Hypothesis

In cells harboring heterozygous SF3B1, U2AF1, SRSF2, or ZRSR2 mutations, the transition from tolerable splicing noise to deterministic oncogenic dysfunction occurs when the burden of mis‑spliced intron‑retention events that generate R‑loops exceeds a cell‑type‑specific threshold. This threshold can be quantified as the product of intron‑retention Shannon entropy (H_IR) and γH2AX foci density (a proxy for replication‑stress‑induced DNA damage). Crossing this point sustains ATR‑Chk1 signaling, exhausts checkpoint recovery, and creates a dependence on PARP‑mediated repair, rendering the cells synthetically lethal to low‑dose splicing modulators combined with ATR inhibitors.

Mechanistic rationale

1. Intron retention as the primary source of R‑loops

   • Mutant spliceosomes increase retention of introns that contain cryptic polyadenylation signals or GC‑rich sequences.
   • These retained introns are poorly exported, persist in chromatin, and anneal to the template strand, forming stable R‑loops.
   • Unlike exon skipping, intron retention directly contributes to transcriptional‑associated DNA lesions because the RNA polymerase II complex encounters a RNA‑DNA hybrid while still engaged with the gene body.

2. Entropy‑R-loop product as a predictor of dysfunction

   • Shannon entropy of intron‑retention events (H_IR) captures the diversity and unpredictability of mis‑splicing across the transcriptome.
   • γH2AX foci density reflects the cumulative DNA‑damage burden resulting from R‑loop‑mediated replication fork collapse.
   • We hypothesize that when H_IR × γH2AX exceeds a critical value (θ), the cell can no longer compensate through RNase H2‑dependent R‑loop resolution or ATR‑mediated fork protection, leading to chronic checkpoint activation.

3. From noise to deterministic phenotypes

   • Below θ, increased transcriptomic heterogeneity produces stochastic protein isoforms that are largely buffered by quality‑control mechanisms (nonsense‑mediated decay, proteasome degradation).
   • Above θ, specific retained introns in key DNA‑repair or replication genes (e.g., BRCA1, FANCD2, MCM5) produce stable, aberrant proteins that actively impair repair, creating a deterministic shift toward genomic instability and proliferation.
   • This explains why only a handful of recurrent mis‑splicing events (3‑5 per cancer type) are sufficient to recapitulate malignant phenotypes despite widespread noise.

4. Therapeutic implication: synthetic lethality with ATR inhibition

   • Cells operating near θ rely heavily on ATR‑Chk1 to manage replication stress; ATR inhibition pushes them into mitotic catastrophe.
   • Low‑dose splicing modulators (e.g., H3B‑8800) amplify intron‑retention entropy, moving wild‑type cells closer to θ but not beyond it, whereas mutant cells are pushed over the edge.
   • Combining a sub‑toxic dose of a splicing modulator with an ATR inhibitor should selectively kill SF3B1/U2AF1/SRSF2/ZRSR2‑mutant cells while sparing normal counterparts.

Experimental plan

• Step 1: Measure H_IR – Perform deep RNA‑seq on isogenic cell lines expressing each spliceosome mutant and corresponding controls. Quantify intron‑retention events per gene, compute Shannon entropy across all retained introns.
• Step 2: Quantify γH2AX – Immunofluorescence microscopy to count γH2AX foci per nucleus under basal conditions and after low‑dose H3B-8800 treatment.
• Step 3: Define θ – Plot H_IR × γH2AX for each condition; identify the value that correlates with onset of persistent ATR‑Chk1 phosphorylation (p‑ATR, p‑Chk1) and loss of colony‑forming ability.
• Step 4: Test synthetic lethality – Treat mutant and wild‑type cells with a matrix of H3B-8800 concentrations and ATR inhibitor (e.g., ceralasertib). Assess viability (CellTiter‑Glo) and apoptosis (caspase‑3/7). Expect synergistic cell death only when the product exceeds θ in mutant cells.
• Step 5: Validate mechanistic link – Knock down RNase H2 or overexpress RNase H1 to modulate R‑loop resolution; observe shifts in θ. Use S9.6 immunoprecipitation followed by sequencing to map R‑loop hotspots and confirm enrichment over retained introns in high‑entropy samples.

If the hypothesis holds, we will have a quantitative, measurable threshold that translates splicing noise into deterministic oncogenic dysfunction, providing a biomarker for patient stratification and a rational combination strategy to exploit splicing‑factor mutant cancers.