Phase-separated condensates of oncogenic transcription factors (MYC, YAP, TAZ) in cancer cells are selectively vulnerable to condensate-disrupting compounds, offering a cancer-selective therapeutic window absent in normal cells where these TFs are not condensate-forming
Adversarial Debate Score
52% survival rate under critique
Expert panel critique
Independent views, each critiquing the hypothesis on its own — the score rewards genuine disagreement and discounts consensus.
Supporting Research Papers
- DNA Partitioning Modulates Liquid-to-Solid Transitions and the Internal Microstructure of FUS Condensates
Protein liquid-liquid phase separation has recently been recognized as an essential process involved in cellular functions, including transcription, translation, and DNA damage repair. However, a furt...
- Mapping high resolution, multidimensional phase diagrams of physiological protein condensates
Biomolecular condensates are membraneless compartments, crucial for organising and regulating diverse cellular processes. Current approaches to study condensate biology either use simplified recombina...
- Targeting DNA Methylation: New Paradigms and the Advent of Gene-Selective Tools
DNA methylation can function as a toxic alkylation reaction exploited by chemotherapeutic agents to induce cancer cell death. However, finely tuned DNA methylation plays a fundamental role in cellular...
- Nanoscale domains govern local diffusion and aging within FUS condensates
Biomolecular condensates regulate cellular physiology by sequestering and processing RNAs and proteins, yet how these processes are locally tuned within condensates remains unclear. Moreover, in neuro...
- ECLIPSE: A Composable Pipeline for Predicting ecDNA Formation, Evolution, and Therapeutic Vulnerabilities in Cancer
Extrachromosomal DNA (ecDNA) represents one of the most pressing challenges in cancer biology: circular DNA structures that amplify oncogenes, evade targeted therapies, and drive tumor evolution in ~3...
Formal Verification
Z3 checks whether the hypothesis is internally consistent, not whether it is empirically true.
This discovery has a Claude-generated validation package with a full experimental design.
Precise Hypothesis
In human cancer cell lines with defined dependency on MYC, YAP, or TAZ (e.g., MYC-amplified neuroblastoma/Burkitt lymphoma; YAP/TAZ-dependent mesothelioma or liver cancer lines), these transcription factors form liquid-liquid phase-separated (LLPS) nuclear condensates detectable by (a) puncta formation of endogenously tagged protein under confocal/super-resolution microscopy, (b) FRAP recovery kinetics consistent with liquid-like mobility (t1/2 < 60s, mobile fraction >50%), and (c) sensitivity to 1,6-hexanediol (>50% puncta dissolution at 5-10% w/v within 5 min). A defined set of condensate-disrupting small molecules (e.g., aliphatic diol analogs, IDR-targeting compounds, or established hits like JQ1-adjacent BET disruptors, or bespoke MYC/YAP condensate perturbants) will reduce condensate puncta number/size by ≥40% at concentrations that reduce cancer cell viability by ≥50% (IC50), while the same compounds at equivalent doses produce ≤10% viability reduction and no significant condensate disruption in matched normal/non-transformed cells (e.g., primary fibroblasts, hTERT-immortalized epithelial cells) where these TFs are expressed at low/diffuse (non-condensed) levels. The therapeutic window (normal cell IC50 / cancer cell IC50) must be ≥5-fold and must correlate (Pearson r ≥0.6, p<0.05) with baseline condensate propensity across a panel of ≥8 cell lines.
- Condensate puncta are observed at similar frequency/size in normal cells (e.g., primary fibroblasts under serum stimulation) as in cancer cells, indicating condensation is not cancer-selective.
- FRAP or hexanediol-sensitivity data show MYC/YAP/TAZ puncta in cancer cells are not liquid-like (e.g., immobile fraction >70%, no hexanediol sensitivity), suggesting aggregates rather than LLPS condensates.
- Condensate-disrupting compounds reduce cancer cell viability but with therapeutic window <2-fold relative to normal cells, or normal cell condensates (if present) are equally disrupted.
- Viability reduction does not correlate with condensate disruption (e.g., compound dissolves condensates without killing cells, or kills cells without measurable condensate change) — indicating the mechanism of action is condensate-independent.
- Genetic disruption of phase separation (IDR truncation/mutation abolishing LLPS in vitro) fails to reduce oncogenic transcriptional output or cell viability in cancer lines, decoupling condensate formation from oncogenic function.
Spine & Adversarial Read
- highThe entire premise that 1,6-hexanediol sensitivity or FRAP dynamics reliably distinguish 'true' phase-separated condensates from other nuclear puncta (aggregates, hubs, or simply chromatin-bound clusters) is contested in the field; McSwiggen et al. and others have shown hexanediol has broad, non-specific effects on transcription and chromatin, meaning the core diagnostic assay may not validly establish LLPS at all.Protocol requires orthogonal confirmation via FRAP kinetics, in vitro droplet reconstitution with purified IDR domains, and concentration-dependent condensate formation (not just presence/absence), but this remains only partially resolved since no single-cell in vivo assay is universally accepted as definitive LLPS proof; the EVP should be read as testing 'condensate-like behavior with therapeutic relevance' rather than rigorously proven biophysical phase separation.
- highWhy is dose-response viability/condensate correlation across a modest panel (n=4-8 cell lines) the chosen methodology rather than a genome-wide CRISPR screen for condensate-dependent essentiality, or direct biophysical measurement (in-cell NMR, single-molecule tracking) of TF diffusion coefficients? The chosen imaging+viability approach is cheaper and faster but is also the most confounded by indirect effects and cannot mechanistically prove condensate-disruption is the cause of death rather than a correlate.The methodology is justified as a Tier 1 minimum-viable, cost-constrained screen intended to establish correlational evidence and go/no-go signal before investing in more expensive orthogonal biophysical or genetic causality studies (IDR-deletion rescue is included as the causal arm); however the EVP does not yet include a genome-wide genetic interaction screen or single-molecule imaging, which would be needed for definitive mechanistic proof and should be flagged as a Tier 2/3 extension, not covered by this budget.
- mediumThe claimed cancer-selectivity may simply reflect known baseline oncogene-dependency differences (cancer cells are MYC/YAP-addicted per DepMap regardless of condensate status) rather than condensate-specific vulnerability; any compound that lowers MYC/YAP activity by any mechanism (not just condensate dissolution) would show a similar selective window, confounding condensate-disruption with generic pathway inhibition.Partially addressed by requiring EC50(condensate)-IC50(viability) correlation and IDR-deletion mutant rescue experiments to isolate the condensate-specific contribution, but a fully rigorous control would require a compound that inhibits TF transcriptional activity WITHOUT disrupting condensates (or vice versa) as a mechanistic decoupling control, which is not yet specified in this protocol and represents a gap.
Experimental Protocol
Minimum viable test (Tier 1, ~4-6 weeks): Use 3 cancer lines with endogenous fluorescent knock-in (or validated antibody IF) for one TF (e.g., MYC in P493-6 B-lymphoma, HCT116, and Burkitt lymphoma Raji) plus 2 normal controls (BJ fibroblasts, RPE1-hTERT). Confirm condensates by confocal microscopy + FRAP + 1,6-hexanediol assay. Treat with 2-3 known condensate disruptors (1,6-hexanediol as tool compound, plus one candidate small molecule, e.g., a BRD4 bromodomain inhibitor known to affect MYC enhancer condensates, or ANKRD1/verteporfin analogs for YAP/TAZ) across dose range (8-point, 3-fold dilution). Measure: condensate count/area (automated image analysis, n≥50 cells/condition, 3 replicates), viability (CellTiter-Glo, 72h), and correlate.
- Cell lines: MYC-driven (P493-6, HCT116, Raji, KP-4), YAP/TAZ-driven (mesothelioma NCI-H226, MDA-MB-231, liver cancer Huh7), normal controls (BJ fibroblast, RPE1-hTERT, primary hepatocytes)
- CRISPR knock-in endogenous fluorescent tagging constructs (mNeonGreen/GFP-MYC, -YAP1, -WWTR1) or validated ChIP-grade antibodies for IF
- Compound library: 1,6-hexanediol, ANKRD1 pathway/verteporfin (YAP), BET inhibitors (JQ1, dBET6), candidate LLPS-disrupting small molecules from published screens (if available)
- Imaging: spinning-disk confocal or super-resolution (Airyscan/SIM) microscope with live-cell chamber for FRAP
- RNA-seq / ChIP-seq datasets (public, e.g., ENCODE, CCLE) for baseline TF expression levels across cell lines to select panel
- CCLE/DepMap dependency data to confirm MYC/YAP/TAZ dependency scores per line
- Image analysis pipeline (CellProfiler or ImageJ/Fiji with custom puncta-detection macro)
- Condensates confirmed liquid-like in ≥3/4 cancer lines (FRAP mobile fraction >50%, hexanediol sensitivity >50% dissolution).
- Condensate-disrupting compound achieves therapeutic window ≥5-fold (normal/cancer IC50 ratio) in ≥2 independent compounds across ≥2 TF types (MYC and YAP/TAZ).
- Pearson correlation r≥0.6 (p<0.05) between condensate EC50 and viability IC50 across the cell panel.
- IDR-deletion mutant shows ≥50% reduction in target gene transcription and ≥40% reduction in oncogenic phenotype rescue vs. WT, confirming condensate-dependence of function.
- Therapeutic window <2-fold, or normal cells show equivalent condensate disruption/toxicity.
- No significant correlation (r<0.3 or p>0.05) between condensate disruption and viability loss.
- Puncta are hexanediol-insensitive and FRAP-immobile (indicating solid aggregates, not LLPS) in >50% of cancer lines tested.
- IDR-deletion mutants retain full oncogenic transcriptional activity, decoupling phase separation from function.
ROI Projection
Implementation Sketch
# Tier 1 pipeline (per cell line, per compound) for cell_line in [cancer_panel + normal_panel]: reporter = generate_or_validate_TF_tag(cell_line, TF) # CRISPR KI or IF baseline_condensates = image_and_quantify(reporter, method="confocal") liquidity_score = FRAP_assay(reporter) + hexanediol_assay(reporter) for compound in [hexanediol, JQ1, verteporfin_analog, candidate_X]: for dose in dose_range(8, log_scale): condensate_metric = quantify_puncta(reporter, compound, dose, t=1-4h) viability = celltiter_glo(cell_line, compound, dose, t=72h) EC50_condensate = fit_dose_response(condensate_metric) IC50_viability = fit_dose_response(viability) store(cell_line, compound, EC50_condensate, IC50_viability, liquidity_score) correlation = pearson(EC50_condensate_all, IC50_viability_all) therapeutic_window = IC50_viability[normal] / IC50_viability[cancer] # Orthogonal genetic validation mutant_TF = delete_IDR(TF) rescue_line = express_in_null_background(mutant_TF) transcription_output = RNAseq(rescue_line, target_genes) phenotype_rescue = viability_assay(rescue_line) vs WT_rescue
- Day 10: If baseline imaging fails to show reproducible, quantifiable puncta above diffuse nuclear background in ≥2 cancer lines, abort/redesign imaging protocol before compound testing.
- Day 20: If FRAP/hexanediol data indicate condensates are not liquid-like (immobile, hexanediol-insensitive) in majority of lines, pause and reassess whether "condensate" framing is valid before proceeding to compound screening.
- Day 35: If preliminary dose-response shows no separation between cancer and normal cell viability curves (window <2-fold) for the first 2 compounds tested, do not proceed to full 8-compound/8-dose matrix — reassess compound selection or hypothesis validity.
NAMED_EXPERTS: []
CLOSEST_EXISTING_WORK: []
NOVELTY_NARROWING_REQUIRED: true
SPINE_STATEMENT: This hypothesis tests whether pharmacological disruption of MYC/YAP/TAZ phase-separated condensates selectively kills cancer cells over normal cells because only cancer cells harbor condensate-forming (rather than diffuse) forms of these transcription factors.