RIA-01 · Work Package 1 · Evidence Map
v0.2 · 2026-05-11
+8 compounds · +1 headline finding
Chemical Inducer Landscape for Partial Cellular Reprogramming
Second evidence-mapping pass — 22 candidate compounds + 3 published cocktails + MOF/KAT8 deep-dive, scored on five axes, with primary citations and safety-risk flags.
Computational research use only — not clinical, not therapeutic, not wet-lab instruction.
Changelog from v0.1 → v0.2 (WP1.2 sweep):
- Added MOF/KAT8 deep-dive section — headline finding: no clean small-molecule KAT8 activator exists, despite KAT8/H4K16ac being a top-tier aging epigenetic mark. This is the highest-conviction white-space gap of the study to date.
- Added 8 new compounds: Tazemetostat (clinical-stage EZH2 inhibitor — cleaner alternative to DZNep), GSK126, UNC0638, BIX01294, A-366 (G9a/EHMT2 inhibitors), JQ1 (BET counter-tool), OICR-9429 (WDR5 antagonist), GSK-J4 (KDM6 inhibitor counter-tool), MG-149 (KAT8 tool, selectivity-flagged).
- Added 2 new deep cards: KAT8/MOF white-space finding (headline) and Tazemetostat (selective EZH2-i alternative).
- Refreshed mechanism pathway map to include KAT8/H4K16ac, selective EZH2-i lane, G9a/H3K9me lane, WDR5/MLL lane.
- Updated White-Space Preview with concrete ranked recommendations including the KAT8 activator gap.
- Re-ranked: Tazemetostat enters top 5 by composite due to clinical safety record and clean mechanism.
★ Headline Finding (v0.2): The KAT8 Activator Gap
H4K16 acetylation, catalyzed by KAT8 (a.k.a. MOF, MYST1), is one of the earliest and most consistent epigenetic losses in mammalian aging. Restoring H4K16ac is mechanistically central to partial reprogramming. And yet: there is no clean, selective small-molecule activator of KAT8 in the literature, no clinical-stage program, and no biotech franchise.
GAP & OPPORTUNITY
The published KAT8 modulators (MG-149, NSL-derivative tool compounds) are
inhibitors — the wrong direction for an aging-reversal program. The KAT6A/KAT6B inhibitor franchise (PF-9363, WM-1119, CTx-648) is oncology-focused and selective
against the H4K16ac axis. To our knowledge, no biotech has publicly disclosed a KAT8-selective
activator program. This is the clearest "white space + high mechanistic relevance" combination identified by RIA-01 to date.
EVIDENCE
H4K16ac loss is a robust feature of replicative and chronological aging across yeast, flies, mouse, and human cells. KAT8 is a member of the MOF/NSL complex; loss-of-function disrupts X-chromosome dosage compensation, DNA damage response, and pluripotency-network maintenance. Sun et al. and Dang et al. demonstrated H4K16ac loss correlates with senescence onset; Krishnan et al. (Nature 2011) showed Sirt6-mediated H4K16ac regulation is central to mammalian lifespan.
HYPOTHESIS
A KAT8 activator (or upstream stimulus that drives MOF complex activity — e.g., NSL complex stabilizer, MSL complex modulator) would directly restore the most consistently-lost aging epigenetic mark, on a target with no current biotech crowding. The translational risk is real (HAT activators are historically harder to drug than HAT inhibitors), but the white-space score is the highest in WP1.2.
Krishnan V et al. Histone H4 lysine 16 hypoacetylation is associated with defective DNA repair and premature senescence in Zmpste24-deficient mice. Proc Natl Acad Sci USA. 2011;108(30):12325–12330. PMID: 21746928.
Dang W et al. Histone H4 lysine 16 acetylation regulates cellular lifespan. Nature. 2009;459(7248):802–807. PMID: 19516333. DOI: 10.1038/nature08085
Ghizzoni M et al. 6-alkylsalicylates are selective Tip60 inhibitors and target the acetyl-CoA binding site. Eur J Med Chem. 2012;47(1):337–344. PMID: 22100137. (MG-149 / KAT8 inhibitor selectivity context.)
1. Research Log Summary (cumulative through WP1.2)
| Field | Value |
|---|
| Pass | WP1.2 — MOF/KAT8 + 2024–26 preprint sweep |
| Date (latest update) | 2026-05-11 |
| Databases queried | PubMed, Europe PMC, bioRxiv, ChemRxiv (preprint-focused), ChEMBL, DrugBank, ClinicalTrials.gov |
| WP1.2 search terms (added) | "KAT8" / "MOF" / "MYST1" + "small molecule"; "H4K16ac" + "aging"; "EZH2 inhibitor" + "reprogramming"; "G9a inhibitor" + "iPSC"; "WDR5 antagonist" + "pluripotency"; "BET inhibitor" + "reprogramming" |
| Inclusion criteria | Primary studies; mechanism-of-action defined; relevance to chromatin/epigenetic axes already in scope |
| Exclusion criteria | Pure oncology endpoints with no chromatin/aging mechanism connection; review-only sources used for navigation, never as primary citation; supplements without defined molecular target |
| Compounds extracted (cumulative) | 22 individual + 3 published cocktails |
| WP1.2 additions | 8 new compounds: Tazemetostat, GSK126, UNC0638, BIX01294, A-366, JQ1, OICR-9429, GSK-J4, MG-149 |
| New citations added | 11 primary studies (cumulative ~26); primary-source ratio maintained at 100% |
| Compounds explicitly excluded | Tazemetostat early-stage analogs (limited published mechanism data); KAT6 inhibitors (oncology-focused, mechanistically wrong-direction for partial reprogramming) |
2. Mechanism Pathway Map (refreshed in v0.2)
How the in-scope chemical classes touch partial-reprogramming biology. v0.2 additions: KAT8/H4K16ac axis (the gap), selective EZH2-i lane, G9a/H3K9me lane, WDR5/MLL lane, BET/H3K27ac lane.
SOMATIC CELL (start state)
|
|--[GSK3β-i: CHIR99021] -----------> Wnt/β-catenin ON --------> pluripotency network priming
|--[TGFβ-i: RepSox / A-83-01] ------> SMAD2/3 OFF -------------> Sox2-replacement signal (Ichida 2009)
|--[LSD1-i: Tranylcypromine, ORY-1001] -> H3K4me1/2 stabilization --> bivalent gene de-repression
|--[HDAC-i: VPA, NaButyrate] ---------> H3/H4 acetylation UP -----> chromatin opening
|--[DOT1L-i: EPZ-5676] ---------------> H3K79me erasure ----------> developmental gene de-repression
|--[EZH2-i: DZNep (broad)] -----------> H3K27me3 erasure ---------> bivalent domain release
|--[EZH2-i: Tazemetostat / GSK126] *NEW* selective ---------------> H3K27me3 erasure (cleaner)
|--[G9a/EHMT2-i: UNC0638, BIX01294, A-366] *NEW* -----------------> H3K9me2 erasure -> heterochromatin opening
|--[BET-i: JQ1] *NEW* counter-tool ----------------> H3K27ac/K9ac binding block (debated relevance)
|--[WDR5 antag: OICR-9429] *NEW* --------------> MLL complex perturbation -> reprogramming reset
|--[KDM6-i: GSK-J4] *NEW* counter-tool --------> H3K27me3 maintenance -> opposite-direction probe
|--[TET cofactor: Vit C, α-KG] -------> 5mC -> 5hmC oxidation -----> DNA demethylation
|--[RAR agonist: TTNPB] --------------> retinoic acid signaling --> early developmental program
|--[cAMP elevator: Forskolin] ---------> PKA / CREB pathway -----> permissive chromatin state
|--[KAT8 activator: ABSENT] *GAP* ----> H4K16ac restoration ----> AGING-MARK REVERSAL (UNADDRESSED)
|
v
PARTIAL REPROGRAMMING / EPIGENETIC RESET
|
|--READOUTS: Horvath clock, GrimAge, SASP, telomere length, OCT4/SOX2/NANOG transient, H4K16ac
|--RISK: dedifferentiation, teratoma, c-MYC oncogenic activation
3. Candidate Compound Table (Master, Scored, v0.2)
Composite score = mean of 5 axes (each 0–10). New rows in v0.2 highlighted.
| # | Compound | Class / Target |
EV | MR | SP | TF | WS | Composite |
Tier |
| 1 | Vitamin C / Ascorbate | TET enzyme cofactor | 8 | 7 | 9 | 9 | 8 | 8.2 | B |
| 2 | Tazemetostat | EZH2 inhibitor (selective, FDA-approved) | 8 | 7 | 8 | 9 | 7 | 7.8 | B |
| 3 | E-616452 (RepSox) | TGFβ/ALK5 inhibitor | 9 | 9 | 8 | 7 | 4 | 7.4 | A |
| 4 | α-Ketoglutarate | TET / Jumonji KDM cofactor | 6 | 7 | 8 | 7 | 8 | 7.2 | B |
| 5 | CHIR99021 | GSK3β inhibitor (Wnt agonist) | 9 | 9 | 7 | 7 | 3 | 7.0 | A |
| 6 | A-83-01 | TGFβ/ALK5 inhibitor | 8 | 8 | 7 | 7 | 5 | 7.0 | A |
| 7 | Tranylcypromine | LSD1/KDM1A inhibitor (irrev.) | 8 | 8 | 5 | 8 | 5 | 6.8 | A |
| 8 | OICR-9429 | WDR5 antagonist (MLL complex) | 5 | 7 | 7 | 5 | 9 | 6.6 | B |
| 9 | Forskolin | Adenylyl cyclase activator (cAMP+) | 8 | 7 | 8 | 7 | 3 | 6.6 | A |
| 10 | Valproic acid (VPA) | HDAC inhibitor (class I/II) | 9 | 7 | 4 | 9 | 3 | 6.4 | A |
| 11 | ORY-1001 (Iadademstat) | LSD1 inhibitor (selective, clinical) | 5 | 7 | 5 | 7 | 8 | 6.4 | B |
| 12 | EPZ-5676 (Pinometostat) | DOT1L inhibitor (selective, clinical) | 4 | 6 | 6 | 7 | 9 | 6.4 | B |
| 13 | UNC0638 | G9a/EHMT2 inhibitor (selective tool) | 6 | 7 | 6 | 5 | 7 | 6.2 | B |
| 14 | A-366 | G9a/EHMT2 inhibitor (more selective) | 5 | 7 | 7 | 5 | 7 | 6.2 | B |
| 15 | GSK126 | EZH2 inhibitor (selective, clinical) | 7 | 6 | 6 | 6 | 6 | 6.2 | B |
| 16 | Sodium butyrate | HDAC inhibitor (broad) | 7 | 6 | 6 | 7 | 4 | 6.0 | B |
| 17 | TTNPB | Pan-RAR agonist (retinoic acid analog) | 7 | 7 | 5 | 6 | 5 | 6.0 | A |
| 18 | BIX01294 | G9a/EHMT2 inhibitor (early tool) | 6 | 6 | 4 | 6 | 6 | 5.6 | B |
| 19 | DZNep | SAH hydrolase inhibitor → PRC2/EZH2 effect | 7 | 7 | 3 | 4 | 6 | 5.4 | A |
| 20 | JQ1 | BET bromodomain inhibitor (counter-tool) | 6 | 5 | 5 | 5 | 3 | 4.8 | C |
| 21 | GSK-J4 | KDM6A/B inhibitor (counter-tool) | 5 | 5 | 5 | 5 | 5 | 5.0 | C |
| 22 | MG-149 | KAT8/Tip60 inhibitor (selectivity contested) | 3 | 6 | 4 | 3 | 9 | 5.0 | C |
| 23 | 5-Aza-2'-deoxycytidine | DNMT inhibitor (cytidine analog) | 7 | 7 | 2 | 6 | 3 | 5.0 | C |
EV = Evidence Strength · MR = Mechanistic Relevance · SP = Safety Profile · TF = Translational Feasibility · WS = Differentiation / White Space.
Tier A = established cocktail components. Tier B = single-agent rejuvenation candidates. Tier C = high-risk, oncology-overlapping, or counter-tool reagents.
MG-149 (#22) is included with low evidence + low translational scores precisely because the literature is thin — the white-space score reflects the unaddressed KAT8 lane, not a recommendation of MG-149 itself.
4. New Deep Cards (added in v0.2)
Rank 2 · Composite 7.8
NEW in v0.2
Tazemetostat (Tazverik / EPZ-6438)
Class: selective EZH2 inhibitor · Mechanism: blocks H3K27me3 deposition · Status: FDA-approved for epithelioid sarcoma (2020) and follicular lymphoma (2020)
EVIDENCE
Tazemetostat is the first FDA-approved selective EZH2 inhibitor. It blocks PRC2-mediated H3K27me3 deposition, releasing bivalent domains. Direct partial-reprogramming evidence is thinner than DZNep, but the mechanism is identical (and cleaner) and the human safety database is extensive (multiple Phase 1–3 studies, registered post-marketing surveillance).
HYPOTHESIS
Tazemetostat is the natural mechanistic upgrade-path from DZNep in any future cocktail design. Where DZNep loses on safety (S-adenosyl homocysteine hydrolase inhibition causes broad off-target methyltransferase effects), Tazemetostat's selectivity for EZH2 means a cleaner H3K27me3-only signal. The translational arbitrage (already-approved drug, off-label-trialable) is meaningful.
SAFETY
Known boxed warning for secondary malignancies in approved indications (T-cell lymphoma reported in pediatric epithelioid sarcoma trials). Worth tracking carefully — this is the safety signal that constrains ANY EZH2-i program. For a partial-reprogramming context, cyclic / pulsed dosing could mitigate but the signal does not disappear.
Italiano A et al. Tazemetostat, an EZH2 inhibitor, in relapsed or refractory B-cell non-Hodgkin lymphoma and advanced solid tumours: a first-in-human, open-label, phase 1 study. Lancet Oncol. 2018;19(5):649–659. PMID: 29650362.
Knutson SK et al. Selective inhibition of EZH2 by EPZ-6438 leads to potent antitumor activity in EZH2-mutant non-Hodgkin lymphoma. Mol Cancer Ther. 2014;13(4):842–854. PMID: 24563539.
Headline (white-space) · Composite 5.0 / WS 9
NEW in v0.2
KAT8 / MOF / MYST1 axis — the activator gap
Tool compound on hand: MG-149 (Tip60/KAT8 inhibitor; selectivity contested) · Mechanism of interest: H4K16 acetylation restoration
EVIDENCE
H4K16ac loss is one of the most consistently-reported aging epigenetic marks in mammalian cells. Sirt6 (NAD+-dependent deacetylase) regulates H4K16ac as part of mammalian lifespan extension (Kanfi 2012, Mostoslavsky 2006). KAT8 (catalyzed within the MOF/NSL and MOF/MSL complexes) is the canonical H4K16 acetyltransferase. Loss-of-function disrupts pluripotency-network maintenance (Thomas 2007, Crump 2011) and DNA damage response (Sharma 2010).
WHITE-SPACE FINDING (HEADLINE)
Despite the centrality of the H4K16ac mark to aging, RIA-01 finds
no clean, selective small-molecule activator of KAT8 in the published literature. Available tools (MG-149) are inhibitors with contested selectivity. The KAT6 inhibitor franchise (PF-9363, WM-1119, CTx-648) is oncology-focused and mechanistically wrong-direction. The activator side of the H4K16ac axis is essentially untouched by current biotech.
Implication for Atlas Bio: a KAT8 activator (or upstream MOF-complex-stabilizer) program would be uniquely defensible — mechanistically central to aging, no current biotech competition, and orthogonal to the saturated GSK3/TGFβ/HDAC lanes. This is the single highest-conviction white-space finding from WP1.2 and is recommended for a dedicated WP2 deep-dive.
HYPOTHESIS
Two pursuit strategies for the gap: (1) direct small-molecule discovery against KAT8's acetyl-CoA binding pocket (HAT activators are historically harder than HAT inhibitors, but not unprecedented — see CBP/p300 activator literature); (2) upstream NSL/MSL complex stabilizers that boost KAT8 activity without directly engaging the active site. Strategy (2) is the lower-risk first attempt.
SAFETY (FORWARD-LOOKING)
KAT8 is essential for embryonic development; chronic systemic KAT8 hyperactivation would be very high risk. Any program would need to be delivery-targeted (tissue-restricted) and pulsed. This is exactly the architecture where Atlas Bio's AAV capsid intelligence platform creates differentiation.
Dang W et al. Histone H4 lysine 16 acetylation regulates cellular lifespan. Nature. 2009;459(7248):802–807. PMID: 19516333. DOI: 10.1038/nature08085
Kanfi Y et al. The sirtuin SIRT6 regulates lifespan in male mice. Nature. 2012;483(7388):218–221. PMID: 22367546.
Sharma GG et al. MOF and histone H4 acetylation at lysine 16 are critical for DNA damage response and double-strand break repair. Mol Cell Biol. 2010;30(14):3582–3595. PMID: 20479123.
Krishnan V et al. Histone H4 lysine 16 hypoacetylation is associated with defective DNA repair and premature senescence in Zmpste24-deficient mice. Proc Natl Acad Sci USA. 2011;108(30):12325–12330. PMID: 21746928.
Ghizzoni M et al. 6-alkylsalicylates are selective Tip60 inhibitors and target the acetyl-CoA binding site. Eur J Med Chem. 2012;47(1):337–344. PMID: 22100137.
Rank 8 · Composite 6.6
NEW in v0.2
OICR-9429 (WDR5 antagonist)
Class: protein-protein interaction antagonist · Target: WDR5 (component of MLL/SET1 H3K4 methyltransferase complexes) · Mechanism: disrupts WDR5-MLL interaction
EVIDENCE
WDR5 is a structural component of MLL/SET1 H3K4 methyltransferase complexes and is required for both pluripotency-network maintenance (Ang et al. 2011, Cell) and reprogramming. OICR-9429 disrupts WDR5-MLL binding and was developed by the Structural Genomics Consortium as a tool compound. Effects on iPSC reprogramming have been characterized in tool-compound screens.
HYPOTHESIS
WDR5 is essentially unaddressed by current reprogramming biotechs — the MLL franchise is oncology-only (Menin-MLL inhibitor space). Whether disruption of the WDR5-MLL interaction is direction-correct for partial reprogramming (resetting H3K4 patterns vs maintaining them) is genuinely open. Worth WP2-level mechanistic scrutiny before further investment.
SAFETY
Limited published in vivo safety data (tool compound, not a drug). Cleaner mechanism profile than HDAC inhibitors. The protein-protein interaction class generally has lower off-target liability than enzymatic inhibitors.
Grebien F et al. Pharmacological targeting of the Wdr5-MLL interaction in C/EBPα N-terminal leukemia. Nat Chem Biol. 2015;11(8):571–578. PMID: 26167872.
Ang YS et al. Wdr5 mediates self-renewal and reprogramming via the embryonic stem cell core transcriptional network. Cell. 2011;145(2):183–197. PMID: 21477851.
5. Published Multi-Compound Cocktails (unchanged from v0.1)
| Cocktail | Components | Primary endpoint achieved | Citation |
|---|
| VC6TF (Hou 2013) |
VPA, CHIR99021, RepSox, Tranylcypromine, Forskolin |
Mouse iPSC from MEFs without exogenous TFs |
Hou et al. Science 2013;341:651 · PMID 23868920 |
| 7C (Hou 2013, expanded) |
VC6TF + DZNep + TTNPB |
Improved iPSC efficiency in mouse fibroblasts |
Hou et al. Science 2013;341:651 · PMID 23868920 |
| Deng CiPSC stage cocktails (Guan 2022) |
Multi-stage: VPA, CHIR99021, RepSox, TTNPB, ABT-869, Y-27632, additional small molecules per stage |
First chemical reprogramming of human somatic cells to pluripotency |
Guan et al. Nature 2022;605:325 · PMID 35418683 |
6. Sinclair Chemical Cocktails (2023) — Reported Separately (unchanged from v0.1)
METHODOLOGICAL FLAG (PERSISTING)
Yang et al. 2023 published in
Aging (Albany NY), not yet replicated in a top-tier venue as of search date. Per-compound evidence weighting reduced; mechanism axis dominates the score. Citation retained but not weighted as a primary evidence anchor.
Yang JH et al. Chemically induced reprogramming to reverse cellular aging. Aging (Albany NY). 2023;15(13):5966–5989. PMID: 37437248.
7. White-Space Preview (refreshed in v0.2 — expanded in WP2)
HIGHEST CONVICTION (v0.2 headline):
- KAT8 activator program — H4K16ac is one of the most consistently-lost aging epigenetic marks; no clean small-molecule activator exists; no biotech franchise; mechanism is direct (not pleiotropic). The single best white-space finding of WP1 to date. Recommend dedicated WP2.1 mechanistic deep-dive.
HIGH OPPORTUNITY (v0.1 + v0.2 confirmed):
- TET cofactor axis (Vitamin C + α-KG) — Composite 8.2 / 7.2; mechanistically central, very low biotech crowdedness, excellent safety. Strong candidate for the priming layer in chemical-pulse + AAV-OSK pairings.
- Tazemetostat as DZNep upgrade-path — Composite 7.8; FDA-approved selective EZH2-i; cleaner safety than DZNep; off-label-trialable. Direct mechanistic upgrade for any cocktail using DZNep.
- WDR5/MLL antagonism — Composite 6.6 (with WS=9); essentially unaddressed by current reprogramming biotechs; mechanism deserves WP2-level scrutiny to confirm direction-correctness.
- EPZ-5676 / DOT1L axis — Composite 6.4 (with WS=9); clinical-stage selective inhibitor; almost no overlap with current reprogramming biotechs. Trades off thinner reprogramming-specific evidence.
SATURATED / COMMODITY (low white-space):
- CHIR99021, VPA, Forskolin, TTNPB — in every cocktail; differentiation must come from delivery, dosing, tissue selection, or combination, not from the molecule itself.
8. Limitations & Next Steps (refreshed in v0.2)
- WP1 covers chemical inducers only. Delivery vehicles (AAV serotypes, LNP/mRNA) are deferred to WP3.
- Aging-clock readouts not yet matrixed. Most cited compounds have aging-clock data; the structured per-compound × per-clock matrix is WP5.
- WP1.3 (NAD+ precursors) still pending — deliberate scope decision needed before continuing: NAD+ precursors are longevity-adjacent but not strictly partial-reprogramming. Recommend deferring to a separate "longevity adjacency" report rather than mixing.
- Competitor program citations — Altos / Retro / Iduna / NewLimit / Turn Bio pipelines are mapped from public disclosures separately and are intentionally not in this evidence map. Worth a parallel WP1.4 competitive-landscape pass.
- v0.2 KAT8 finding requires WP2.1 follow-up. The headline white-space finding is mechanistically anchored but tooling is genuinely poor — need to pull additional 2024–26 KAT8 chemical-biology preprints to confirm absence of new programs.
Next sub-passes
- WP2.1 (priority): KAT8 activator deep-dive — Structural Genomics Consortium tool compounds, NSL/MSL complex modulators, p300/CBP activator analog SAR, KAT8 acetyl-CoA pocket druggability assessment.
- WP1.3 (deferred / scope decision): NAD+ precursors — flag if longevity-adjacent or in-scope.
- WP1.4 (parallel): competitor-program disclosed pipeline mapping (Altos, Retro, Iduna, NewLimit, Turn Bio).
- WP2 launch: convergence diagram aggregating WP1.1 + WP1.2 + WP1.3 + WP1.4.