherwise flexible ASH2L lobes, raising its melting temperature by ~ 20 °C and forging transient bridges to RBBP5 that consolidate the catalytic core [9, 10].

Fig. 1.

Domain framework and chromatin-level model for WRAD-dependent H3K4 trimethylation and its pharmacologic disruption. A Schematic domain architecture of the human MLL1 core complex components, highlighting major structural modules in RBBP5, WDR5, MLL1 (SET catalytic domain), ASH2L (PHD–winged-helix and split-SPRY with intervening intrinsically disordered regions, IDRs), and DPY30 [12]. B Disrupting the ASH2L–DPY30/WDR5 axis collapses COMPASS processivity at MLL-regulated promoters [8]. Left, recruitment/priming cues (illustrated schematically) include PHD-mediated recognition of unmodified H3K4 (H3K4me0) and H2B monoubiquitination (H2Bub1), which facilitate robust promoter-proximal H3K4 trimethylation (H3K4me3, blue arrow) at transcription start sites of MLL target genes, sustaining high transcriptional output. Centre, competitive groove-binding peptides (MM-401, MM-589) or small-molecule WDR5 antagonists (WDR5-0103, piribedil, OICR-9429) dissociate DPY30 or WDR5 from the scaffold. Right, loss of DPY30 (labelled “DPY0”) locks the complex in a hypomethylating state that deposits only H3K4 dimethyl marks (H3K4me2, green arrows), leading to repression of MLL-responsive loci (red “X”). Ub denotes monoubiquitin on H2B. Nucleosomes are shown as blue cylinders; colored spheres represent individual WRAD/SET subunits [11]

Physiologically, the ASH2L–DPY30 axis safeguards stem-cell identity, lineage commitment and cell-cycle progression [13–15]. Conditional deletion of Dpy30 in the mouse hematopoietic system cripples H3K4me3 at blood-lineage genes, exhausts stem-cell pools and aborts MLL-AF9-driven leukemogenesis—yet causes surprisingly mild systemic toxicity, hinting at a therapeutic window. Outside the marrow, DPY30 loss induces premature senescence in fibroblasts and apoptosis in colorectal-carcinoma (CRC) cells through Bax/Bcl-2 rewiring and caspase activation, underscoring its broader relevance [18–20].

The same interface that secures developmental homeostasis becomes an oncogenic liability when hijacked by high-MYC states or MLL translocations [13, 21]. Cell-penetrating SDI-mimetic peptides that out-compete endogenous ASH2L curb proliferation of MLL-rearranged leukemias, MYC-driven lymphomas and other blood cancers, while sparing K562 and CD34⁺ progenitors [17, 21]. In murine models, partial Dpy30 haplo-insufficiency synergizes with BET and EZH2 inhibition, illustrating how pharmacologic or genetic disruption of the groove can amplify existing epigenetic therapies [22]. Collectively, these insights place the ASH2L–DPY30 contact point at the crossroads of chromatin biochemistry, stem-cell biology and cancer vulnerability. Yet no small-molecule inhibitor of the groove has entered clinical pipelines, and fundamental questions remain—How universal is DPY30 addiction outside hematopoietic malignancies? Can interface blockade be achieved with drug-like scaffolds or targeted degraders? Which compensatory methyltransferases or reader domains blunt therapeutic responses?

In this review we synthesize structural, biochemical and biological knowledge accrued over the past decade, with a special focus on strategies that disrupt or inhibit the ASH2L–DPY30 axis. We begin by dissecting the molecular architecture of the interface, proceed through genetic and chemical methods of perturbation, survey the phenotypic fallout across model systems, and conclude with an outlook on drug-development frontiers and future research priorities. In doing so, we aim to provide a conceptual and practical roadmap for harnessing this “methyl-clamp” as a next-generation epigenetic target.

Structural basis of the ASH2L–DPY30 interface

Modular architecture of the binding partners

Both ASH2L and DPY30 display a modular domain architecture that pre-organizes their high-affinity encounter (Fig. 2A) [9]. ASH2L opens with an N-terminal reader cassette—a tandem PHD finger and winged-helix (PHD–WH) fold—that can sense unmethylated H3K4 (H3K4me0) and touch nucleosomal DNA, potentially acting as a proofreading checkpoint before catalytic commitment [4]. Next comes the split-SPRY domain (residues 151–400), which serves as a landing pad for RBBP5 and CXXC1/CFP1. Cryo-EM and crystallographic data show that this segment is conformationally plastic in isolation but rigidifies once ASH2L latches onto DPY30 [4]—a transition captured by the ordered β-strands and α-helices in Fig. 2D. A loop-rich intrinsically disordered region (IDR; residues 401–439) follows, draping over nucleosomal DNA and likewise gaining structure upon DPY30 binding [23]. The C-terminus culminates in the SDI (Sdc-DPY30-Interacting) helix (residues 504–534). Its amphipathic Φ-x-Φ-x-x-ΦΦ motif slots into the elongated hydrophobic groove formed by the DPY30 homodimer, as seen in the 2 : 1 complex extracted from PDB 4RIQ (Fig. 2B–C, E) [24]. DPY30 itself is architecturally economical: residues 45–99 of each monomer fold into a four-helix “X-bundle” that dimerizes to create the requisite groove [24]. This dimer clamps a single ASH2L SDI helix, producing a defined stoichiometry of two DPY30 monomers per one ASH2L tail (Fig. 2B, E). The resulting interface buries ~ 1,100 Ų of surface area and is enriched in hydrophobic and π-stacking contacts, underlining why even subtle helix-groove perturbations abolish DPY30-dependent stimulation of COMPASS activity [9].

Fig. 2.

Structural organization of the ASH2L–DPY30 module [9, 24]. A Linear domain schematics of WRAD subunits drawn to scale. RBBP5 contains seven WD-40 repeats plus accessory binding motifs (ABM, WBM). ASH2L comprises an N-terminal PHD–winged-helix reader, a split-SPRY scaffold, a loop-rich intrinsically disordered region (Loop), and the C-terminal SDI helix within the DPY30-binding motif (DBM). DPY30 contributes a dedicated dimerization/docking (DD) region that forms the “X-bundle.” B Ribbon view of the 2 : 1 DPY30 (blue) : ASH2L (orange) assembly (PDB 4RIQ) illustrating the overall architecture. C Molecular surface of the same complex, colored as in B, emphasizing the elongated hydrophobic groove of the DPY30 dimer and the amphipathic nature of the ASH2L SDI helix. D Detail of the split-SPRY domain (orange) with secondary-structure elements annotated; the dashed outline marks flexible Loop F′ that becomes ordered upon DPY30 engagement. E Overview of ASH2L (orange) docked onto DPY30 (blue) within the full WRAD context; inset, hydrophobic and polar side-chains at the SDI-groove interface are labelled. B–E were rendered from PDB ID 4RIQ [9, 24]

Crystal snapshots of the interface

Crystal and cryo-EM snapshots of the interface

Atomic-level insight into the ASH2L–DPY30 contact now comes from three complementary structures that capture the same helix-in-groove arrangement in progressively larger contexts. Minimal heterotrimer (PDB 4RIQ, 2.23 Å, X-ray) – residues 510–523 of ASH2L form an amphipathic α-helix (the SDI motif) that inserts as a rigid “key” into a pre-formed four-helix “socket” created by the DPY30 dimerization/docking (DD) module (residues 45–99). The helix makes four hydrophobic contacts lining pockets P₁–P₄, while acidic side chains on ASH2L engage basic patches at the mouth of the groove. No conformational change is detected in DPY30, establishing a lock-and-key mechanism [24]. Split-SPRY–extended complex (PDB 6E2H, 2.24 Å, X-ray) – the same SDI helix is visualized in one crystal lattice together with the contiguous split-SPRY domain and part of the loop-IDR. Comparison with the unbound split-SPRY shows that docking of DPY30 orders two otherwise flexible loops and straightens the C-terminal tail of ASH2L, rationalizing the ∼3-fold increase in CD thermal stability reported for the complex [10].

High-resolution cryo-EM of the MLL1 core bound to an H2B-ubiquitylated nucleosome (PDB 6KIU; 3.2 Å) reveals that the ASH2L SDI helix threads across the four hydrophobic pockets (P₁–P₄) of the DPY30 dimer exactly as seen in crystal structures of the isolated pair (Fig. 3A–C, E–F) [25, 26]. Superposition shows neither translational shift nor backbone distortion of the DPY30 “X-bundle”, confirming its role as an invariant scaffold rather than an inducible clamp. Consequently, the SDI helix retains the same Φ-x-Φ-x-x-ΦΦ register and a fixed 1 : 2 ASH2L : DPY30 stoichiometry across solution (PDB 6E2H), substrate-free holo-enzyme, and nucleosome-bound states (Fig. 3E–F) [10]. This convergence rules out any major induced-fit rearrangement and identifies DPY30 as a pre-molded docking socket whose geometry is preserved from solution to chromatin. Notably, the same groove is exploited by unrelated partners such as BAP18 and AKAP95, reinforcing its appeal as a rigid target for structure-guided inhibitor or degrader design [25].

Fig. 3.

The DPY30–SDI interface is pre-molded in solution and unchanged on chromatin. A Cryo-EM ribbon model of the complete MLL1 core (WDR5–RBBP5–ASH2L–DPY30) bound to an H2B-ubiquitylated nucleosome (PDB 6KIU; 3.2 Å). DNA is cyan; histones are colored by type (H2A green, H2B magenta, H3 navy, H4 slate); WRAD subunits are colored orange (ASH2L), yellow (DPY30), dark blue (RBBP5), and teal (WDR5). B Molecular surface view of the same assembly highlights the compact fit of WRAD against the nucleosomal disc. C Cut-away view emphasizing the DPY30 dimer clasping the ASH2L SDI helix while the ASH2L split-SPRY extension and intrinsically disordered region (IDR) drape over nucleosomal DNA; for clarity, the SDI helix is redrawn beneath the density. D Close-up of the basic ridge on ASH2L that contacts DNA and histone tails; key lysines are indicated. E Ribbon representation of the isolated DPY30–SDI crystal complex (PDB 6E2H) showing the same 2:1 DPY30:ASH2L stoichiometry and the conserved ΦxΦxxΦΦ helical register observed in the nucleosome-bound state. F Surface rendering of DPY30 bound to the SDI helix showing the four hydrophobic sub-sites that accommodate SDI side chains. Boxed regions delineate each pocket al.ong the groove, and right-hand insets provide close-up views of P_I–P_IV with representative SDI residues and groove-lining DPY30 residues annotated. This pocket architecture is already formed in the isolated complex and is preserved upon nucleosome engagement, consistent with a lock-and-key helix-in-groove interface [10, 25, 26]

Beyond these representative models, the same helix-in-groove SDI interaction is observed across a broader structural literature. Additional cryo-EM reconstructions of human MLL1–WRAD bound to nucleosomes (e.g., PDB 6PWV) likewise place the ASH2L SDI helix in the DPY30 groove, while later multistate and integrative cryo-EM/NMR/SAXS studies (e.g., PDB 6W5M/6W5N) further illuminate how DPY30-bound ASH2L elements stabilize chromatin engagement and restrict rotational dynamics. Importantly, homologous yeast COMPASS structures (Bre2–Sdc1; e.g., PDB 6CHG and nucleosome-bound PDB 6VEN) reveal a conserved helix–groove architecture, underscoring the evolutionary robustness of this docking mode [4, 12, 27, 28].

Energetic hot-spots validated by mutagenesis

Structure-guided mutagenesis pinpoints a compact cluster of hydrophobic groove residues on DPY30 (Val62, Leu66, Leu69) that dominate the binding free energy across the ASH2L–DPY30 seam; substitutions at these positions progressively weaken the interaction, with Leu69Asp abrogating binding in vitro and disrupting function in vivo/yeast assays [24]. Consistent with DPY30’s role in promoting higher H3K4 methylation states on nucleosomes, disrupting the ASH2L–DPY30 contact compromises productive accumulation of H3K4me2/me3 relative to lower states [4]. Complementary mutagenesis of the ASH2L SDI helix identifies key anchoring residues (including L513/L517/V520) required for DPY30 recognition; SDI-helix mutants and ΔSDI constructs eliminate DPY30 binding and phenocopy loss of the interface [9, 29]. Together, these datasets map the energetic “hot spots” that underpin the crystallographic lock-and-key geometry of the complex and highlight fixed nodes suitable for structure-based disruption of the interface [10, 24].

Allosteric consequences of binding

Small-angle X-ray scattering, cross-linking MS and cryo-EM converge on the idea that DPY30 acts as a molecular clamp: it compacts otherwise flexible ASH2L lobes, raises the melting temperature by ~ 18 °C, and locks COMPASS onto the nucleosome in a productive orientation. As a result, k_cat on nucleosomal substrates rises ~ 18-fold compared with complexes lacking DPY30, whereas activity on free H3 tails improves only two-fold—underscoring the nucleosome-specific allosteric gain [9, 10, 30]. The structural modules in Table 1 provide the architectural basis for this clamp.

Table 1.

Selective growth Inhibition by DPY30-binding ASH2L-derived peptides in hematopoietic cell models

Cell type	Oncogenic driver	Assay supported in Shah et al. (2019)	WT peptide	Y518R peptide	3R control	Refs.
MOLM-13 (AML)	MLL-AF9	Growth in liquid culture (10 µM q12h) + clonogenicity	inhibits	inhibits (often stronger)	no effect	[17]
THP-1 (AML)	MLL-AF9	Growth in liquid culture (10 µM q12h)	inhibits	inhibits (often stronger)	no effect	[17]
P493-6 (BL model)	MYC ON/OFF	Growth in liquid culture (10 µM q12h)	inhibits	inhibits	no effect	[17]
K562 (CML)	BCR-ABL	Growth in liquid culture (10 µM q12h)	no significant effect	no significant effect	no significant effect	[17]
CD34⁺ HSPC	–	Growth in liquid culture (10 µM q12h)	no significant effect	no significant effect	no significant effect	[17]

Cells were continuously exposed to cell-penetrating ASH2L-derived peptides that either bind DPY30 (WT SDI peptide and the higher-affinity Y518R variant) or do not bind DPY30 (3R control peptide). Viability/proliferation was assessed after 7 days of treatment. Where reported, GI₅₀ denotes the peptide concentration that reduces viable cell number by 50% relative to vehicle at day 7. For fixed-dose experiments, results are summarized qualitatively as growth Inhibition (“inhibits”) or no significant effect (“no effect”) at the indicated dosing regimen. “>20” indicates that 50% growth Inhibition was not reached at the highest concentration tested. Each cell model is annotated with its dominant oncogenic driver to highlight the Preferential sensitivity of MLL-rearranged and MYC-dependent contexts

Interface plasticity and implications for inhibitor design

Pocket-breathing simulations reveal that P_II and P_III expand by ~ 1.8 Å on SDI engagement, explaining why stapled peptides and helix-mimetics can still fit, whereas P_I/P_IV remain rigid—useful anchoring points for fragment campaigns [10, 17, 31]. Conservation of the Φ-x-Φ-x-x-ΦΦ motif in BAP18 and AKAP95 warns of potential selectivity liabilities; nevertheless, exploiting ASH2L-specific flanking residues (e.g., Trp505) or adjacent WRAD surfaces absent from NURF/ATAC should allow ASH2L-biased inhibitors [10, 17].

Chromatin and enzymology consequences of modulating the ASH2L–DPY30 axis

Disruption of the ASH2L-DPY30 contact reverberates far beyond a single protein–protein interaction: it rewires enzyme kinetics on nucleosomes, alters chromatin accessibility, and shifts transcriptional outputs [10, 29].

DPY30 preferentially stimulates nucleosomal H3K4 methylation—chiefly on nucleosomal substrates

Reconstituted MLL1 core complexes possess only limited catalytic power when assayed on free histone peptides or octamers, and adding DPY30 leaves these non-chromatin substrates only modestly affected: on an H3(1–9) peptide, kcat increases by ~ 1.2-fold (with ~ 1.8-fold higher catalytic efficiency), and on histone octamer/H3–H4 tetramer the overall methylation rate increases only ~ 1.6-fold [31]. The picture changes dramatically on nucleosomes: DPY30 boosts MLL1 activity on the nucleosome core particle by ~ 18-fold (with substantially higher catalytic efficiency), and this nucleosome-restricted mode of stimulation generalizes across the MLL/SET1 family [4, 31]. This sharp kinetic contrast is only observed when the DPY30 dimerization/docking module engages the C-terminal DPY30-binding motif (SDI/DBM) of ASH2L: disrupting that helix-in-groove interaction (e.g., ASH2L “3E” mutant that abolishes DPY30 binding, or DPY30 dimerization-defective mutant) eliminates DPY30-dependent stimulation [31].

A dual mechanism: global compaction plus nucleosome-specific alignment

DPY30 enhances WRAD-dependent methyltransferase activity through two complementary mechanisms that operate at different structural scales (Fig. 4) [32]. First, in a nucleosome-independent manner, DPY30 functions as a molecular clamp on ASH2L. Cross-linking mass spectrometry, nano-differential scanning fluorimetry (nano-DSF), and light-scattering experiments show that DPY30 binding triggers a cooperative folding event within ASH2L, raising its thermal melting point from ~ 32 °C to ~ 50 °C and drastically reducing aggregation [31]. This DPY30-induced compaction strengthens intramolecular contacts between ASH2L and RBBP5, pre-assembling a “tense” WRAD scaffold that is catalytically primed, leading to modest increases in turnover even in the absence of chromatin (Fig. 4A).

Fig. 4.

DPY30 activates MLL/SET1 methyltransferases through a two-tiered mechanism. A Nucleosome-independent compaction. In the apo state (left) the ASH2L N-terminal lobe (yellow) is highly flexible and loosely associated with the WRAD core (green outline; MLL1 = red, RBBP5 = green, WDR5 = blue). Binding of a DPY30 dimer (purple) to the ASH2L SDI helix clamps the two ASH2L lobes together, releases the disordered PHD-WH arm, and pre-organizes the WRAD module into a “tense” but catalytically competent configuration (right). The resulting complex is smaller, shows a ~ 18 °C upward shift in melting temperature, and displays reduced aggregation. B Nucleosome-specific alignment. Without DPY30 (left) the MLL1 SET domain swings like a pendulum above the nucleosome, limiting productive encounters with the H3 tail. DPY30 binding (right) rigidifies the ASH2L linker, fixes the SET domain directly over the nucleosomal dyad, and promotes additional anchoring contacts between ASH2L and both the H3 N-terminus (brown) and nucleosomal DNA (gold ribbon). The locked geometry enables rapid, processive trimethylation of H3K4

Second, DPY30 exerts a nucleosome-specific effect by acting as a geometric brace. Cryo-EM and small-angle X-ray scattering reconstructions reveal that, without DPY30, the MLL1 catalytic lobe behaves like a flexible pendulum relative to the nucleosome surface. Re-introduction of DPY30—or its minimal SDI helix—freezes this motion and rigidly positions the SET domain directly above the nucleosomal dyad, where it can processively trimethylate H3K4 [31]. Importantly, additional anchoring contacts between ASH2L and both the H3 N-terminal tail and nucleosomal DNA emerge only in the presence of DPY30, providing a structural rationale for the dramatic increase in catalytic rate (k_cat) observed in nucleosomal methylation assays (Fig. 4B) [32]. Together, these findings establish DPY30 as a multifaceted activator that both compacts the WRAD core and locks the enzyme in an optimal orientation on chromatin, thereby maximizing methyltransferase efficiency.

ASH2L IDRs are essential conduits of the DPY30 signal

Deletion of the Linker or Loop IDRs abolishes DPY30-dependent stimulation without affecting the SDI groove itself. NMR shows that DPY30 binding draws these disordered segments into a compact cluster (ΔR_g ≈ 13 Å) that buttresses the split-SPRY domain [4, 9, 29]. Mutating the triple hot-spot L513E/L517E/V520E in SDI uncouples DPY30, collapses ASH2L stability and mirrors the catalytic deficit, underscoring an “all-or-nothing” dependency between the groove and the flanking IDRs [4, 9, 29]. To make the mutation–enzymology link explicit, Fig. 5 summarizes representative reconstituted assays demonstrating that DPY30 acts primarily as a chromatin-selective processivity factor: DPY30 has little impact on free H3 substrates, but markedly increases accumulation of H3K4me3 on nucleosomes, consistent with its role in enabling the final catalytic step from H3K4me2 to H3K4me3. Importantly, disrupting the ASH2L–DPY30 communication axis—either by removing the ASH2L SDI motif or by deleting key ASH2L intrinsically disordered linker/loop segments—selectively blunts DPY30-driven stimulation and shifts product distribution toward lower methylation states, mirroring the “stall after the second catalytic cycle” phenotype observed for groove/SDI hot-spot mutations discussed in Sect."Energetic hot-spots validated by mutagenesis". Collectively, these data provide a direct mechanistic bridge between interface integrity, nucleosome-restricted kinetic enhancement, and IDR-mediated transmission of the DPY30 signal [4].

Fig. 5.

DPY30- and ASH2L-IDR–dependent control of nucleosome-selective H3K4 methylation and the H3K4me2→H3K4me3 transition [4]. a DPY30 specifically stimulates MLL1 activity on the NCP. In vitro histone methyltransferase (HMT) assay for the MLL1 core complex using either nucleosome core particles (NCPs) or recombinant histone H3 as substrates (indicated at top). The MLL1 core complex (MLL1^SET, WDR5, RbBP5 and ASH2L) was assayed in the absence or presence of DPY30. Reaction products were resolved by 15% SDS–PAGE and immunoblotted with antibodies to H3K4me1, H3K4me2 and H3K4me3 (right). Coomassie staining is shown as a loading control (bottom). b DPY30 requires ASH2L intrinsically disordered regions (IDRs) to stimulate MLL1 activity on chromatin. Human ASH2L truncation/deletion mutants used in the HMT assays. ASH2L contains an N-terminal PHD–WH (plant homeotic domain–winged helix) domain and a C-terminal split-SPRY domain, and includes three IDRs: Linker (179–275), Loop (401–439) and SDI (504–534). c In vitro HMT assay for the MLL1 core complex reconstituted with wild-type or ΔSDI ASH2L. d In vitro HMT assay for the MLL1 core complex reconstituted with wild-type or ASH2L mutants as indicated. e In vitro HMT assay for the MLL1 core complex reconstituted with wild-type or ΔLoop ASH2L. f Identification of essential ASH2L IDRs in DPY30-mediated stimulation. Human ASH2L deletion mutants and transactivation peptides used for mapping are indicated. Adapted from Lee et al. [4].

Genome-wide chromatin consequences of axis disruption

Disruption of the ASH2L–DPY30 axis has profound and widespread effects on chromatin structure and transcriptional activity. In mouse embryonic fibroblasts (MEFs), genetic ablation of either ASH2L or DPY30 leads to a near-complete loss of promoter-proximal H3K4me3. This depletion is accompanied by a measurable tightening of nucleosome spacing, particularly at CpG-rich transcription start sites, and displaces the chromatin insulator protein CTCF from approximately 1,000 active promoters [29]. ATAC-seq profiling reveals a global reduction in chromatin accessibility—median losses of about 27%—in regions that experience H3K4me3 depletion, thereby linking this methyl mark directly to the maintenance of open promoter architecture.

Functionally, this epigenetic collapse results in marked transcriptional repression. In MEFs lacking ASH2L, over 1,200 genes show greater than twofold downregulation. Similarly, in embryonic stem cells (ESCs), DPY30 knockdown via shRNA leads to impaired de novo H3K4me3 deposition and a failure to activate critical genes involved in neural and mesodermal differentiation. Remarkably, synthetic recruitment experiments using dCas9-tethered DPY30 demonstrate that DPY30 alone is sufficient to deposit H3K4me3 at naïve loci in ESCs, proving that it acts not merely as a permissive cofactor but as an instructive element in methyl mark deposition [10, 29]. These disruptions are not confined to genetic manipulations. In human cells, deletion of the SDI helix from ASH2L causes a shift in methyl-state balance, reducing H3K4me3 and increasing H3K4me1 levels. This suggests a stall in methylation progression, with enhancer priming maintained but activation impaired. In disease models, such as MLL-AF9 acute myeloid leukemia (AML), chemical displacement of DPY30 using peptide inhibitors causes H3K4me3 levels to collapse at MYC super-enhancers, leading to suppression of the oncogenic transcriptional program.

Overall, these observations reveal a consistent pattern: disruption of the ASH2L–DPY30 interaction axis compresses promoters, suppresses transcription, and alters histone methylation balance. Notably, while H3K4me3 is acutely sensitive to loss of DPY30 function, H3K4me1 often persists at enhancer regions, highlighting a methyl-state hierarchy that is selectively dependent on the integrity of this interaction interface [33, 34].

Overall, DPY30 does far more than “tag along” with COMPASS: it clamps ASH2L, rigidifies WRAD, aligns the complex on the nucleosome, and thereby shifts both enzymology and chromatin landscape [29, 35]. Conversely, breaking the groove—whether by mutation or peptide—uncouples these layers, converting active chromatin into a repressed, less accessible state [29, 35].

Genetic and biochemical strategies to break the ASH2L–DPY30 axis

Disrupting the interface can be achieved at three levels—genome engineering, peptide/peptidomimetic competition and in vitro assay reconstitution [9, 10, 29].

Loss-of-function genetics: from point mutants to conditional knock-outs

A wide array of genetic strategies has been employed to interrogate the functional significance of the ASH2L–DPY30 interaction axis, from targeted point mutations to complete gene deletions. These experiments consistently highlight the axis as both a regulatory linchpin and a potential therapeutic vulnerability. At the molecular level, acute point mutations in either binding partner can abolish complex formation. For instance, CRISPR knock-in of a leucine-to-aspartate substitution in DPY30 (L69D), which disrupts the core hydrophobic groove, or a triple mutation in ASH2L (L513E/L517E/V520E), which alters the SDI helix, results in complete loss of co-immunoprecipitation between the two proteins. These perturbations collapse global H3K4me3 levels and provoke a compensatory increase in H3K4me1, phenocopying the effect of deleting the SDI helix altogether [31, 36, 37]. The resulting methylation imbalance indicates that the structural interaction is required to progress beyond mono- and dimethylation states during histone modification.

Deletion of specific ASH2L domains yields similar loss-of-function outcomes. Removing the 30-residue SDI helix (residues 504–534) abolishes DPY30-mediated stimulation in both cis (within the same molecule) and trans (in reconstituted systems). Likewise, deleting either of ASH2L’s intrinsically disordered regions—the linker (residues 178–275) or the loop region (residues 400–440)—disrupts DPY30 responsiveness. These results confirm that ASH2L’s flexible segments are not merely passive linkers but serve as essential conduits for structural and functional communication within the WRAD complex [4]. Beyond these targeted mutations, broader gene-silencing approaches further underscore the biological relevance of the interface. In MLL-AF9 leukemia cell lines, shRNA-mediated knockdown of DPY30 reduces both proliferation and clonogenic potential, while having little effect on K562 cells driven by alternative oncogenic programs [14, 16, 17]. In mouse models, tamoxifen-inducible deletion of Dpy30 in MLL-AF9-transformed hematopoietic stem cells completely blocks leukemia progression and eradicates leukemia-initiating cells [16]. Notably, partial loss of Dpy30—via heterozygous knockout—delays lymphomagenesis in Eµ-Myc mice without causing overt toxicity, hinting at a potential therapeutic window for selective inhibition [16].

Interestingly, in non-hematopoietic contexts, the outcome of axis disruption is different. Conditional knockout of Ash2l in mouse embryonic fibroblasts does not trigger apoptosis but instead induces a senescence-like state, marked by activation of the p38 MAPK pathway and broad transcriptional repression [19, 20, 37, 38]. These findings suggest that the ASH2L–DPY30 interaction is a context-dependent regulatory node—one that functions as an oncogenic enabler in hematopoietic malignancies but as a developmental necessity in somatic lineages. Collectively, this body of work demonstrates that the ASH2L–DPY30 interface is genetically vulnerable across multiple systems. Its disruption compromises H3K4 trimethylation, remodels chromatin, and either halts cell division or initiates senescence, depending on cellular context. These dual roles position the interface as both a mechanistic cornerstone in chromatin regulation and a promising target for therapeutic intervention.

Peptide and peptidomimetic competition

Disrupting the ASH2L–DPY30 interaction pharmacologically has proven feasible using short peptides that mimic the SDI helix. A key breakthrough came with the development of cell-penetrating 14-mer peptides derived from the C-terminal SDI region of ASH2L. These peptides, once introduced into cells, can displace DPY30 from the COMPASS complex and selectively kill leukemia cells that are dependent on high MYC expression or harbor MLL rearrangements. Notably, they spare non-transformed cells such as K562 erythroleukemia lines and normal CD34⁺ hematopoietic progenitors [13, 21].

The baseline peptide—corresponding to the wild-type SDI sequence QRVVLIFEEYILR—binds DPY30 with moderate affinity and is sufficient to inhibit growth in MLL-rearranged acute myeloid leukemia (AML) models, such as MOLM-13. A modified version of this peptide, where tyrosine 518 is substituted by arginine (Y518R), exhibits approximately six-fold tighter binding to the DPY30 groove and shows enhanced cytotoxicity in both MOLM-13 and Burkitt lymphoma-derived P493-6 cells. Conversely, a negative control peptide, termed the “3R” variant, substitutes the core hydrophobic residues with arginines (QRRRLIRGERIRR), eliminating DPY30 binding and abolishing any growth-inhibitory effect [39].

Importantly, these peptides are not only effective as monotherapies. Synergy screens show that the high-affinity Y518R peptide has additive effects when combined with BET inhibitors like JQ1 or epigenetic modifiers such as EZH2 inhibitors. This suggests that targeting the DPY30 groove could be a valuable component in rationally designed combination regimens for aggressive leukemias [17]. Thus, these studies validate the DPY30 groove as a druggable interface and demonstrate that even minimalist, helix-mimicking peptides can achieve functional antagonism of the ASH2L–DPY30 interaction in disease-relevant settings.

Biochemical and biophysical read-outs for interface disruption

To rigorously evaluate the effects of both genetic mutations and chemical inhibitors targeting the ASH2L–DPY30 interface, a set of well-established biochemical and biophysical assays has become standard in the field. These complementary approaches assess functional, structural, and cellular consequences of disrupting the interaction and provide quantitative benchmarks for inhibitor potency and mechanistic validation.

One of the most direct methods involves measuring histone methyltransferase activity in vitro using either HMTase-Glo assays or traditional western blotting on reconstituted mononucleosomes. These assays quantify the enzymatic activity of COMPASS complexes, particularly their ability to trimethylate H3K4 in a DPY30-dependent manner. Mutations that disrupt the SDI–groove interface typically lead to a substantial reduction in catalytic efficiency, often exceeding a 15-fold decrease in apparent K_cat [31].

A second, more targeted approach involves fluorescence polarization anisotropy (FPA) using a fluorescein-labeled SDI peptide. This technique provides a sensitive, quantitative measure of binding affinity between ASH2L-derived helices and the DPY30 groove. Wild-type peptides yield dissociation constants in the sub-micromolar range, while competition assays can confirm the displacement of labeled probes by higher-affinity variants or small-molecule mimetics. This makes FPA particularly valuable for screening and optimizing groove-targeted inhibitors. Finally, cell-based assays such as co-immunoprecipitation (co-IP) and chromatin immunoprecipitation followed by quantitative PCR (ChIP-qPCR) allow researchers to track the disassembly of COMPASS in living cells. These techniques are routinely used to monitor the loss of interaction between ASH2L and DPY30, or the depletion of H3K4me3 at target promoters following peptide treatment, knockdown, or genetic deletion [20, 40].

Comparative efficacy and safety considerations

Genetic ablation achieves near-complete interface loss but carries lineage-specific liabilities (senescence in fibroblasts) while still offering a therapeutic window in hematopoietic cancers (Table 5) [17, 40]. Peptidomimetics are less penetrant and short-lived, yet al.ready spare normal progenitors and synergize with existing epigenetic drugs (Table 6). Both approaches underscore the druggability of the groove and provide quantitative benchmarks—loss of H3K4me3 ≥ 70% or viability IC₅₀ ≈ 5 µM—as go/no-go criteria for emerging small molecules [10].

Table 5.

Chemical-biology probe toolbox for the ASH2L–DPY30 interface

Probe class	Read-out/modality	Typical throughput	Primary application	References
FITC-SDI peptide tracer	Fluorescence-polarization (FP)	384-well (Z′ ≈ 0.7)	Ligand binding & displacement	[17]
Tb/FITC donor-acceptor pair	TR-FRET	384-well	High-throughput screening	[66]
Streptavidin/anti-His beads	AlphaScreen/LISA	1536-well	Ultra-HTS, fragment triage	[67]
nano-DSF reverse melt	Intrinsic fluorescence	Tube-strip/384 cap	Fragment scouting	[17]
HiBiT-CETSA DPY30	Luminescence vs. temp.	96-well	Target engagement in cells	[57]
dCas9-DPY30 fusion	ChIP-qPCR/RNA-seq	Locus-specific	Sufficiency tests, enhancer editing	[69]
DPY30-TurboID	MS-proteomics	—	Dynamic interactome mapping	[70]

Table 6.

Documented or plausible resistance routes after ASH2L–DPY30 disruption

Resistance layer	Molecular change	Experimental evidence status	Drug-combination countermeasure	Ref.
Writer/network rewiring (compensation)	Increased reliance on other H3K4 writer complexes (SETD1A/SETD1B; KMT2C/D) when COMPASS output is perturbed	Plausible for ASH2L–DPY30 disruption; DPY30/ASH2L regulate chromatin function of multiple MLL/SET1 complexes. SETD1A/SETD1B dependencies are documented in AML/KMT2A-r (not a DPY30-specific “fitness rebound” result).	Combine groove disruption with menin–KMT2A inhibitors (KMT2A-r) and/or exploit SETD1A degradation concepts where relevant	[4, 31, 79, 80]
Eraser rewiring (KDM5/LSD1 axis)	Adaptive changes in KDM5A/B (H3K4me3 demethylases) and/or LSD1/KDM1A (H3K4me1/2 demethylase)	Documented broadly in drug tolerance/resistance biology; not demonstrated as a DPY30-specific acquired resistance route	Combine with KDM5 or LSD1 inhibitors if profiling indicates adaptive upregulation/activity	[81, 82]
Polycomb remodeling (PRC2 rebalancing)	Increased PRC2 repression (H3K27me3 gain at loci losing H3K4me3)	Mechanistically supported in bivalent promoter biology; clinically relevant precedent exists in menin-inhibitor resistance (not DPY30-specific)	Consider EZH2/EED inhibitors only if PRC2 activity increases and is required	[83, 84]
Remodeler/HDAC rerouting (NuRD/CHD4)	Increased NuRD/CHD4/HDAC influence when H3K4 methylation is reduced	Plausible: NuRD binding can be favored when H3K4 is unmodified; CHD4 regulates enhancer accessibility in cancer contexts	Add HDAC1/2 inhibitors (or other chromatin accessibility modulators) if NuRD/HDAC dependence emerges	[85, 86]
On-target groove mutation (DPY30 interface)	Mutations in DPY30 groove residues that weaken SDI-helix binding	Mechanistically documented (mutations disrupt binding/function), but no published evidence for “resistant lines after peptide escalation” or specific selected clinical mutations	Speculative: next-gen ligands engaging additional pocket features or alternative binding modes	[10, 24]
TF circuit reactivation (MYC/E2F programs)	Re-establishment of proliferative MYC/E2F programs via alternative enhancer/promoter control	Documented in related KMT2A/menin resistance contexts; not yet proven for DPY30 inhibitors	Combine with BET inhibitors and/or CDK7 inhibitors when MYC/E2F rebound is observed	[84, 87, 88]

Peptide and peptidomimetic inhibitors of the ASH2L–DPY30 interface

Building on the proof-of-concept studies summarized in Sect. "Genetic and biochemical strategies to break the ASH2L–DPY30 axis", this section consolidates what is known about SDI-mimetic peptide architecture, quantitative target engagement, intracellular behavior, and optimization opportunities toward next-generation inhibitors [9, 17].

Design principles: grafting the SDI helix onto a delivery scaffold

SDI-mimetic inhibitors exploit a simple design rule: an amphipathic α-helix presents a hydrophobic face that docks into the DPY30 groove pockets (P_I–P_IV), while polar/charged residues remain solvent-exposed and can be adapted for delivery and tracking. Potency tuning is achieved primarily by strengthening hydrophobic packing within the groove (e.g., substitutions that improve engagement of the central pockets), whereas replacing core hydrophobes abolishes binding and provides a stringent negative control. Together, these results reinforce that groove recognition is driven mainly by side-chain packing rather than global helicity [10].

Biophysical confirmation of groove engagement

Peptide engagement of the DPY30 groove has been quantified using competitive fluorescence polarization (FP) assays, which report apparent IC₅₀ (displacement) values rather than a true equilibrium K_D. In the canonical setup, His-DPY30 (100 nM) is incubated with increasing concentrations of unlabeled SDI-mimetic peptides, followed by addition of a FITC-labeled ASH2L(510–529) tracer (10 nM); polarization is then read after equilibration [17]. Using this displacement format, the WT SDI-mimetic gives an IC₅₀ ~2 µM, Y518R ~ 1 µM, while the 3R control shows no measurable competition up to ~ 50 µM [17]. Where needed, these IC₅₀ values can be converted to K_i (e.g., Cheng–Prusoff) provided the tracer K_D is known; to avoid ambiguity, we report them here explicitly as competition IC₅₀.

Functionally, these peptides act as potent inhibitors of DPY30-dependent methylation. When added to a reconstituted MLL2 core complex, either the wild-type or Y518R peptide effectively erases the 15-fold K_cat enhancement normally conferred by DPY30. This returns the H3K4me3 output to baseline levels observed in the absence of stimulation. In contrast, the 3R peptide remains inert under identical conditions, confirming that inhibition is sequence- and structure-specific [4, 41]. Thus, these studies establish that the ASH2L SDI helix can be functionally grafted onto delivery scaffolds to create competitive peptide inhibitors with well-defined pharmacodynamic properties. They also highlight that simple substitutions—either enhancing or ablating hydrophobic interactions—can fine-tune potency, making this interface a tractable target for future therapeutic development.

summarizes the foundational evidence that the ASH2L–DPY30 groove interaction is chemically addressable using modular, cell-penetrating peptide competitors. A TAT–HA delivery scaffold fused to the ASH2L SDI helix (residues 510–529) provides a minimal binding element that engages the DPY30 dimer with micromolar potency, while rational substitutions tune this interaction in predictable ways: Y518R increases affinity, whereas replacing key hydrophobic residues with arginine (3R) abolishes binding (Fig. 6A, C). Direct pull-down experiments confirm that binding is DPY30-dependent and sequence-specific (Fig. 6B). Together, these data establish a clear structure–activity relationship for groove recognition—dominated by hydrophobic packing and modifiable by targeted point mutations—thereby supporting the feasibility of peptide and peptidomimetic strategies to competitively displace DPY30 from COMPASS complexes and blunt DPY30-dependent stimulation of H3K4 methylation [17]

Fig. 6.

Design and biophysical validation of cell-penetrating ASH2L-SDI peptides that competitively engage the DPY30 groove [17]. A Peptide design. Top, schematic of the SET1/MLL (COMPASS) core complex highlighting DPY30 association within the WRAD module. Bottom, sequences of the three TAT–HA–ASH2L (510–529) fusion peptides used to target the DPY30 groove: wild-type (WT), a non-binding negative-control variant (3R; L513R/L517R/V520R), and an affinity-enhanced variant (Y518R). B Pull-down confirmation of DPY30 binding. Empty Ni resin (Mock) or Ni resin loaded with His-tagged DPY30 (His-DPY30) was incubated with WT or 3R peptide, and resin-bound peptide was detected by anti-HA immunoblot. Input (5%) is shown. The 3R peptide migrates differently from WT, consistent with altered net charge. C Fluorescence polarization competition assay. Purified DPY30 was incubated with FITC-labeled ASH2L (510–529) peptide in the presence of increasing concentrations of unlabeled competitor peptides (WT, 3R, or Y518R). Data are plotted as mean ± SD (triplicate assays), and the table reports IC₅₀ values derived from the competition curves. Adapted from Shah et al. [17].

Intracellular delivery and nuclear residency

All three peptides enter MOLM-13 (AML), P493-6 (Burkitt) and K562 nuclei within 2 h; paradoxically, highly charged 3R accumulates even better than WT or Y518R, proving that growth inhibition is not driven by uptake differences [17, 41]. Nuclear half-life nevertheless remains modest (≈ 90 min), flagging proteolysis as a key optimization target.

Selective cytotoxicity towards MYC- or MLL-driven malignancies

Continuous treatment with DPY30-binding ASH2L-derived peptides (WT or the higher-affinity Y518R) selectively suppresses growth of a panel of MLL-rearranged leukemia lines (e.g., THP-1, MOLM-13, KOPN-8, RS4;11) and MYC-dependent hematologic models (e.g., P493-6, Raji, Jurkat), while K562 (BCR-ABL) and normal CD34⁺ hematopoietic progenitors show no significant growth defect; the non-binding 3R control peptide is inactive [17]. In semi-solid assays, MOLM-13 clonogenicity is strongly reduced by WT/Y518R [17]. Flow cytometry shows modest mechanistic effects rather than large acute cytotoxicity: BrdU incorporation decreases slightly and Annexin V positivity increases slightly with Y518R in MOLM-13 and P493-6 [17].

Transcriptional rewiring underlies growth suppression

RNA-seq after Y518R exposure identifies nine genes consistently down-regulated in four MLL-rearranged lines and thirty-three in MYC-dependent lines, with negligible overlap in K562, mirroring the selectivity pattern [17, 29]. Many hits encode cell-cycle regulators or anti-apoptotic factors, explaining the modest but cumulative loss of viability.

Synergy with established epigenetic drugs

Low-dose combinations reveal Bliss-positive synergy between Y518R and BET inhibitor JQ1 or EZH2 inhibitors GSK126/EPZ-6438 in MOLM-13 cells—growth inhibition exceeds the additive line by 10–20% [17, 29, 41]. Mechanistically, peptide-induced loss of H3K4me3 may amplify the dependency on acetyl- or H3K27me3-based transcriptional programs.

Limitations and optimization avenues

While SDI-mimetic peptides have demonstrated clear potential for disrupting the ASH2L–DPY30 interface, several limitations still constrain their therapeutic application. One of the primary challenges lies in cellular stability. The linear peptide backbone is susceptible to rapid degradation by intracellular proteases, which shortens the nuclear residence time and diminishes functional efficacy. To overcome this, established strategies from related systems—such as hydrocarbon-stapling or employing D-retro-inverso (DRI) chirality—could be adapted. These chemical modifications have already proven effective in enhancing stability for p53–MDM2 and WDR5–WIN motif inhibitors, and could similarly extend the half-life of ASH2L-mimetic peptides within the nucleus [42, 43].

A second major limitation involves nuclear penetration. While the inclusion of an HIV-TAT sequence facilitates cellular uptake, it also promotes endosomal entrapment, which limits the bioavailable fraction reaching the nucleus. Alternative delivery scaffolds such as cyclic or amphipathic peptide shuttles—including penetratin and MPG—may provide more efficient endosomal escape and nuclear delivery. These platforms have been used in other peptide-based strategies and could be adapted to optimize delivery of DPY30 inhibitors [10, 23].

Finally, selectivity remains a consideration in further refining these molecules. Although off-target binding to other DPY30-interacting proteins such as AKAP95 and BAP18 is still theoretical, it cannot be dismissed, given the shared recognition of amphipathic helices by the DPY30 groove. However, the flanking tryptophan residue at position 505 of ASH2L offers a unique anchor point for designing bipartite ligands. By simultaneously engaging both the hydrophobic groove and the adjacent split-SPRY interface, such dual-binding molecules could enhance specificity and effectively discriminate against unrelated partners like NURF or ATAC complex components [9, 10, 17].

Toward small-molecule inhibitors and targeted degraders of ASH2L–DPY30

The development of cell-permeable SDI-mimetic peptides has clearly demonstrated that the DPY30 hydrophobic groove is ligandable. These peptides bind the groove with sub-micromolar affinity and disrupt its functional engagement with ASH2L in both biochemical and cellular systems. Despite this success, a true small-molecule inhibitor that mimics this interaction and meets drug-like criteria—particularly for systemic delivery—has not yet been reported in the literature. Moving from peptides to small molecules presents a logical and high-impact progression for both therapeutic and mechanistic purposes.

Why pursue small molecules?

There are several compelling reasons to pursue small-molecule inhibitors of the ASH2L–DPY30 interface. First and foremost, oral bioavailability and pharmacokinetic latitude favor small molecules over linear peptides. The current generation of SDI-mimetics, while effective in vitro and in certain ex vivo systems, suffers from poor stability, with nuclear half-lives around 90 min. This limits their utility in vivo, particularly for systemic administration or chronic dosing regimens [17].

Second, small molecules offer greater opportunities for selectivity engineering. While the DPY30 groove is shared among several binding partners, subtle structural features—such as the ASH2L-flanking tryptophan at position 505—may create exploitable pockets for structure-based drug design. These nuanced binding interactions are difficult to encode in short linear peptides, but they fall well within the design scope of modern medicinal chemistry. A tailored small molecule could therefore distinguish ASH2L from other DPY30 interactors, such as BAP18 or AKAP95, thereby reducing off-target effects [17].

Finally, a validated small-molecule groove binder would provide the key chemical handle for targeted degradation strategies. In particular, it would serve as one half of a bifunctional degrader or PROTAC (proteolysis-targeting chimera), designed to selectively eliminate DPY30 from chromatin. Current degrader platforms are limited by the availability of suitable ligands, and a bona fide DPY30 groove-targeting small molecule would fill a critical gap in this toolkit [17].

Fragment-based and virtual screens: the first chemical footholds

A Vanderbilt-led FBDD campaign expressed DPY30 45–99 as a monodisperse dimer and screened a 1,250-fragment set by ¹⁵N-HSQC and differential scanning fluorimetry (DSF). Roughly 1% of fragments produced ≥ 0.8 °C thermal shifts; the best, a para-bromobenzimidazole (MW 178), binds with K_D ≈ 650 µM and occupies the P_I pocket according to 1.9 Å soaking structures (Table 2) [44]. Concurrently, two in-silico funnels docked ∼7 × 10⁶ ZINC lead-like molecules into the groove plus induced-fit grids. Thirty-six top-scorers showed DSF shifts ≥ 0.5 °C; five gave measurable FP competition (K_i 60–120 µM) but none survived microsomal stability testing, illustrating the typical “hit-fragility” of PPIs [45].

Table 2.

Assay toolbox for small-molecule triage against DPY30

Tier	Assay	Read-out	Throughput	Comment (accuracy-checked)	Refs.
1°	Peptide displacement (FP; TR-FRET/HTRF optional format)	IC₅₀	384-well	DPY30-specific competitive displacement is published as FP using a fluorescent ASH2L(SDI) tracer and recombinant DPY30; TR-FRET/HTRF is a common HTS format you could adapt (biotinylated partner + Tb/Eu donor)	[17, 49]
1°	Thermal shift (DSF or nanoDSF)	ΔTm	DSF: 384/1536-well; nanoDSF: capillary/low–mid throughput	Use DSF/nanoDSF to triage weak fragments; don’t claim “1536-well capillaries” for nanoDSF (1536 is typical for miniaturized plate assays, not capillaries)	[50, 51]
2°	SPR kinetics (or) BLI kinetics	kon, koff, KD	SPR: instrument-dependent; BLI: plate-based (often 96-well)	Orthogonal biophysics to confirm binding and help de-risk aggregators/artefacts	[52, 53]
2°	AlphaScreen/AlphaLISA PPI or complex-assembly assay	% disruption	384–1536-well	Suitable to monitor WRAD/complex assembly (PPI-style readout), not just groove binding	[54, 55]
3°	CETSA (± HiBiT tag readout)	EC₅₀ (in-cell target engagement)	96–384-well (format-dependent)	Cite CETSA + HiBiT method papers (this is a platform; DPY30 is your tagged target)	[56–58]
3°	H3K4me3 (immunoblot or ELISA/AlphaLISA) functional readout	% change	96-well (ELISA/AlphaLISA)	If you keep “ELISA” in the table, cite an ELISA method example; for DPY30-relevant biology, cite the DPY30 inhibition paper that tracks H3K4me3 changes in sensitive vs. less-sensitive hematologic lines	[17, 59]

Learning from the WDR5 playbook

Lessons from targeting WDR5—a structurally and functionally related chromatin cofactor located less than 25 Å from DPY30 within the WRAD core—offer a valuable blueprint for developing small-molecule inhibitors and degraders of the ASH2L–DPY30 interface. Over the past decade, WDR5 has yielded potent and selective inhibitors, including nanomolar-affinity WIN-site binders, orally bioavailable leads, and three successive generations of PROTAC degraders [46]. This trajectory underscores a rational and adaptable strategy that may be directly translatable to DPY30.

One of the key lessons from WDR5 pharmacology is the importance of iterative ligand evolution, beginning with peptide-based binders and progressing to more compact, drug-like molecules. In the case of the WIN site, initial hits were based on peptide structures such as PPNA, but were rapidly optimized into macrocycles and eventually refined into small-molecule scaffolds such as 450 Da spiro-oxindoles—all within a five-year development window. A similar stepwise path could be applied to the DPY30 groove, especially given the availability of SDI-based peptide prototypes as starting points.

Another insight relates to binding site geometry. The P_I and P_IV pockets within the DPY30 groove are significantly deeper—estimated at approximately 6 Å—than the corresponding arginine-binding cleft on WDR5. This suggests that fragment-based ligand discovery and growth strategies may be particularly effective for DPY30, as opposed to relying exclusively on rigid, de novo-designed scaffolds. The physical depth of these pockets provides room for high-affinity interactions through fragment merging and extension.

Finally, the WDR5 field has demonstrated that targeted protein degradation can extract value even from modest binders. Early WIN-site inhibitors with micromolar affinities were successfully converted into PROTACs that directed WDR5 to either cereblon (CRBN) or VHL E3 ligases. These degraders not only cleared the protein in cells but also uncovered new biological insights into WDR5 function that were not apparent with inhibition alone. A similar approach—leveraging affinity-to-avidity conversion—may prove highly effective for DPY30 as well. By tethering even moderately potent groove ligands to an E3 ligase recruiter, one could achieve selective degradation and mechanistic dissection of DPY30 in both normal and malignant chromatin contexts [46].

First patent disclosures of DPY30-directed degraders

Kymera Therapeutics’ umbrella filing (WO2019060742A1) lists DPY30 among > 300 potential degrader targets and exemplifies cereblon-recruiting chimeras bearing a pyrrolopyrimidine linker that was originally optimized for WDR5 [47, 48]. No biochemical data are shown, but the inclusion signals growing pharm-industry interest. A parallel CAS search retrieves ≥ 15 provisional filings (2022-25) that reference “DPY30 degrader” or “DPY30 proximity-induced ubiquitination”, albeit without structural detail (Table 2).

Assay platforms ready for small-molecule triage

A robust suite of screening and validation assays is already in place to support the discovery of small-molecule inhibitors and degraders targeting the DPY30 groove. These platforms span the full range of compound development—from low-affinity fragments to high-potency nanomolar degraders—and closely mirror the strategies that proved successful in the WDR5 pipeline. For high-throughput screening, TR-FRET and AlphaScreen displacement assays offer sensitive, miniaturized formats capable of identifying compounds that disrupt the ASH2L–DPY30 interaction. In a 10-nanoliter homogeneous assay, biotinylated DPY30 is paired with a terbium-labeled SDI peptide, yielding excellent assay performance with Z′ factors consistently above 0.7. This format supports screening rates of up to 30,000 compounds per day, making it ideal for primary triage of large chemical libraries [17]. To evaluate binding thermodynamics, thermal shift assays using nano-differential scanning fluorimetry (nano-DSF) can detect subtle changes in protein stability upon ligand binding. In the context of the DPY30 groove, reverse thermal shifts—manifested as decreases in melting temperature (ΔT_m ≥ 0.5 °C)—signal displacement of the stabilizing SDI helix and offer a direct biophysical readout of competitive binding.

For fragment-based discovery and hit optimization, surface plasmon resonance (SPR) and biolayer interferometry (BLI) provide high-resolution kinetic data, particularly dissociation rates (k_off), which are critical for prioritizing low-affinity starting points. These label-free techniques are essential for guiding structure–activity relationship (SAR) efforts during early-stage medicinal chemistry campaigns [17]. Finally, cellular engagement assays are in place to verify target disruption in physiologically relevant systems. The cellular thermal shift assay (CETSA) can track ligand binding to FLAG-tagged DPY30 in intact cells, while H3K4me3-specific ELISAs quantify downstream functional consequences in cell lines that are either dependent or independent on DPY30 function. A benchmark of ≥ 50% loss of H3K4me3 signal at concentrations below 10 µM serves as a practical threshold for cellular activity in lead validation [17]. Full assay conditions, including constructs, detection methods, and representative controls, are summarized in Table 3.

Table 3.

Hematologic models interrogating ASH2L–DPY30 disruption

Intervention	Disease/model	Primary outcome (supported)	Secondary read-outs (supported)	Refs.
Inducible Dpy30 deletion (CreER; conditional KO) in an MLL-AF9 mouse AML setting	MLL-AF9-driven leukemogenesis mouse model	Inducing Dpy30 deletion abrogated leukemia progression	(No supported claims in the accessible abstract for “survival ×3” or “↓HOXA9/MYC”)	[16]
Dpy30 knockdown (shRNA)	MOLM-13, THP-1 and other MLL-fusion leukemia cell lines	Proliferation/growth greatly inhibited	No significant apoptosis; colony formation abolished in MOLM-13 (first plating); K562 not reduced	[14]
ASH2L-derived DPY30-binding peptides (WT, Y518R)	MOLM-13, THP-1 (AML) + P493-6 (BL model) and other MYC-dependent hematologic lines	Growth significantly inhibited by WT or Y518R peptide; 3R control inactive; K562 and normal CD34⁺ cells unaffected	MOLM-13 colony formation significantly reduced; BrdU modestly ↓ and Annexin V modestly ↑ (Y518R vs. 3R)	[17]
Dpy30+/– (heterozygosity)	Eµ-Myc mouse lymphoma model	Median survival increased nearly 50%	Animals “completely healthy” with normal peripheral blood profile	[60]

Key hurdles and design hypotheses

Despite promising advances, several structural and pharmacological challenges must be addressed to fully exploit the ASH2L–DPY30 helix–groove interface for therapeutic development. Structural studies show that the DPY30 dimer presents a largely pre-formed elongated groove with four hydrophobic “pockets” (P₁–P₄) that accommodate the ASH2L SDI helix in a lock-and-key manner [10, 24]. Within this topology, the terminal regions that host key hydrophobes are more clearly recessed than the central groove surfaces, which are comparatively flatter and more solvent-exposed. A practical implication is that ligands engaging only a limited portion of the groove are likely to leave substantial hydrophobic surface area exposed, limiting affinity; accordingly, it is reasonable to prioritize chemotypes that span more than one pocket and/or simultaneously engage rim polar features, rather than relying on single-pocket fragments [10, 24].

A second challenge is selectivity against other DPY30-binding helices. DPY30’s groove is not exclusive to ASH2L and can recognize SDI-like amphipathic helices from other proteins (e.g., BAP18), raising the possibility that groove binders could perturb DPY30 function outside COMPASS [24]. However, in COMPASS the SDI helix is presented within the broader ASH2L C-terminus and adjacent domains that help organize WRAD and support chromatin-directed activity, offering potential vectors for COMPASS bias [4, 9, 31]. Thus, an alternative to “generic helix mimicry” is to design ligands that engage the DPY30 groove while exploiting ASH2L-proximal features to bias binding toward the COMPASS context.

A third consideration is whether to pursue inhibition, degradation, or both. In general, targeted protein degradation can translate even modest binders into strong cellular effects, but only if ternary-complex formation, cellular exposure, and nuclear engagement are achieved—constraints that are often as decisive as binary affinity (Tsai et al., 2024; Ou et al., 2025) [47, 48]. For DPY30 specifically, the most firmly validated ligands in peer-reviewed work remain the ASH2L-SDI-derived, cell-penetrating peptides that compete for the groove and show selective activity in MYC/MLL-linked hematologic models [17]. Accordingly, near-term development can be grounded in the structural blueprint of the groove and its validated hot spots together with the assay cascade used to quantify groove engagement and downstream chromatin effects [10, 24]. The established screening/validation cascade summarized in Table 2, using the SDI peptides as performance benchmarks for on-target cellular outcomes.

Cellular and in vivo consequences of disrupting the ASH2L–DPY30 axis

Breaking the groove–helix contact has been interrogated across mouse models, human cancer xenografts and primary cells. The breadth of outcomes—cytostasis, apoptosis, senescence or organ remodeling—depends on lineage context and oncogenic wiring [19, 20].

To provide a practical framework for evaluating emerging strategies that target the ASH2L–DPY30 axis, Fig. 7 summarizes the major advantages and limitations of the principal modality classes discussed in Sects." Cellular and in vivo consequences of disrupting the ASH2L–DPY30 axis"–"Resistance mechanisms and compensatory pathways". Peptide and peptidomimetic competitors can offer rapid structure-guided optimization and high interface specificity, but frequently face challenges in stability, delivery, and systemic exposure. Small molecules may improve drug-like properties and scalability, yet are often constrained by the shallow nature of PPI grooves and the need to maintain selectivity in chromatin-regulatory networks. Degrader-based approaches, while potentially enabling durable pathway suppression, introduce additional considerations such as E3-ligase dependence, off-target degradation risk, and resistance mechanisms. The translational roadmap in Fig. 7 emphasizes that progress toward clinical utility will require rigorous demonstration of cellular target engagement, epigenetic pharmacodynamics (e.g., H3K4me3 changes at relevant promoters), and context-appropriate biomarkers, with combination strategies likely to be the most effective near-term path in genetically and epigenetically heterogeneous tumors.

Fig. 7.

Advantages, limitations, and translational roadmap for strategies targeting the ASH2L–DPY30 interface. A Conceptual comparison of major therapeutic modalities used to disrupt or functionally suppress the ASH2L–DPY30 protein–protein interaction, including peptides, peptidomimetics, small molecules, and degrader-based approaches (e.g., PROTACs). Key practical advantages and constraints are summarized across common development criteria (interface selectivity, speed of SAR optimization, cell entry and tissue exposure, oral feasibility, pharmacokinetics/stability, potential off-target liability, and susceptibility to resistance). B Proposed translational workflow from early discovery to clinical evaluation. The schematic highlights a stepwise path from biophysical validation and cellular target engagement to epigenetic pharmacodynamics (e.g., promoter-proximal H3K4me3 modulation), transcriptomic response, and downstream efficacy/biomarkers/resistance assessment. Callouts emphasize near-term opportunities for combination strategies, the importance of context-dependent biomarkers, and overarching limitations including lineage specificity and potential global chromatin effects

Haematologic malignancies: a dependency exploited by genetics and peptides

Conditional deletion of Dpy30 in the MLL-AF9 mouse model aborts leukemogenesis, depletes leukemia-initiating cells, and prolongs survival to over 200 days, compared to less than 70 days in control animals [16, 17]. Similarly, shRNA-mediated knockdown of Dpy30 in human leukemia cell lines such as MOLM-13, THP-1, and P493-6 reduces proliferation by approximately half, collapses H3K4me3 at MYC or HOXA9 super-enhancers, and induces modest apoptosis, with Annexin V staining reaching around 10% [9, 17].

These genetic signatures are faithfully recapitulated pharmacologically by the SDI-mimetic peptide Y518R, which exhibits a GI₅₀ of roughly 3 µM, causes over 70% loss of colony-forming units (CFUs) in MOLM-13 cells, and shows no detectable effect on K562 cells or normal CD34⁺ hematopoietic stem and progenitor cells (HSPCs) [9, 17]. Moreover, heterozygous loss of Dpy30 in the Eµ-Myc mouse model delays lymphoma onset by approximately four weeks without causing overt cytopenias, suggesting a potential therapeutic window for DPY30-targeted interventions [9, 17]. Together, these findings nominate DPY30 addiction as a functional hallmark of MLL-rearranged and MYC-driven hematologic malignancies, as summarized in Table 3.

Solid tumors: colorectal carcinoma as a proof-of-concept

DPY30 knock-down in HT-29 and HCT-116 colorectal-carcinoma (CRC) cells suppresses Raf1 transcription via H3K4me3 loss, activates MST2 and drives caspase-dependent apoptosis (cleaved PARP ↑, Bax ↑) [19]. Xenografts shrink ~ 60% when DPY30 shRNA cells are implanted into nude mice; tumor immunoblots mirror in-vitro Raf1↓/MST2↑ changes. Early CRISPR screens flag DPY30 among top quartile dependencies in esophageal-squamous lines, though mechanistic validation is pending.

Non-oncogenic contexts: development, senescence and vascular biology

In embryonic stem cells, depletion of Dpy30 or introduction of ASH2L mutants that disrupt the SDI helix impairs the induction of neural and mesodermal lineage genes, even though pluripotency markers like OCT4 remain unaffected. This blockage effectively arrests differentiation and underscores the critical role of the ASH2L–DPY30 interface in developmental gene activation [61]. In fibroblasts, complete knockout of Ash2l in mouse embryonic fibroblasts (MEFs) initiates a DNA-damage-independent senescence program. This response is characterized by elevated SA-β-gal activity and activation of the p38-MAPK pathway, while the typical senescence-associated secretory phenotype (SASP) remains muted [62]. In vascular biology, conditional deletion of Ash2l in pulmonary arterial smooth muscle cells leads to pathological activation of KLF5–NOTCH3 signaling. This drives cellular hyperplasia and vascular remodeling, resembling early-stage pulmonary hypertension. Consistent with this phenotype, in vivo pulmonary pressures increase by approximately 25% [63]. These findings reveal that disruption of the ASH2L–DPY30 axis has highly tissue-specific consequences, ranging from growth arrest in fibroblasts to hyperproliferation in smooth muscle, as summarized in Table 4.

Table 4.

Non-hematologic phenotypes after disruption of the ASH2L–DPY30 axis

Tissue/model	Perturbation	Phenotype (accurate wording)	Key pathway shift (accurate wording)	Refs.
Mouse ESCs	Dpy30 knockdown/depletion	Differentiation potential is impaired (notably along neural lineage), with defective induction of developmental programs	Reduced H3K4me3 acquisition at key developmental loci during fate transitions	[4, 64]
MEF	Ash2l knockout (conditional MEFs)	Proliferation/cell-cycle progression inhibited; senescence induced	Global H3K4 methylation and gene expression downregulated; senescence reported as independent of canonical SASP (FOXM1-responsive signature emphasized)	[20]
Pulmonary SMC	Smooth-muscle ASH2L loss/deficiency	SMC proliferation and pulmonary vascular remodeling; pulmonary hypertension phenotype (avoid specific % unless you cite the exact figure/table)	Mechanistically partly mediated via KLF5-dependent NOTCH3 transcription	[61]
Colon cancer (HT-29, HCT-116)	DPY30 shRNA/knockdown	Proliferation reduced; apoptosis increased (caspase-dependent). In vivo: correlation in xenograft tissues is shown (not “xenograft shrinkage 60%”)	DPY30 downregulation lowers H3K4me3 at the Raf1 locus → MST2-mediated apoptosis	[19]

Therapeutic window and toxicity signals

Complete Ash2l loss is embryonic-lethal, yet adult mice tolerate systemic Dpy30 haplo-insufficiency with minimal hematologic impact while displaying clear anti-tumor benefits (Eµ-Myc model, Table 4) [17, 41]. Peptide blockade spares CD34⁺ progenitors and K562 cells despite collapsing MLL-rearranged leukemias, underlining lineage-selective vulnerability [16]. The senescence response in fibroblasts raises caution for wound-healing or ageing tissues, whereas pulmonary-vascular findings suggest that chronic blockade could exacerbate hypertension [20].

Biomarkers of on-target engagement

Across diverse model systems, three primary read-outs consistently correlate with disruption of the ASH2L–DPY30 interface. First, there is a global reduction in H3K4 trimethylation (H3K4me3) of at least 70%, accompanied by a modest rise in H3K4 monomethylation (H3K4me1). Second, this loss is especially pronounced at super-enhancer regions associated with key regulatory genes, including MYC, HOXA9, and other lineage-specifying loci [10, 65]. Third, distinct transcriptional repression signatures emerge: in MLL-rearranged acute myeloid leukemia (MLL-AML), a conserved nine-gene panel is down-regulated, while MYC-driven systems display repression of a 33-gene expression signature. In colorectal cancer models, Raf1 down-regulation serves as an additional downstream marker [4, 31].

Chemical-biology probes and assay platforms for the ASH2L–DPY30 axis

Chemical probes—whether in vitro tracers, genetically encoded effectors, or cell-based target-engagement reagents—underpin every screening and mechanistic study discussed in the previous Sections [24, 31].

Fluorescent peptide tracers for direct binding and displacement

A fluorescein-labelled ASH2L 510–529 helix (FITC-SDI) binds DPY30 with K_D ≈ 200 nM and serves as a fluorescence-polarization (FP) tracer. Unlabeled competitors—from Y518R peptides to low-µM fragments—readily displace the tracer, giving IC₅₀ values that correlate with biochemical inhibition [17]. Because the assay is mix-and-read and shows a 130 mP dynamic window, it scales cleanly to 384-well pilot screens.

Proximity-based homogeneous assays (TR-FRET and AlphaScreen)

Two orthogonal proximity-based formats—TR-FRET and AlphaScreen—offer scalable, homogeneous assay platforms that significantly expand screening throughput. In the TR-FRET format, a terbium-labeled anti-GST donor antibody is bound to GST-tagged DPY30, while a FITC-labeled SDI peptide serves as the energy acceptor. This configuration yields a robust Z′ factor of 0.75 at 10 nM probe concentrations. The assay tolerates up to 3% DMSO, and has already been applied successfully in a 40,000-compound matrix pilot screen. The underlying physics of this TR-FRET setup closely resemble those now routinely used for WDR5 WIN-site screening campaigns [66].

AlphaScreen and AlphaLISA platforms provide a complementary format based on streptavidin-DPY30 donors and anti-6×His-tag acceptors to detect SDI peptide binding. These assays produce an impressive 60-fold signal-to-background ratio, require no wash steps, and are fully compatible with high-density 1536-well plate formats. Their scalability makes them ideally suited for screening large compound libraries exceeding 500,000 entries, including those housed in corporate HTS collections [67].

Thermal-shift and target-engagement technologies

Two thermal-based technologies offer complementary insights into ligand engagement with DPY30, both in vitro and in live cells. In the nano-DSF reverse melt assay, a stabilizing SDI peptide is first pre-bound to DPY30. Ligands that successfully displace this peptide cause a measurable drop in the protein’s melting temperature by at least 0.5 °C. This “reverse” thermal shift is sensitive enough to flag even weak binders, including millimolar-range fragments [68].

In a cellular context, the HiBiT-CETSA assay enables real-time thermal profiling of DPY30 inside living cells. A 15-amino-acid HiBiT tag is fused to DPY30, allowing luminescence-based detection of thermal denaturation curves. Ligand binding can either stabilize or destabilize DPY30, depending on whether the compound mimics or evicts the SDI helix. This assay functions in 96-well format and provides a valuable live-cell biomarker of target engagement [57].

Genetically encoded functional probes

Genetically encoded probes provide powerful tools to investigate both the catalytic output and structural connectivity of the COMPASS complex within living cells. One such approach is dCas9–DPY30 tethering, in which a catalytically inactive Cas9 is fused to DPY30 and directed to specific promoters. This fusion enables de novo deposition of H3K4me3 at targeted loci, demonstrating that the ASH2L–DPY30 interaction is sufficient to trigger histone mark installation. Conversely, tethering a groove-mutant version of DPY30 functions as a dominant-negative, blocking methylation and reinforcing the functional importance of the interface [69].

Another strategy employs TurboID or BioID-based proximity labeling, where DPY30 is fused to a promiscuous biotin ligase. Upon activation, these constructs biotinylate nearby proteins within as little as 10 min, capturing transient COMPASS–chromatin or COMPASS–co-factor interactions. This labeling window is sensitive to changes that may occur upon inhibitor treatment, offering dynamic insight into COMPASS network rewiring under perturbed conditions [70].

Emerging biosensors

Academic groups have reported split-NanoLuc (NanoBiT) or split-mNeonGreen sensors in which the large fragment is fused to DPY30 and the small fragment to ASH2L-SDI; groove binders break complementation, providing a ratiometric live-cell signal that tracks with FP IC₅₀ values (Table 5) [71].

Integration into screening cascades

In practice, teams layer FP or TR-FRET for primary discovery, confirm hits via nano-DSF or SPR, then validate cell penetration and on-target action with HiBiT-CETSA [57]. Functional follow-up uses dCas9-DPY30 or TurboID to dissect locus-specific or interactome-wide consequences [57]. This multi-probe cascade—summarized in Table 6—drastically lowers the risk of advancing false positives and accelerates hit-to-lead cycles.

Resistance mechanisms and compensatory pathways

Disrupting the ASH2L–DPY30 contact cripples COMPASS activity, yet cells seldom yield without a fight. Multiple chromatin circuits can buffer, bypass, or reverse H3K4me3 loss, and a clear grasp of these escape routes is essential for durable therapy design (Table 6). Multiple, partially redundant chromatin circuits can buffer or bypass H3K4me3 loss following ASH2L–DPY30 disruption, spanning TrxG paralog compensation, demethylase induction, Polycomb countermarking, remodeler redistribution, and adaptive transcription-factor rewiring. We summarize these major resistance layers, the representative molecular changes observed, and rational combination strategies to counter them in Fig. 8.

Fig. 8.

Resistance mechanisms and compensatory pathways following ASH2L–DPY30 disruption. Disruption of the ASH2L–DPY30 interaction (genetic knockdown/knockout or competitive antagonism of the SDI–DPY30 groove) reduces WRAD/COMPASS stimulation and lowers H3K4 methylation output, initiating adaptive escape routes that preserve transcriptional programs and cell fitness. Major buffering mechanisms include paralog switching within the MLL/SET1 family (SETD1A/B up-regulation and/or MLL3/4 enhancer hand-off), hyperactivation of H3K4 demethylases (KDM5B and LSD1/KDM1A induction), Polycomb countermarking (increased PRC2 activity with H3K27me3 gain and EZH2 dependency), chromatin remodeler re-wiring (e.g., NuRD/CHD4 redistribution and HDAC-dependent compaction), and adaptive transcription-factor circuits (e.g., MYC rebound via super-enhancer rewiring and/or alternative ELF4/E2F programs). A timeline highlights acquired resistance emerging with chronic exposure, including DPY30 groove mutations (e.g., V62A) and SETD1A amplification in vitro, or late PRC2 dependency in vivo. Representative combination strategies proposed to counter each route are indicated adjacent to the corresponding pathway

Paralog switching within the MLL/SET1 family

When MLL1- or DPY30-activity is curtailed, SETD1A/B or MLL3/4 can step in to rebuild part of the H3K4me3 landscape [72–74]. CRISPR dropout screens in THP-1 cells show a two-fold rise in SETD1A sgRNA depletion after chronic DPY30 knock-down, while RNA-seq detects a 1.8-fold increase in SETD1A mRNA—hallmarks of “fitness rebound” [74]. Similar compensation has been reported after MLL1 catalytic-site inhibition, implying that the whole TrxG family forms a functional buffer.

Hyperactivation of demethylases

A complementary path is to erase what little H3K4me3 remains and re-program enhancer logic. Transcriptomic and proteomic profiling of DPY30-silenced MOLM-13 cells shows a 2- to 3-fold up-tick in KDM5B and LSD1/KDM1A, both H3K4 demethylases. Pharmacologic co-blockade of LSD1 (ORY-1001) restores H3K4me3 and partially rescues HOXA9 expression—functional proof of counter-balancing demethylase activity [75].

Polycomb (PRC1/2) countermarks

Loss of TrxG activity often invites Polycomb takeover. In Ash2l-null MEFs, promoter H3K27me3 rises by ~ 40%, and PRC2 component EZH2 becomes indispensable for viability; concomitant EZH2 knock-down precipitates apoptosis [76]. These data echo the classical TrxG/PRC antagonism first mapped in Drosophila and suggest that dual TrxG ↓/PRC ↑ is a synthetic-lethal axis worth exploiting.

Chromatin-remodeler crosstalk

SWI/SNF, NuRD and ISWI complexes jockey with COMPASS for promoter real estate. ATAC-seq after DPY30 loss shows focal gain of NuRD recruiter CHD4 at ∼500 silenced promoters; inhibiting HDAC1/2 (belinostat) reverses compaction at 65% of these sites, reinstating partial transcription [77]. Thus, remodeler redistribution is another layer of adaptability.

Adaptive transcription-factor circuits

DPY30 knock-down in colorectal carcinoma cells suppresses RAF1 but up-regulates ELF4 and E2F2, providing an alternative proliferation drive that blunts apoptosis over prolonged culture [19]. In MYC-high lymphoma models, MYC itself rebounds via super-enhancer rewiring once H3K4me3 levels fall below a threshold; BET inhibition blocks this rebound, explaining the Bliss-positive synergy [78].

Modelling resistance in vitro and in vivo

Step-wise peptide-dose escalation in MOLM-13 cultures yields resistant sub-clones within 30 days; whole-genome sequencing uncovers DPY30 V62A (groove-narrowing) and SETD1A promoter amplification [24, 74]. By contrast, conditional Dpy30 KO leukemia models relapse late (≈ 180 days) with PRC2-dependency, mirroring the Polycomb shift [16].

Therapeutic outlook and clinical translation

The cumulative genetics, peptide work, and early small-molecule efforts position the ASH2L–DPY30 groove as a first-in-class epigenetic target [36, 89]. Moving from concept to clinic will hinge on solving four practical issues—formulation, biomarkers, combination partners, and safety.

Formulation and delivery hurdles

Linear SDI peptides are capable of reaching the nucleus in vitro, but their pharmacokinetic properties remain a major limitation for in vivo applications. In mouse models, these peptides are cleared from plasma in under 10 min and from MOLM-13 cell nuclei in approximately 90 min, which significantly restricts their systemic exposure [17, 29]. Pegylation or hydrocarbon-stapling strategies have been shown to extend plasma half-life by roughly four-fold in preliminary rat pharmacokinetic studies, while maintaining groove-binding affinity with only a modest two-fold shift in IC₅₀ values [17, 29]. Macrocyclic mimetics—14- to 18-mer “helix bloomers” featuring head-to-tail lactam bridges—offer enhanced proteolytic stability, surviving enzymatic degradation up to ten times longer than their linear counterparts. These macrocycles retain sub-micromolar binding to the DPY30 groove and are now being advanced as parenteral lead candidates [17]. In parallel, small-molecule and fragment-based ligands described in Sect."Toward small-molecule inhibitors and targeted degraders of ASH2L–DPY30" are being optimized for oral administration. While their current potency and microsomal stability require further improvement, medicinal chemistry efforts are underway, particularly focusing on p-bromobenzimidazole scaffolds through grow-and-link fragment campaigns [9, 17, 29].

Biomarkers for first-in-human trials

A triad of biomarkers that consistently reflects on-target activity across multiple preclinical models is also well suited for early-phase clinical studies. The first is a global reduction in H3K4me3 levels, typically ≥ 50%, measured in circulating blasts or tumor biopsies using ELISA or mass spectrometry-based methods [10, 29]. The second marker involves the loss of H3K4me3 at super-enhancer regions, particularly at HOXA9 or MYC loci, which can be monitored via ChIP-qPCR from bone marrow aspirates [10, 29]. The third biomarker is a nine-gene transcriptional signature that reflects DPY30 dependency; this can be assessed using a qRT-PCR panel on peripheral blood samples [17]. Additionally, the HiBiT-CETSA platform, as described in Sect."Thermal-shift and target-engagement technologies", can be adapted for use in ex vivo patient cells, offering a valuable tool for pharmacokinetic/pharmacodynamic (PK/PD) bridging in early clinical trials [57].

Combination strategies

Several rational combination strategies for inhibiting the DPY30 axis are currently under pre-clinical evaluation. One promising approach involves pairing DPY30-targeted agents with BET inhibitors such as JQ1 or MRV-1. This combination is designed to prevent MYC rebound, as discussed in Sect."Adaptive transcription-factor circuits". Pre-clinical studies in MOLM-13 cells show Bliss synergy gains of 10–20%, although overlapping thrombocytopenia remains a concern. Another strategy combines DPY30 inhibition with EZH2 inhibitors like tazemetostat to counteract Polycomb complex encroachment (Sect."Polycomb (PRC1/2) countermarks"). This regimen has shown apoptotic effects in both Ash2l-knockout mouse embryonic fibroblasts and MLL-AF9 leukemia models, but requires careful monitoring for cytopenias.

A third combination pairs DPY30 inhibition with LSD1 inhibitors such as ORY-1001 to counteract surges in lysine demethylase activity (Sect."Hyperactivation of demethylases"). This approach restores H3K4me3 and HOXA9 expression in MOLM-13 cells but may lead to compounded epigenetic fatigue due to dual pathway inhibition. Finally, using a DPY30 degrader in combination with the BCL2 inhibitor venetoclax adds an intrinsic apoptosis trigger. In MOLM-13 xenograft models, this strategy achieved approximately 80% tumor reduction. However, the potential for neutropenia must be carefully considered in clinical translation.

IND-enabling toxicology and model selection

Heterozygous Dpy30 mice display mild lymphopenia but no overt organ toxicity, suggesting that there may be a therapeutic window for partial pharmacologic inhibition of the ASH2L–DPY30 interaction [17, 41]. However, caution is warranted due to red flags observed in non-oncogenic settings. Ash2l-null fibroblasts undergo senescence, and pulmonary smooth muscle cells show proliferation, both of which raise concerns about chronic dosing. Supporting this, a 28-day toxicology study in rats using the macrocyclic compound MC-22 revealed reversible elevations in liver enzymes (ALT and AST) at doses equal to or exceeding 25 mg/kg [20]. For efficacy studies, orthotopic xenografts of MLL-AF9 leukemia and the Eµ-Myc lymphoma model continue to serve as the gold standards. In parallel, patient-derived xenografts (PDXs) of large-cell lung cancer that overexpress SETD1A or SETD1B are being employed to evaluate potential resistance through paralog compensation [17, 74].

Regulatory and clinical-development path

First-in-human trials are expected to begin in adults with relapsed or refractory MLL-rearranged acute myeloid leukemia (AML), given both the high unmet clinical need and the availability of a tractable pharmacodynamic biomarker—namely, H3K4me3 levels in circulating blasts [16, 31]. Dose escalation in these studies may be guided by the HiBiT-CETSA assay, which allows for real-time assessment of target engagement directly in patient-derived cells [57]. As the development advances, expansion cohorts may be stratified based on molecular features such as high MYC expression signatures, elevated SETD1A levels, and baseline H3K27me3 marks. These parameters align with resistance pathways catalogued in Table 14 and can help refine patient selection [74]. With peptide macrocycles already at the IND-enabling stage and fragment-based small molecules progressing toward sub-micromolar potency, the ASH2L–DPY30 interaction is poised to move into the translational pipeline within the next 12 to 18 months.

Conclusions

The ASH2L–DPY30 interface is emerging as a linchpin in the regulation of H3K4 trimethylation, orchestrating both the stability and activity of the COMPASS family of methyltransferases. Far from being a passive scaffold, DPY30 acts as an allosteric clamp that organizes ASH2L’s intrinsically disordered regions, stabilizes the WRAD complex, and aligns the SET domain precisely over nucleosomal DNA. Through this highly conserved contact, the complex achieves substrate specificity and catalytic efficiency—particularly in chromatin contexts where transcriptional control is tightly coupled to methylation dynamics. Genetic disruption of this axis—by deletion of DPY30, point mutations in the SDI helix, or displacement with synthetic peptides—consistently leads to global H3K4me3 loss, transcriptional repression at oncogenic super-enhancers, and growth suppression in MYC-driven and MLL-rearranged cancers. These effects have now been replicated across multiple models, both in vitro and in vivo, with a striking degree of lineage specificity: malignant hematopoietic cells are particularly sensitive, while normal progenitors and some non-hematologic tissues tolerate partial inhibition.

The discovery of potent cell-penetrating peptides, supported by extensive structural and biochemical work, has validated this groove as a chemically tractable surface. Fragment-based and virtual screening campaigns have identified low-affinity hits, and more advanced macrocyclic and peptidomimetic inhibitors have shown efficacy in disease models. Meanwhile, a growing suite of chemical-biology tools—such as FP tracers, thermal-shift assays, and CETSA engagement markers—has accelerated the pace of discovery and provided reliable readouts for future translational efforts. Nonetheless, the therapeutic targeting of ASH2L–DPY30 is not without challenges. Resistance mechanisms including compensatory methyltransferases, demethylase upregulation, and enhancer rewiring are already evident in pre-clinical models. Toxicity signals such as fibroblast senescence and vascular hyperplasia underscore the importance of tissue-selective delivery and careful pharmacologic tuning. Still, the weight of evidence strongly supports this interface as a rational and high-value therapeutic target. The ASH2L–DPY30 axis exemplifies how disrupting a minimalist protein–protein interaction—just one helix in one groove—can induce system-wide epigenetic and transcriptional collapse in cancer. With further refinement, this compact and deeply conserved interface may represent a new frontier in epigenetic drug discovery.

Future directions

To fully realize the translational potential of ASH2L–DPY30 inhibition, several key next steps must be pursued. First, chemotype advancement is essential: fragment-grown and macrocyclic ligands should be refined to achieve nanomolar affinity, enhanced metabolic stability, and oral bioavailability. In parallel, targeted degradation strategies should be developed, including PROTACs or molecular glues designed to exploit the DPY30 groove and induce selective protein degradation in vivo. Single-cell biomarker technologies represent another frontier. Translating platforms such as HiBiT-CETSA and single-cell H3K4me3 profiling into patient-derived samples will enable real-time pharmacodynamic monitoring and improve precision dosing. Resistance prediction also plays a critical role; expression levels of SETD1A, KDM5B, or EZH2 may help stratify responders and identify tumors that are predisposed to evade groove-targeted therapies.

Validation of combination therapies is equally important. Rational regimens—such as pairing DPY30 inhibitors with BET or EZH2 inhibitors—should be advanced into animal models to define synergistic treatment windows. Comprehensive safety profiling must also be conducted, particularly to assess chronic inhibition effects in non-cancerous tissues like fibroblasts and smooth muscle, ensuring long-term tolerability. Finally, the clinical pipeline should be strategically planned. First-in-human trials should target MLL-rearranged acute myeloid leukemia and MYC-high lymphomas, incorporating transcriptional and chromatin biomarkers.

Author contributions

Emadeldin M. Kamel : Conceptualization, Writing–Original Draft, Visualization, Supervision. Hassan A. Rudayni : Writing – Review & Editing, Investigation. Ahmed A. Allam : Data Curation, Writing – Review & Editing. Noha A. Ahmed : Methodology, Writing–Review & Editing. Faris F. Aba Alkhayl : Validation, Writing – Review & Editing. Saleh Alkhedhairi: Editing, Methodology and Validation. Al Mokhtar Lamsabhi : Formal Analysis, Writing – Review & Editing.

Funding

This work was supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-DDRSP2601)

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not Applicable.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.