A quantitative splicing entropy score predicts synthetic lethality to spliceosome modulators in SF3B1-mutant tumors

Dr. Hugo Mora (MD)· Mar 26, 2026

Mechanism: High Splicing Entropy Score (SES) in SF3B1-mutant tumors indicates excessive mis-splicing, leading to nuclear stress and R-loop accumulation. Readout: Readout: Spliceosome inhibitors cause lethal transcriptomic collapse in high SES tumors, leading to significant tumor regression and decreased tumor viability.

Hypothesis

A computable splicing entropy score (SES) that integrates intron retention, cryptic splice‑site usage, and exon skipping across the transcriptome can distinguish adaptive splicing diversity from deleterious transcriptomic entropy in SF3B1‑mutant cancers. Tumors with SES above a defined threshold will exhibit synthetic lethality when treated with spliceosome‑targeting agents such as SF3B1 inhibitors or CLK2/3 inhibitors, whereas tumors below the threshold will be resistant.

Mechanistic basis

Heterozygous spliceosome mutations reprogram splice site recognition, generating a mixture of functional and aberrant isoforms. Low‑level isoform variation can support tumor adaptability (e.g., drug‑resistant NT5C2 variants) while excessive mis‑splicing overwhelms RNA‑processing capacity, leading to nuclear stress, R‑loop accumulation, and activation of the p53‑dependent checkpoint. The SES captures the burden of GC‑rich intron retention and cryptic 3′ splice‑site activation that directly correlates with splicing‑factor condensate gelation under ATP‑limiting conditions. When SES exceeds the cellular buffering capacity, further pharmacological inhibition of spliceosome kinetics pushes the system past a catastrophic entropy point, causing lethal transcriptomic collapse.

Testable predictions

In a panel of SF3B1‑mutant cell lines, SES calculated from RNA‑seq will inversely correlate with IC50 values of H3B‑8800 and E7107.
CRISPR‑mediated reduction of intron‑retention–prone genes (e.g., BRD4, MAPK1) will lower SES and confer resistance to spliceosome inhibitors without altering mutant allele frequency.
Elevating cellular ATP via overexpression of mitochondrial uncoupling protein 2 will decrease SES‑dependent sensitivity, linking energy status to the entropy threshold.
Patient‑derived xenografts with pretreatment SES above the 75th percentile will show tumor regression upon CLK2/3 inhibitor treatment, while those below will exhibit stable disease.

Experimental design

Generate isogenic SF3B1‑WT and SF3B1‑K700E lines in hematopoietic progenitors. Perform deep RNA‑seq, compute SES as (weighted sum of retained intron reads + cryptic 3′ splice‑site junction reads) / total mapped reads, weighting GC‑rich introns higher.
Treat cells with dose‑response curves of H3B‑8800, E7107, and a CLK2/3 inhibitor. Fit IC50 and compare to SES using Spearman correlation.
Use lentiviral shRNA to knock down selected retained‑intron genes; re‑measure SES and drug sensitivity.
Modulate ATP levels with oligomycin or ATP‑boosting compounds; reassess SES and drug response.
Implant SF3B1‑mutant patient‑derived xenografts in NSG mice, baseline RNA‑seq to assign SES, randomize to CLK2/3 inhibitor or vehicle, monitor tumor volume and perform endpoint RNA‑seq to verify SES changes. Statistical significance set at p<0.05, with power analysis targeting 80% detection of a 30% IC50 shift.

References: Spliceosome mutations reprogram splice site recognition, Pan-cancer analyses identify >15,000 cancer-specific splice variants, Emerging therapies include PRMT5, CLK2/3 inhibitors, and ASOs, Synthetic introns target SF3B1-mutant cells

Hypothesis

Mechanistic basis

Testable predictions

In a panel of SF3B1‑mutant cell lines, SES calculated from RNA‑seq will inversely correlate with IC50 values of H3B‑8800 and E7107.

CRISPR‑mediated reduction of intron‑retention–prone genes (e.g., BRD4, MAPK1) will lower SES and confer resistance to spliceosome inhibitors without altering mutant allele frequency.

Elevating cellular ATP via overexpression of mitochondrial uncoupling protein 2 will decrease SES‑dependent sensitivity, linking energy status to the entropy threshold.

Patient‑derived xenografts with pretreatment SES above the 75th percentile will show tumor regression upon CLK2/3 inhibitor treatment, while those below will exhibit stable disease.

Experimental design

Generate isogenic SF3B1‑WT and SF3B1‑K700E lines in hematopoietic progenitors. Perform deep RNA‑seq, compute SES as (weighted sum of retained intron reads + cryptic 3′ splice‑site junction reads) / total mapped reads, weighting GC‑rich introns higher.

Treat cells with dose‑response curves of H3B‑8800, E7107, and a CLK2/3 inhibitor. Fit IC50 and compare to SES using Spearman correlation.

Use lentiviral shRNA to knock down selected retained‑intron genes; re‑measure SES and drug sensitivity.

Modulate ATP levels with oligomycin or ATP‑boosting compounds; reassess SES and drug response.

Implant SF3B1‑mutant patient‑derived xenografts in NSG mice, baseline RNA‑seq to assign SES, randomize to CLK2/3 inhibitor or vehicle, monitor tumor volume and perform endpoint RNA‑seq to verify SES changes. Statistical significance set at p<0.05, with power analysis targeting 80% detection of a 30% IC50 shift.