A Transcriptional Entropy Threshold Determines ATR Inhibitor Sensitivity in Spliceosome-Mutant Cancers

Dr. Hugo Mora (MD)· Mar 21, 2026

Mechanism: Spliceosome mutations increase intron retention and R-loop formation, activating ATR-Chk1 signaling when a retained-intron threshold is surpassed. Readout: Readout: ATR inhibitors reduce cell viability by over 70% in cells exceeding this threshold, correlating with decreased p-Chk1 and γ-H2AX signals.

Hypothesis

In hematopoietic malignancies bearing heterozygous SF3B1, U2AF1, SRSF2, or ZRSR2 mutations, oncogenic potency is not dictated by a single misspliced transcript but by the quantitative burden of co‑transcriptional splicing errors that generate R‑loops. When the ratio of retained introns to total nascent transcripts exceeds a definable threshold, ATR‑Chk1 signaling becomes constitutively activated, creating a synthetic‑lethal dependence on ATR activity. Below this threshold, cells tolerate splicing noise; above it, genome instability drives clonal expansion and determines therapeutic response to ATR inhibitors.

Mechanistic rationale

Mutant spliceosomes alter the kinetic coupling between RNA polymerase II elongation and exon definition. Elevated Pol II elongation rates, already observed in cancer increased transcription rates, reduce the time window for spliceosome assembly, increasing the probability of intron retention, especially at GC‑rich promoters prone to R‑loop formation. Retained introns expose nascent RNA that can hybridize with the template DNA strand, stabilizing R‑loops. These structures impede replication fork progression, activate ATR‑Chk1, and promote a mutational phenotype that favors clonal expansion.

Thus, spliceosome mutations act as “entropy amplifiers”: they increase the stochastic production of aberrant transcripts, and the functional impact emerges only when the cumulative error load surpasses a biophysical limit that overwhelms RNA‑processing and DNA‑repair capacities.

Testable predictions

Quantitative threshold – In isogenic cell lines expressing heterozygous SF3B1, U2AF1, SRSF2, or ZRSR2 mutants, nascent RNA‑seq (PRO‑seq) coupled with intron‑retention profiling will reveal a linear relationship between Pol II elongation rate and retained‑intron fraction. Cells will exhibit a sharp increase in γ‑H2AX and p‑Chk1 signal once the retained‑intron/total‑nascent ratio exceeds ~0.12 (empirically derived from dose‑response curves). Below this ratio, ATR inhibition will have minimal effect; above it, ATR inhibitors will reduce viability >70 % (IC₅₀ shift >5‑fold) SF3B1 modulator synergy.
R‑loop dependence – Overexpression of RNase H1 in mutant cells will lower R‑loop levels (measured by S9.6 immunofluorescence) without altering the specific misspliced isoforms of MAP3K7 or other known targets. This manipulation will raise the retained‑intron threshold required for ATR inhibitor sensitivity, demonstrating that R‑loops, not individual transcripts, mediate the phenotype.
Clinical correlation – Patient‑derived MDS/AML samples with matched DNA‑seq and RNA‑seq will be stratified by the retained‑intron/total‑nascent ratio. Those above the predicted cutoff will show higher expression of ATR‑Chk1 pathway genes and greater ex vivo sensitivity to ATR inhibitors (e.g., ceralasertib), independent of the identity of the spliceosome mutation splicing factor dysregulation as driver.

Falsifiability

If experiments show that ATR inhibitor response correlates tightly with the abundance of a single misspliced transcript (e.g., MAP3K7 downregulation) and remains unchanged when global intron‑retention levels are experimentally decoupled from R‑loop formation (via RNase H1 titration), the hypothesis would be refuted. Likewise, if the retained‑intron/total‑nascent ratio fails to predict ATR inhibitor sensitivity across multiple genetic backgrounds, the entropy‑threshold model would be invalid.

By linking transcriptional kinetics, spliceosome fidelity, and genome‑stability signaling through a quantifiable metric, this framework shifts the debate from “which transcript matters” to “how much splicing noise is tolerable” and offers a concrete biomarker for ATR‑targeted therapy in spliceosome‑mutant cancers spliceosome mutation landscape.

Hypothesis

Mechanistic rationale

Testable predictions

Quantitative threshold – In isogenic cell lines expressing heterozygous SF3B1, U2AF1, SRSF2, or ZRSR2 mutants, nascent RNA‑seq (PRO‑seq) coupled with intron‑retention profiling will reveal a linear relationship between Pol II elongation rate and retained‑intron fraction. Cells will exhibit a sharp increase in γ‑H2AX and p‑Chk1 signal once the retained‑intron/total‑nascent ratio exceeds ~0.12 (empirically derived from dose‑response curves). Below this ratio, ATR inhibition will have minimal effect; above it, ATR inhibitors will reduce viability >70 % (IC₅₀ shift >5‑fold) SF3B1 modulator synergy.

R‑loop dependence – Overexpression of RNase H1 in mutant cells will lower R‑loop levels (measured by S9.6 immunofluorescence) without altering the specific misspliced isoforms of MAP3K7 or other known targets. This manipulation will raise the retained‑intron threshold required for ATR inhibitor sensitivity, demonstrating that R‑loops, not individual transcripts, mediate the phenotype.

Clinical correlation – Patient‑derived MDS/AML samples with matched DNA‑seq and RNA‑seq will be stratified by the retained‑intron/total‑nascent ratio. Those above the predicted cutoff will show higher expression of ATR‑Chk1 pathway genes and greater ex vivo sensitivity to ATR inhibitors (e.g., ceralasertib), independent of the identity of the spliceosome mutation splicing factor dysregulation as driver.

Falsifiability