1,200+ U2AF1 mutations reveal distinct clinicobiological features of the S34/Q157 co-mutation subgroup Free

The U2AF1 spliceosome gene recurrently acquires S34 (predominantly S34F/Y) and Q157 (primarily Q157P/R) mutations in myeloid malignancies. These hotspot mutations occupy distinct zinc finger domains, differentially modulating splice site selection and potentially driving tumorigenesis through divergent pathways. Clinically, S34 mutations are more prevalent than Q157, with each conferring unique phenotypic and prognostic profiles. Typically, S34 and Q157 mutations are mutually exclusive in myeloid neoplasms, and S34/Q157 co-mutations are exceptionally rare, leaving their molecular and clinical implications poorly characterized.Here, we describe the largest cohort (n=41) of U2AF1 S34/Q157 co-mutated myeloid neoplasms among >1,200 U2AF1-mutant cases, revealing distinct molecular and clinical profiles through systematic comparisons with single-mutant groups.

We evaluated 1,263 newly diagnosed myeloid neoplasm patients (2020-2025) with U2AF1 S34/Q157 mutations detected by next-generation sequencing (NGS). The cohort comprised 888 MDS, 223 AML, 86 MPN, and 66 MDS/MPN cases. Patients were stratified into three groups: S34+Q157 co-mutation (co-mut, n=41), S34-only (n=1,022), Q157-only (n=200). Comparative analyses of molecular and clinical characteristics were performed across groups.

The co-mut cohort comprised 33 S34F+Q157R, 7 S34Y+Q157R, and 1 S34F+Q157P cases. The S34-only group predominantly harbored S34F/Y (n=1019), while the Q157-only group consisted mostly of Q157P/R (n=194). Intriguingly, Q157R was less common than Q157P in Q157-only cases (58 vs 136, p<0.001). However, in the co-mut group, Q157R was highly enriched (97.6%). This suggesets that Q157R and Q157P may have distinct pathogenic mechanisms.

The disease composition across the co-mut group, S34-only group, and Q157-only group was as follows: MDS (61% vs 72.4% vs 58%), AML (29.3% vs 17.6% vs 13.5%), MPN (3.4% vs 4.6% vs 18%), and MDS/MPN (3.4% vs 4.8% vs 7.5%). AML prevalence was significantly elevated in co-mut vs single-mut groups (p<0.05), indicating S34/Q157 synergy in AML pathogenesis. Additionally, MPN predominated in Q157-only cases (p<0.05 vs other groups).

Based on variant allele frequency (VAF) patterns, the co-mutation cohort could be stratified into two subgroups: a balanced VAF subgroup (difference <10%; 58.5% of cases) and a skewed VAF subgroup (difference ≥10%; 41.5% of cases). Notably, nearly all cases in the skewed subgroup exhibited a higher VAF for S34 than for Q157, with only one exception showing the opposite pattern (Q157 > S34). The balanced subgroup demonstrated a significantly higher incidence of AML (37.5% vs. 17.6%, p< 0.05), reinforcing the S34/Q157 synergistic hypothesis.

Further analysis demonstrated that co-mut patients exhibited intermediate median age (53 years) between S34-only (49 years) and Q157-only (63 years) groups, and the co-mut group was significantly lower than the Q157-only group (p<0.001). Furthermore, S34Y carriers were youngest (median 45 vs 51 years for S34F, p<0.001), while Q157P patients were oldest (64 vs 60 years for Q157R). Male predominance (69.0-85.4%) was most prominent in co-mut cases (85.4%), with females being consistently older except in Q157R subgroup (M:61.5 vs F:57.5 years).

Hematologic parameters revealed co-mut patients had significantly higher Hb (89 vs 74-76 g/L, p<0.05) and RBC (2.96×10^12/L vs 2.45-2.43×10^12/L, p<0.01) compared to single-mutant groups, while maintaining intermediate WBC counts between S34-only and Q157-only groups. Platelet levels showed no intergroup differences.

Co-mut cases showed frequent concomitant mutations in ASXL1, SETBP1, ETV6, RUNX1,and NRAS. Compared to S34-only, SETBP1, ETV6, IDH2, and CEBPA were significantly enriched (all p<0.05). Similarly, ETV6 enrichment persisted versus Q157-only cases, whereas JAK2 mutations predominated in Q157-only patients (p<0.05), reflecting distinct mutational landscapes. These co-mutation results may be confounded by intergroup disease distribution differences.

Our NGS analysis identified U2AF1 S34/Q157 co-mutated myeloid neoplasms with distinct molecular patterns (Q157R predominance) and clinical phenotypes (AML enrichment). By characterizing this co-mut cohort, we reveal the molecular complexity underlying the pathogenesis of myeloid neoplasms and highlight that different U2AF1 mutation subtypes likely exert unique molecular mechanisms in disease development.