Abstract
Objectives
SRSF2 mutations are known to be associated with poor outcomes in myelodysplastic neoplasm, but studies on their prognostic impact on acute myeloid leukemia (AML) remain limited. In this retrospective study, we analyzed clinical and pathologic characteristics of patients with AML and correlated the outcomes with SRSF2 mutations.
Methods
We characterized the morphologic, immunophenotypic, molecular, and clinical findings in AML with mutated SRSF2 and compared them with SRSF2 wild-type (WT) myeloid neoplasms (MNs).
Results
Using next-generation sequencing, we identified 134 patients with MNs and SRSF2 mutations (85 with AML and 49 with MNs) in addition to 342 SRSF2-WT AMLs. Fifty-two (62%) patients with altered SRSF2 demonstrated a variable degree of morphologic dysplasia. The most frequent immunophenotypic aberrancies in SRSF2-mutant AML included diminished CD33 expression and overexpression of CD7, CD56, or CD123, similar to WT AML. More IDH1/2 (P = .015) and NPM1 (P = .002) mutations were seen in SRSF2-mutant AML than in SRSF2-mutant non-AML. Further, more IDH1/2, ASXL1, RUNX1, and STAG2 mutations were observed in SRSF2-mutant AML than in SRSF2-WT AML (P < .0001 to P = .001). Finally, patients with SRSF2-mutant AML showed a significantly worse overall survival (OS) than patients with SRSF2-WT AML (P < .0001), but this worse OS appeared to be rescued by allogeneic stem cell transplant (allo-SCT).
Conclusions
Acute myeloid leukemia with altered SRSF2 shows a variable degree of morphologic dysplasia without uniform immunophenotypic aberrancies. SRSF2 mutations appear to be independent poor prognostic factors, but allo-SCT has improved the clinical outcomes in patients with SRSF2-mutant AML.
Keywords: somatic mutation, flow cytometry, NGS, SRSF2, AML
KEY POINTS.
The prognostic impact of SRSF2 mutations is evaluated in patients with acute myeloid leukemia (AML).
SRSF2 mutations appear to be independent poor prognostic factors for overall survival in patients with AML.
Poor clinical outcome appears to be rescued by allogeneic stem cell transplant in patients with altered SRSF2 AML.
INTRODUCTION
The serine and arginine rich splicing factor 2 (SRSF2) gene is a member of the splicing factor gene family, along with splicing factor 3B subunit 1 (SF3B1); U2 small nuclear RNA auxiliary factor 1 (U2AF1); and zinc finger, CCCH-type, RNA binding motif and serine/arginine rich 2 (ZRSR2).1 These specific splicing factor genes are thought to be important for early spliceosome assembly and are commonly altered in hematologic malignancies.1,2SRSF2 mutation has been linked directly to myeloid and erythroid dysplasia in Srsf2-mutant mice, which supports the role of SRSF2 in the pathogenesis of myelodysplastic syndrome (MDS).3,4SRSF2 mutations are reported in various myeloid neoplasms (MNs) and up to 25% of acute myeloid leukemia (AML) cases.4,5
SRSF2 mutations are known to be associated with poor outcomes in patients with MDS and a higher risk of progression to AML.6-8 Similarly, SRSF2 mutations are independently associated with poor survival in patients with primary myelofibrosis.9 Although common in chronic myelomonocytic leukemia (CMML), the prognostic impact of SRSF2 on CMML remains elusive.10
Recent emerging evidence suggests that SRSF2 mutations may be associated with poor outcomes in patients with AML.1,4,11,12 Specifically, the updated 2022 European LeukemiaNet (ELN) risk stratification for AML includes SRSF2 mutations in the adverse category if no co-occurring favorable factors are identified.11 Another recent study, however, has indicated that splicing factor gene mutations are not independent prognostic factors in AML when other covariates, such as age, history of myeloid disorders, cytogenetics, and concurrent mutations, are considered.13
Importantly, previous studies have demonstrated that patients with AML and MN who have SRSF2 mutations can have overall survival (OS) similar to patients with SRSF2 wild type (WT) following a hematopoietic stem cell transplant (SCT).14,15 These studies demonstrate the possible opportunity for using SRSF2 mutation status in patient stratification for treatment decisions. Relatively few studies have directly assessed SRSF2 mutations in AML, but the limited available literature has evaluated the prognostic impact of SRSF2 mutation and its potential effect on clinical treatment decisions. In this retrospective study, we characterized morphology, immunophenotypic, molecular genetic, and clinical findings of SRSF2-mutant AML and nonacute MNs, including any myeloproliferative neoplasm (MPN), CMML, and MDS. We demonstrated that SRSF2 appeared to be an independent worse prognostic factor in AML and allogeneic SCT (allo-SCT) should be considered in SRSF2-mutant AML to improve clinical outcomes.
METHODS
Case Selection
We retrospectively identified patients carrying SRSF2 mutations using a keyword search in our electronic health record system from 2017 to 2022. All CMML, MPN, MDS, and AML diagnoses were made according to the 2022 World Health Organization (WHO) classification and International Consensus Classification (ICC).16,17 Selected cases included 85 patients with AML and 49 patients with SRSF2-mutant non-AML cancer. We also included 342 patients with SRSF2-WT AML as a control group for clinical, molecular genetic, and immunophenotypic comparison.
Clinical, morphologic, immunophenotypic, molecular, and cytogenetic data were reviewed. Perl iron stain was evaluated in each case. Morphologic dysplasia and blast count were evaluated by 4 hematopathologists. This retrospective study was institutional review board approved.
Multiparametric Flow Cytometric Study
Flow cytometry (FC) results were evaluated in 74 altered SRSF2 and 57 SRSF2-WT AML cases. Eight-color multiparametric FC (MFC) was performed using BD FACSCanto II instruments (BD Biosciences) according to the manufacturer’s instructions. For initial diagnosis of AML (25,000 events), an acute leukemia panel was performed with the following antibody combinations: (1) CD13/CD15/CD33/CD34/CD45/CD56/HLA-DR, (2) CD11b/CD11c/CD14/CD34/CD64/CD45/CD117, (3) CD19/CD20/CD5/CD10/CD38/κ/λ/CD45, (4) CD1a/CD2/CD3/CD4/CD5/CD7/CD8/CD45, and (5) CyMPO/cyCD3/CD19/cyCD22/cyCD79/CD34/CD45/nTDT. A population of CD34+ or CD117+ myeloblasts or CD64+/CD14+ monocytoid cells with overexpressed lymphoid antigen (CD2, CD5, CD7, or CD56 on blasts or CD56 on monocytoid cells) or mature myeloid antigen (CD11b, CD11c, or CD15 on blasts) or decreased myeloid antigen expression (CD13, CD33, or CD117 on blasts) was considered diagnostic of MN (<20% blasts/equivalents) and AML (≥20% blasts/equivalent).
If previously treated AML was evaluated, an MRD panel (500,000 events, on average) was performed with the following antibody combinations: (1) CD7/CD33/CD11b/CD34/CD13/CD45/CD38/CD19, (2) HLA-Dr/CD117/CD4/CD34/CD123/CD45/CD38/CD19, (3) HLA-DR/CD56/CD36/CD34/CD64/CD45/CD14/CD19, and (4) CD2/CD22/CD5/CD34/CD15/CD45/CD38/CD19. Data analysis was performed using FCS Express, release 5, software (De Novo Software). CD34+, CD34− myeloblasts, monocytic, and granulocytic populations were separately evaluated, following the previously published analytic strategy.18-21 In brief, leukemic blasts were identified as blasts with original leukemia-associated immunophenotype or blasts with immunophenotypes markedly deviating from normal.
In the acute leukemia panel, increased expression was defined as greater than 20% of blasts expressing these antigens that were normally not present. Reduced or negative expression was defined as less than 20% or 0% of blasts expressing these antigens that were normally present. In the AML MRD panel, an aberrantly expressed antigen was defined as normalized median fluorescence intensity and percentage of antigen expression on blasts or monocytes deviating from the normal range established in our laboratory.
Next-Generation Sequencing Method
Nucleic acid was purified from fresh bone marrow or peripheral blood samples using QIAamp DNA Blood Mini Kit (Qiagen). Forty nanograms of input DNA underwent a multiplex polymerase chain reaction targeting all coding exons within 141 myeloid-related genes using the QIAseq Targeted DNA Human Myeloid Neoplasms Panel (Qiagen), as previously described.21 Prepared libraries were subjected to next-generation sequencing (NGS) on a NextSeq 500 instrument (Illumina) to generate FASTQ files. Reads were mapped to a GRCh37 reference using the CLC Genomics Workbench (Qiagen) to generate variant call files, which were processed using Clinical Insight-Interpret (Qiagen) to assess pathogenicity based on American College of Medical Genetics/Association for Molecular Pathology guidelines.22 Quality control metrics, such as depth of coverage, variant allele frequency (VAF), and average quality scores for reported variants, were evaluated individually for pathogenic and likely pathogenic calls compared with variants of unknown significance.
Statistical Analysis
We performed χ2 tests to compare frequencies of MFC abnormality, somatic mutation, and abnormal cytogenetics between groups using the online GraphPad calculator (GraphPad Software). We performed t tests for quantitative data. Patients’ OS was analyzed using the Kaplan-Meier method, and multivariate analysis was performed using Cox regression tests (IBM SPSS). P < .05 was considered statistically significant.
RESULTS
Clinicopathologic Characteristics of Patients
The SRSF2-mutant group included 99 male and 35 female patients (134 total), at a median age of 70 years (range, 47-89 years). The control group with SRSF2-WT and TP53 consisted of 155 male and 121 female patients (276 total), at a median age of 62 years (range, 18-90 years). The control group with SRSF2-WT and mutant-TP53 had 24 male and 42 female patients (66 total), at a median age of 67.5 years (range, 31-86 years). The patient characteristics and genetic abnormalities are summarized in Supplementary Table 1 and Table 2 (all supplementary material is available at American Journal of Clinical Pathology online).
Patient characteristics of the SRSF2-mutant group, including age, sex, NGS molecular findings, WHO classification, and ICC classification are summarized in TABLE 1. The patients were predominantly male. There were 49 patients with non-AML cancer and 85 patients with AML. Among patients in the non-AML group, MDS was the most common subtype (n = 23 [46.9%]). These subtypes included MDS with low blasts and single-lineage dysplasia (n = 4 [8.2%]), MDS with low blasts and multilineage dysplasia (n = 6 [12.2%]), MDS with low blasts and SF3B1 mutation (n = 2 [4.2%]), MDS with low blasts and 5q deletion (n = 1 [2.1%]), MDS with increased blasts (MDS-IB, n = 8 [16.3%]), and MDS with biallelic TP53 inactivation (n = 1 [2.1%]), according to the 2022 WHO classification. One case of MDS with low blasts, more than 15% ring sideroblasts, and SF3B1-WT, however, could not be optimally classified based on the current WHO scheme. This patient was classified as MDS, not otherwise specified (NOS), with multilineage dysplasia according to the ICC classification. Cases with MDS-IB and more than 10% blasts were classified as MDS/AML in the ICC classification.
TABLE 1.
Clinicopathologic Characteristics of Patients With SRSF2 Mutations
| Characteristic | Values | |||
|---|---|---|---|---|
| Patients, No. | 134 | |||
| Age at diagnosis, median (range), y | 70 (47-89) | |||
| Male/female, No. | 99/35 | |||
| Normal karyotype, % (n/N) | 63.4 (85/134) | |||
| VAF, median % (range) | 41 (4.3%-54%) | |||
| Concomitant mutations other than SRSF2, median (range) | 3 (0-7) | |||
| Classifications | WHO-5 th | ICC | ||
| CCUS, No. (%) | 1 (0.7) | 1 (0.7) | ||
| MDS, No. (%) | 23 (17.2) | 17 (12.7) | ||
| MDS-IB | 8 (6.0) | MDS-EB | 3 (2.2) | |
| MDS-LB-SLD | 4 (3.0) | MDS-NOS-SLD | 4 (3.0) | |
| MDS-LB-MLD | 6 (4.4) | MDS-NOS-MLD | 7 (5.1) | |
| MDS-LB-RS | 1 (0.7) | |||
| MDS-SF3B1, No. (%) | 2 (1.4) | 2 (1.4) | ||
| MDS-del(5q), No. (%) | 1 (0.7) | 1 (0.7) | ||
| MDS-biTP53 | 1 (0.7) | MDS/AML with mutated TP53 | 1 (0.7) | |
| MDS/AML, No. (%) | — | MDS/AML | 5 (3.7) | |
| MDS/MPN, No. (%) | 18 (13.4) | 18 (13.4) | ||
| CMML | 13 (9.7) | 13 (9.7) | ||
| CMML-1 | 11 (7.9) | 11 (7.9) | ||
| CMML-2 | 2 (1.5) | 2 (1.5) | ||
| MDS/MPN-N | 1 (0.7) | aCML | 1 (0.7) | |
| MDS/MPN, NOS | 4 (3.0) | 4 (3.0) | ||
| MPN, No. (%) | 6 (4.4) | 6 (4.4) | ||
| PMF | 4 (3.0) | 4 (3.0) | ||
| MPN, NOS | 2 (1.5) | 2 (1.5) | ||
| BPDCN, No (%) | 1 (0.7) | 1 (0.7) | ||
| AML, No. (%) | 85 (63.4) | 85 (63.4) | ||
| AML-NPM1 | 10 (7.4) | 9 (6.7)a | ||
| AML-MR | 59 (44) | AML-MR gene mutations | 26 (19.4) | |
| AML from MDS | 22 (16.4) | |||
| AML from CMML | 6 (4.4) | |||
| Blast phase of MPN | 10 (7.4) | 10 (7.4) | ||
| AML with mutated TP53 | 6 (4.4) | |||
| AML-pCT | 4 (3.0) | t-AML | 4 (3.0) | |
| AML-DDX41 | 2 (1.5) | 2 (1.5) | ||
aCML, atypical chronic myeloid leukemia; AML, acute myeloid leukemia; AML-DDX41, acute myeloid leukemia with germline DDX41 mutation; AML-MR, acute myeloid leukemia myelodysplasia related; AML-NPM1, acute myeloid leukemia with NPM1 mutation (WHO) or mutated NPM1 (ICC); AML-pCT, acute myeloid leukemia post–cytotoxic therapy; BPDCN, blastic plasmacytoid dendritic cell neoplasm; CCUS, clonal cytopenia of undetermined significance; CMML, chronic myelomonocytic leukemia; ICC, International Consensus Classification; MDS, myelodysplastic syndrome; MDS-del(5q), myelodysplastic syndrome with 5q deletion; MDS-biTP53, myelodysplastic syndrome with biallelic TP53 inactivations; MDS-EB, myelodysplastic syndrome with excess blasts; MDS-IB, myelodysplastic neoplasm with increased blasts; MDS-LB-MLD, myelodysplastic neoplasm with low blasts and multilineage dysplasia; MDS-NOS-MLD, myelodysplastic syndrome, not otherwise specified, with multilineage dysplasia; MDS-LB-RS, myelodysplastic neoplasm with low blasts and ring sideroblasts; MDS-LB-SLD, myelodysplastic neoplasm with low blasts and single-linage dysplasia; MDS/MPN-N: myelodysplastic syndrome/myeloproliferative neoplasm with neutrophilia; MDS-SF3B1, myelodysplastic syndrome with mutated SF3B1; MPN, myeloproliferative neoplasm; NOS, not otherwise specified; PMF, primary myelofibrosis; t-AML, therapy-related acute myeloid leukemia; VAF, variant allele frequency; WHO, World Health Organization.
aOne case of AML with mutated TP53 and NPM1 classified as AML with mutated TP53.
There were 13 cases of CMML (26.5%); 1 case of MDS/MPN with neutrophilia (MDS/MPN-N, 2.1%); 4 cases of MPN/MDS, NOS (8.2%); 6 cases of MPN (12.2%); 1 case of clonal cytopenia of undetermined significance (CCUS, 2.1%); and 1 case of blastic plasmacytoid dendritic cell neoplasm (2.1%) according to the current WHO classification. The case of MDS/MPN-N in the 2022 WHO classification remained as atypical chronic myeloid leukemia in the ICC classification.
When compared with the SRSF2-WT AML group, patients with SRSF2-mutant AML were older (median age, 70 years in SRSF2 positive vs 64 years in SRSF2 negative, P < .0001) and significantly male predominant (63% in SRSF2 positive vs 51% in SRSF2 negative, P = .0001). Based on the 2022 WHO classification, AML myelodysplasia related (AML-MR, n = 59 [69.4%]) was the most common subtype, followed by AML with NPM1 mutation (AML-NPM1, n = 10 [11.8%]), blast phase of MPN (n = 10 [11.8%]), AML post–cytotoxic therapy (n = 4 [4.7%]), and AML with germline DDX41 mutation (n = 2 [2.4%]). The AML-MR group also included 4 cases of MDS that progressed to AML when acquiring NPM1 mutations.
Based on the ICC classification, AML-MR gene mutations were the most common (n = 24 [28.2%]), followed by AML progressed from MDS (n = 22 [25.9%]), blast phase of MPN (n = 10 [11.8%]), AML with mutated NPM1 (n = 9 [10.5%]), AML progressed from CMML (n = 6 [7.1%]), AML with mutated TP53 (n = 6 [7.1%]), therapy-related AML (n = 4 [4.7%]), and AML with germline DDX41 mutation (n = 2 [2.4%]). One case with both NPM1 and TP53 mutations was classified as AML with mutated TP53.
The majority of patients with AML (53%) had a prior history of MDS (n = 24), CMML (n = 7), MPN (n = 10), or cytotoxic exposure (n = 4). Forty-eight (56.5%) of the patients with AML showed a normal diploid karyotype. The median VAF of SRSF2 mutation was 41%. In all 134 patients, the median number of concomitant mutations other than SRSF2 was 3 (range, 0-7).
Morphologic Characterization of Patients With SRSF2-Mutant AML
The average blast count on bone marrow evaluations was 49% (range, 5%-95%). Two patients with AML had less than 20% blasts in bone marrow and more than 20% circulating blasts. All 14 patients with AML-NPM1 had a blast count above 20%. Fifty-two (of 84 [62%]) patients demonstrated a variable degree of morphologic dysplasia, including single-lineage (n = 16: dyserythropoiesis, 3; dysmegakaryopoiesis, 4; dysgranulopoiesis, 9), bilineage (n = 20: dysplasia of erythroid and megakaryocytic lineages, 1; dysplasia of erythroid and granulocytic lineages, 3; dysplasia of granulocytic and megakaryocytic lineages, 16), and trilineage dysplasia (n = 16). The frequency of morphologic dysplasia was reduced in patients with de novo AML (n = 17/39 [44%], P = .002) or AML-NPM1 (n = 5/14 [36%], P = .007) compared with secondary AML (n = 35/45 [78%]).
FC Characterization of SRSF2-Mutant AML
Flow cytometry data were available and evaluated in 74 (of 85) patients. Patients with SRSF2-positive/NPM1-positive AML (n = 12) showed characteristically similar immunophenotypic findings to patients with SRSF2-negative/NPM1-positive AML, suggesting that NPM1 is driving the AML phenotype. They either showed typical CD34-negative/HLA-DR (negative or reduced) blasts (n = 9) or blasts with monocytic differentiation (n = 3).
The most commonly expressed aberrant antigen was reduced or no expression of CD33 (n = 21 [28%]), followed by overexpression of CD7 (n = 16 [22%]), overexpression of CD56 (n = 12 [16%]), overexpression of CD123 (n = 10 [14%]), overexpression of CD15 or CD64 (n = 6 [8%]), overexpression of CD11b (n = 5 [7%]), overexpression of TdT (n = 4 [5%]), overexpression of CD2 or CD22 (n = 2 [3%]), and overexpression of CD19 (n = 1 [1%]). However, SRSF2 did not appear to drive any specific phenotype because aberrantly expressed antigens were not significantly different from SRSF2-WT AML (n = 57).
Characterization of Somatic Mutations in Patients with SRSF2-Mutant AML
The median VAF of SRSF2 mutations was 41% (range, 4.3%-54%) for patients with AML and 41% (range, 10%-56%) for patients with non-AML cancer, respectively, suggesting founding clones in all MNs occurring before other subclones. Among the SRSF2-mutant groups, specific mutational data were available for 36 (73%) patients with non-AML cancer and 74 (87%) patients with AML. The vast majority of mutations occurred at amino acid 95, including 33 patients with non-AML cancer and 72 patients with AML. Within the 33 patients with non-AML cancer with a mutation at amino acid 95, 12 (36%) had a P95H mutation, 10 (30%) had a P95L mutation, and 8 (24%) had a P95R mutation. The remaining patients showed fewer common mutations, and 1 (3%) had a deletion at amino acid 95. Among the 72 patients with AML who had a mutation at amino acid 95, 36 (49%) had a P95H mutation, 14 (19%) had a P95L mutation, and 15 (20%) had a P95R mutation. The remaining mutations were less common, and 6 (8%) patients had a deletion at amino acid 95.
Co-occurring mutations were evaluated in patients with SRSF2-mutant disease. The median number of co-mutations alongside SRSF2 was 3 for patients with AML (range, 0-7) and 3 for patients with non-AML disease (range, 0-7). The most common co-occurring mutations in patients with AML were IDH1/2 (n = 38 [45%]), ASXL1 (n = 38 [45%]), RUNX1 (n = 21 [25%]), and TET2 (n = 18 [21%]). The most common co-occurring mutations in patients with non-AML cancer were ASXL1 (n = 22 [45%]), TET2 (n = 16 [35%]), and IDH1/2 (n = 11 [22%]). Among the 13 cases of CMML that did not progress to AML, 2 (15%) showed an ASXL1 co-mutation; among the 7 cases of CMML that progressed to AML, 3 (43%) showed an ASXL1 co-mutation.
Co-occurring mutations were compared between patients with non-AML disease and patients with SRSF2-mutant AML. IDH1/2 and NPM1 mutations were significantly more common in the group with AML TABLE 2 (P = .015 and P = .002). FLT3 and CEBPA mutations showed a trend to occur more frequently in patients with AML (P = .09). Co-occurring mutations were also compared between the patients with SRSF2-mutant and SRSF2-WT AML; IDH1/2, ASXL1, RUNX1, and STAG2 mutations were significantly more common in the patients with SRSF2-mutant AML TABLE 2 (P < .0001 to P = .001). Finally, a complex karyotype occurred less frequently in the SRSF2-mutant AML group (P = .002).
TABLE 2.
Comparison of Genetic Abnormalities in SRSF2-Positive Myeloid Neoplasms vs SRSF2-Negative AML Cases
| Gene |
SRSF2-positive AML, No. (%) (n = 85) |
SRSF2-positive non-AML cancer, No. (%) (n = 49) |
SRSF2-negative AML, No. (%) (n = 342) |
P value SRSF2-positive AML vs SRSF2-positive non-AML cancer |
P value SRSF2-positive AML vs SRSF2-negative AML |
|---|---|---|---|---|---|
| IDH1/2 | 38 (45) | 11 (22) | 42 (12) | .015 | <.0001 |
| AXSL1 | 38 (45) | 22 (44) | 48 (14) | 1.0 | <.0001 |
| FLT3 | 6 (7) | 0 (0) | 28 (8) | .09 | 1.0 |
| N/KRAS | 13 (15) | 7 (14) | 48 (14) | 1.0 | .73 |
| RUNX1 | 21 (25) | 10 (20) | 36 (11) | .68 | .001 |
| CEBPA | 6 (7) | 0 (0) | 15 (4) | .09 | .40 |
| NPM1 | 14 (16) | 0 (0) | 48 (14) | .002 | .57 |
| BCOR | 7 (8) | 1 (2) | 15 (4) | .26 | .17 |
| TP53 | 11 (13) | 2 (4) | 64 (19) | .13 | .26 |
| STAG2 | 16 (19) | 8 (16) | 18 (5) | .80 | .0002 |
| TET2 | 18 (21) | 17 (35) | 52 (15) | .1 | .19 |
| Complex karyotype | 7 (8) | 2 (4) | 79 (23) | .49 | .002 |
AML, acute myeloid leukemia.
The relationship of co-occurring mutations other than SRSF2 was evaluated in all 3 patient groups FIGURE 1. ASXL1 mutations were enriched in SRSF2-mutant groups (P = .0001 and P < .0001) and tended to be co-mutated with STAG2 mutations in both SRSF2-mutant and SRSF2-WT AMLs (P = .01 and P = .001) and with SETBP, ZRSR2, EZH2, and RUNX1 mutations in SRSF2-WT AMLs (P = .0001 to P = .01). ASXL1 and RUNX1 mutations were also mutually exclusive with NPM1 mutations in both SRSF2-mutant and SRSF2-WT AMLs (P = .01 and P = .0002). IDH1/2 mutations tended to be co-mutated with NPM1 mutations in both AML groups (P = .04 and P = .003) and DNMT3A mutation in the SRSF2-WT AML group (P = .02). Finally, NPM1 mutation was positively associated with normal karyotype in both AML groups. TP53 mutation was positively associated with abnormal karyotype in SRSF2-WT AMLs.
FIGURE 1.
Co-occurring mutations in all patients. A, Circos plot of co-occurring genetic alterations in all patients with non-AML cancer and mutated SRSF2.B, Circos plot of co-occurring genetic alterations in all patients with AML and mutated SRSF2. C, Circos plot of co-occurring genetic alterations in all patients with AML and wild-type SRSF2. AML, acute myeloid leukemia.
Prognostic Impact of SRSF2 Mutation on OS in Patients With AML
The prognostic impact of SRSF2 mutation on OS in patients with AML was evaluated. Because patients with TP53-aberrant AML are known to have dismal clinical outcomes, OS was compared among 4 AML patient groups: SRSF2 negative/TP53 negative, SRSF2 positive/TP53 negative, SRSF2 negative/TP53 positive, and SRSF2 positive/TP53 positive FIGURE 2. SRSF2-positive/TP53-negative patients had significantly decreased OS compared with SRSF2-negative/TP53-negative patients (P < .0001). Patients with TP53-positive AML demonstrated the worst OS (P < .0001), regardless of the status of SRSF2 mutation, compared with patients with TP53-negative AML FIGURE 2. SRSF2 did not confer any prognostic impact on OS in NPM1-mutant AML (P > .05, data not shown) or ASXL1-mutant AML (P > .05, data not shown). Because the AML-MR subtype is known to have an adverse prognosis, SRSF2 mutation status was also assessed in this subgroup of patients, and it did not impact OS (P > .05, data not shown).
FIGURE 2.
OS in patients with AML. OS was compared among 4 patient groups: SRSF2-negative/TP53-negative AML (blue, n = 276), SRSF2-positive/TP53-negative AML (red, n = 74), SRSF2-negative/TP53-positive AML (green, n = 66), and SRSF2-positive/TP53-positive AML (orange, n = 11). ****, P < .0001; ***, P < .001; *, P < .05. AML, acute myeloid leukemia; OS, overall survival.
Clinical Outcomes of Patients With AML Based on Therapy
Overall survival in patients with SRSF2-mutated or SRSF2-WT AML was evaluated when the disease was treated with intensive induction therapy (INT [eg, cytarabine and daunorubicin]) vs hypomethylating agents (HMAs [eg, azacitidine-decitabin]) or venetoclax (VEN). Patients with acute promyelocytic leukemia (APL) or TP53 aberrations were excluded from these analyses. Patients with SRSF2-mutant and patients with SRSF2-WT AML treated with INT had a better OS than the group treated with non-INT therapies, likely because of either younger age or better overall fitness FIGURE 3A (P = .001 and P < .0001). Patients with SRSF2-mutant AML showed similar OS to patients with SRSF2-WT when treated with INT (P = .27) but a worse OS with a non-INT regimen (P = .03).
FIGURE 3.
OS in patients with AML, subdivided by chemotherapy, transplant status, and posttransplant survival. A, OS in patients with AML treated by chemotherapy. OS was compared among 4 patient groups: SRSF2-positive intensive (red, n = 30), SRSF2-positive nonintensive (orange, n = 44), SRSF2-negative intensive (blue, n = 180), and SRSF2-negative nonintensive (green, n = 79). B, OS in patients with AML treated with SCT. OS was compared among 4 patient groups: SRSF2-positive transplant (red, n = 23), SRSF2-positive no transplant (orange, n = 51), SRSF2-negative transplant (blue, n = 112), and SRSF2-negative no transplant (green, n = 147). C, OS in AML posttransplant groups. OS was compared among 4 patient groups: SRSF2-positive intensive (red, n = 13), SRSF2-positive nonintensive (orange, n = 10), SRSF2-negative intensive (blue, n = 100), and SRSF2-negative nonintensive (green, n = 12). ****, P < .0001; ***, P < .001; **, P < .01; *, P < .05. Patients with APL or aberrant TP53 were excluded from these analyses. AML, acute myeloid leukemia; APL, acute promyelocytic leukemia; OS, overall survival; SCT, stem cell transplant.
Finally, SCT outcomes were evaluated and compared. Both SRSF2-mutant and the control AML groups showed improved OS when treated with allo-SCT FIGURE 3B (red vs orange lines and blue vs green lines, P < .0001). In addition, patients who received a transplant showed similar survival rates, regardless of SRSF2 mutation status (P = .34), but patients with SRSF2-mutant disease who did not receive a transplant showed worse OS than patients with SRSF2-WT disease FIGURE 3B (orange vs green lines, P < .0001). More importantly, SRSF2 mutations or non-INT chemotherapy regimens were no longer associated with an unfavorable OS in patients with post-transplant AML FIGURE 3C (P > .05).
Multivariate Analysis of OS in Patients With AML
Multivariate analyses were also performed to assess the effect of multiple covariates on OS. Patients with APL were excluded from these analyses. In a Cox multivariate regression model, older age, adverse cytogenetics (according to 2022 ELN guidelines), and SRSF2, TP53, TET2, and BCOR mutations were associated with worse OS, while allo-SCT was associated with better OS TABLE 3. Although insignificant, INT showed a trend toward better OS (P = .07). We also performed a second Cox regression analysis after excluding all TP53-aberrant patients. In this second analysis, older age, adverse cytogenetics, and SRSF2, TET2, and BCOR mutations were associated with worse OS, while allo-SCT was associated with improved OS.
TABLE 3.
Multivariate Cox Regression for Overall Survival in Patients With AML
| Patients with AML and TP53 aberrations | Patients with AML and no TP53 aberrations | |||||
|---|---|---|---|---|---|---|
| Variable | P value | HR | 95% CI | P value | HR | 95% CI |
| Age >60 y | .017 | 1.50 | 1.08-2.07 | .013 | 1.16 | 1.11-2.34 |
| INT | .07 | 0.72 | 0.58-1.02 | .44 | 0.88 | 0.64-1.22 |
| SCT | <.0001 | 0.32 | 0.23-0.45 | <.0001 | 0.30 | 0.21-0.43 |
| Adverse cytogenetics | <.0001 | 1.85 | 1.32-2.59 | .001 | 1.84 | 1.29-2.64 |
| SRSF2 | .036 | 1.42 | 1.02-1.97 | .047 | 1.42 | 1.0-2.02 |
| TP53 | .021 | 1.56 | 1.07-2.28 | — | — | — |
| RUNX1 | .75 | 0.94 | 0.64-1.38 | .71 | 0.93 | 0.63-1.37 |
| TET2 | .019 | 1.48 | 1.07-2.06 | .027 | 1.49 | 1.05-2.12 |
| FLT3 | .32 | 1.29 | 0.78-2.12 | .34 | 1.28 | 0.77-2.15 |
| IDH1 | .40 | 0.80 | 0.49-1.34 | .75 | 0.91 | 0.49-1.70 |
| IDH2 | .24 | 0.77 | 0.49-1.20 | .64 | 0.84 | 0.52-1.35 |
| NPM1 | .68 | 0.92 | 0.61-1.37 | .55 | 0.88 | 0.59-1.33 |
| STAG2 | .31 | 0.78 | 0.48-1.26 | .51 | 0.84 | 0.50-1.41 |
| BCOR | .008 | 2.14 | 1.22-3.76 | .019 | 2.04 | 1.12-3.72 |
AML, acute myeloid leukemia; HR, hazard ratio; INT, intensive induction therapy; SCT, stem cell transplant.
DISCUSSION
Our study contributes to the growing evidence that patients with AML and SRSF2 mutations may experience poor outcomes.1,11,23 Our analysis revealed that SRSF2 mutations had a significantly adverse impact on OS in patients with AML in both univariate and multivariate analyses, but this poor outcome improved with allo-SCT. Our study showed that SRSF2 mutations, along with TP53, TET2, and BCOR mutations; older age; and adverse cytogenetics, served as independent prognostic markers, in agreement with previous literature reports.1,24 Consistent with previous reports, allo-SCT was also associated with a favorable OS in our multivariate analyses.1,13
In this study, OS was similar between patients with SRSF2-mutant AML and patients with SRSF2-WT AML who received allo-SCT, regardless of the type of initial chemotherapy. Grimm et al14 also showed similar incidence of relapse, event-free survival and OS both in patients with SRSF2-mutant AML and in patients with SRSF2-WT AML. Despite negative association with OS in the patients who received HMA and/or VEN in our study, non-INT regimens can potentially serve as alternative management, bridging older patients and those unfit for allo-SCT. Thus, our data support a potential role for SRSF2 mutation status in risk stratification and treatment decisions.
Because the molecular subclassification of hematologic malignancies continues to advance, it is paramount to analyze the impact of all contributing NGS data. We identified multiple significant co-occurring mutations alongside SRSF2 in this study. IDH1/2 and NPM1 mutations were more common in patients with SRSF2-mutant AML than in patients with non-AML cancer. These findings suggest that IDH1/2 and NPM1 may be acquired mutations after SRSF2 and contribute to AML progression. Indeed, Todisco et al2 demonstrated that IDH1/2 co-mutations along with SRSF2 mutations predicted AML phenotype. Similarly, NPM1-mutant MNs would be classified as AML regardless of blast count in the 2022 WHO classification.16
Among patients with AML, ASXL1 and IDH1/2 were significantly more common in patients with SRSF2-mutant AML than in patients with SRSF2-WT AML, which agrees with previous findings.1,13,14 Johnson et al25 showed that patients with AML who had ASXL1 and SRSF2 co-mutations had poor outcomes in a small-scale study. The same study also demonstrated that co-mutated ASXL1 and SRSF2 AML showed genetic and morphologic features similar to CMML.25 Our study does show that ASXL1 and SRSF2 co-mutations occur in CMML, with 15% in patients with CMML that did not progress to AML and 43% that did. Additional studies failed to show compelling evidence that the interaction between ASXL1 and SRSF2 mutations was of prognostic significance.1,26 Our study further illustrated that ASXL1 mutation failed to exert any synergistically negative impact on OS in patients with SRSF2-mutant AML.
A previous study extensively evaluating bone marrow morphology and FC data along the spectrum of MNs, from CCUS to MDS to AML, showed that all 4 cases of CCUS with SRSF2 mutation demonstrated subdiagnostic dysplasia in 1- or 2-cell lineage.23 The current study continues to show similar results, with significant morphologic dysplasia being identified in patients with secondary AML and SRSF2 mutations. Our study continues to suggest that SRSF2 mutations can be involved in early events, leading to morphologic dysplasia. In addition, updated diagnostic guidelines are beginning to define dysplastic MNs based on molecular analysis, and our study supports the use of SRSF2 mutation status as a marker for dysplasia.16
A limited number of studies have been performed to characterize immunophenotypic findings in SRSF2-mutant AML. One study described monocytic differentiation of myeloid blasts as typical immunophenotypic findings in SRSF2/ASXL1 co-mutated AML cases.25 Our study, however, did not show similar findings, despite ASXL1 mutation being the most common co-mutation. Underexpression of CD33 and overexpression of CD7 and CD56 were the most common immunophenotypic aberrancies in patients with SRSF2-mutant AML, but no specific pattern of immunophenotypic findings could be identified when compared with patients who had SRSF2-WT, likely because of the heterogeneous genetic background of this group of patients.
Limitations of this study include data coming from just 1 institution and the retrospective nature of the study. In summary, SRSF2 mutations continue to be strongly associated with dysplastic myeloid disorders, including AML. In support of recent evidence, our study demonstrates that patients with AML and SRSF2 mutations experience adverse outcomes, but these outcomes can be improved through allo-SCT.
Supplementary Material
Acknowledgments
We thank the staff of the Clinical Molecular Oncology Laboratory and Clinical Flow Cytometry Laboratory for supporting mutation and FC analyses.
Funding: This work was supported in part by the Kansas Institute for Precision Medicine Center of The National Institute of General Medical Sciences and The Centers of Biomedical Research Excellence grant (P20 GM130423), the National Institutes of Health/National Cancer Institute Cancer Center Support grant (P30 CA168524), and the Kansas Bioscience Authority Eminent Scholar grant to A.K.G. A.K.G. is the Chancellors Distinguished Chair in Biomedical Sciences Endowed Professor at the University of Kansas Medical Center.
This article is available for CME credit.Go to academic.oup.com/ajcp/pages/journal_cme to see the latest articles. Thecomplete catalog of journal CME coursescan be found at store.ascp.org.
Contributor Information
Joshua Tatarian, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Natalie Tupper, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Peng Li, Division of Hematopathology, Department of Pathology, University of Utah, Salt Lake City, UT, US.
Julie Feusier, Division of Hematopathology, Department of Pathology, University of Utah, Salt Lake City, UT, US.
Maryam Abdo, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Stephen Hyter, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Patrick R Gonzales, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Da Zhang, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Janet Woodroof, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Sarah Kelting, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Andrew K Godwin, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US; Kansas Institute for Precision Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Wei Cui, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, KS, US.
Conflict of interest disclosure
The authors declare no conflict of interest.
REFERENCES
- 1. van der Werf I, Wojtuszkiewicz A, Meggendorfer M, et al. Splicing factor gene mutations in acute myeloid leukemia offer additive value if incorporated in current risk classification. Blood Adv. 2021;5(17):3254-3265. 10.1182/bloodadvances.2021004556 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Todisco G, Creignou M, Galli A, et al. Co-mutation pattern, clonal hierarchy, and clone size concur to determine disease phenotype of SRSF2(P95)-mutated neoplasms. Leukemia. 2021;35(8):2371-2381. 10.1038/s41375-020-01106-z [DOI] [PubMed] [Google Scholar]
- 3. Kim E, Ilagan JO, Liang Y, et al. SRSF2 mutations contribute to myelodysplasia by mutant-specific effects on exon recognition. Cancer Cell. 2015;27(5):617-630. 10.1016/j.ccell.2015.04.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Visconte V, Nakashima MO, Rogers HJ.. Mutations in splicing factor genes in myeloid malignancies: significance and impact on clinical features. Cancers (Basel). 2019;11(12):1844. 10.3390/cancers11121844 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Marneth AE, Mullally A.. The molecular genetics of myeloproliferative neoplasms. Cold Spring Harb Perspect Med. 2020;10(2):a034876. 10.1101/cshperspect.a034876 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Zheng X, Zhan Z, Naren D, Li J, Yan T, Gong Y.. Prognostic value of SRSF2 mutations in patients with de novo myelodysplastic syndromes: a meta-analysis. PLoS One. 2017;12(9):e0185053. 10.1371/journal.pone.0185053 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Makishima H, Visconte V, Sakaguchi H, et al. Mutations in the spliceosome machinery, a novel and ubiquitous pathway in leukemogenesis. Blood. 2012;119(14):3203-3210. 10.1182/blood-2011-12-399774 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Wu SJ, Kuo YY, Hou HA, et al. The clinical implication of SRSF2 mutation in patients with myelodysplastic syndrome and its stability during disease evolution. Blood. 2012;120(15):3106-3111. 10.1182/blood-2012-02-412296 [DOI] [PubMed] [Google Scholar]
- 9. Vannucchi AM, Lasho TL, Guglielmelli P, et al. Mutations and prognosis in primary myelofibrosis. Leukemia. 2013;27(9):1861-1869. 10.1038/leu.2013.119 [DOI] [PubMed] [Google Scholar]
- 10. Aujla A, Linder K, Iragavarapu C, Karass M, Liu D.. SRSF2 mutations in myelodysplasia/myeloproliferative neoplasms. Biomark Res. 2018;6:29. 10.1186/s40364-018-0142-y [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Döhner H, Wei AH, Appelbaum FR, et al. Diagnosis and management of AML in adults: 2022 recommendations from an international expert panel on behalf of the ELN. Blood. 2022;140(12):1345-1377. 10.1182/blood.2022016867 [DOI] [PubMed] [Google Scholar]
- 12. Wu M, Guo ZW, Huang GN, Ye YB.. Features and impacts on the prognosis of gene mutations in patients with acute myeloid leukemia. Neoplasma. 2021;68(5):1072-1078. 10.4149/neo_2021_201230N1426 [DOI] [PubMed] [Google Scholar]
- 13. Senepati J, Short N, Maiti A, et al. Splicing gene mutations do not affect outcomes in newly diagnosed patients with acute myeloid leukemia. Blood. 2022;140(suppl 1):1427-1429. 10.1182/blood-2022-169934 [DOI] [Google Scholar]
- 14. Grimm J, Jentzsch M, Bill M, et al. Clinical implications of SRSF2 mutations in AML patients undergoing allogeneic stem cell transplantation. Am J Hematol. 2021;96(10):1287-1294. 10.1002/ajh.26298 [DOI] [PubMed] [Google Scholar]
- 15. Hamilton BK, Visconte V, Jia X, et al. Impact of allogeneic hematopoietic cell transplant in patients with myeloid neoplasms carrying spliceosomal mutations. Am J Hematol. 2016;91(4):406-409. 10.1002/ajh.24306 [DOI] [PubMed] [Google Scholar]
- 16. Khoury JD, Solary E, Abla O, et al. The 5th edition of the World Health Organization Classification of Haematolymphoid Tumours: Myeloid and Histiocytic/Dendritic Neoplasms. Leukemia. 2022;36(7):1703-1719. 10.1038/s41375-022-01613-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Arber DA, Hasserjian RP, Orazi A, et al. Classification of myeloid neoplasms/acute leukemia: global perspectives and the international consensus classification approach. Am J Hematol. 2022;97(5):514-518. 10.1002/ajh.26503 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Jaso JM, Wang SA, Jorgensen JL, Lin P.. Multi-color flow cytometric immunophenotyping for detection of minimal residual disease in AML: past, present and future. Bone Marrow Transplant. 2014;49(9):1129-1138. 10.1038/bmt.2014.99 [DOI] [PubMed] [Google Scholar]
- 19. Ouyang J, Goswami M, Tang G, et al. The clinical significance of negative flow cytometry immunophenotypic results in a morphologically scored positive bone marrow in patients following treatment for acute myeloid leukemia. Am J Hematol. 2015;90(6):504-510. 10.1002/ajh.23988 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Xu J, Jorgensen JL, Wang SA.. How do we use multicolor flow cytometry to detect minimal residual disease in acute myeloid leukemia? Clin Lab Med. 2017;37(4):787-802. 10.1016/j.cll.2017.07.004 [DOI] [PubMed] [Google Scholar]
- 21. McGowan PF, Hyter SD, Cui W, Plummer RM, Godwin AK, Zhang D.. Comparison of flow cytometry and next-generation sequencing in minimal residual disease monitoring of acute myeloid leukemia: one institute’s practical clinical experience. Int J Lab Hematol. 2022;44(1):118-126. 10.1111/ijlh.13711 [DOI] [PubMed] [Google Scholar]
- 22. Richards S, Aziz N, Bale S, et al. ; ACMG Laboratory Quality Assurance Committee. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015;17(5):405-424. 10.1038/gim.2015.30 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Gao Y, Jia M, Mao Y, et al. Distinct mutation landscapes between acute myeloid leukemia with myelodysplasia-related changes and de novo acute myeloid leukemia. Am J Clin Pathol. 2022;157(5):691-700. 10.1093/ajcp/aqab172 [DOI] [PubMed] [Google Scholar]
- 24. Montalban-Bravo G, Kanagal-Shamanna R, Class CA, et al. Outcomes of acute myeloid leukemia with myelodysplasia related changes depend on diagnostic criteria and therapy. Am J Hematol. 2020;95(6):612-622. 10.1002/ajh.25769 [DOI] [PubMed] [Google Scholar]
- 25. Johnson SM, Richardson DR, Galeotti J, et al. Acute myeloid leukemia with co-mutated ASXL1 and SRSF2 exhibits monocytic differentiation and has a mutational profile overlapping with chronic myelomonocytic leukemia. Hemasphere. 2019;3(5):e292. 10.1097/HS9.0000000000000292 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Kang D, Jung J, Park S, et al. Genetic characteristics according to subgroup of acute myeloid leukemia with myelodysplasia-related changes. J Clin Med. 2022;11(9):2378. 10.3390/jcm11092378 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.