Spliceosomal gene aberrations are rare, coexist with oncogenic mutations, and are unlikely to exert a driver effect in childhood MDS and JMML

Recent whole exome sequencing studies identified recurrent somatic mutations in spliceosome complex genes in a significant proportion of adult patients with myelodysplastic syndrome (MDS).^1-3  The biologic significance of such mutations remains to be fully elucidated; some mutations (eg, those involving SF3B1) may correlate with a better overall survival.²  The spectrum of spliceosome genes commonly affected includes SF3B1 (mutated in up to 83% of refractory anemia with ringed sideroblasts), U2AF35 (affected in up to 18% of refractory cytopenia with multilineage dysplasia), and SRSF2 (abnormal in up to 31% of chronic myelomonocytic leukemia [CMML] cases).¹  The lesions occur as somatic heterozygous missense mutations. They cluster to 2 hotspots in U2AF35, targeting protein residues S34 or Q157, or to 1 hotspot in SRSF2 (targeting P95) or SF3B1 (targeting K700) each. The genes code for components of the splicing machinery responsible for processing of pre-mRNA to mature RNA.^1,4  We hypothesized that the disruption of the spliceosome complex might play a central role in the pathogenesis of childhood MDS and juvenile myelomonocytic leukemia (JMML). We used targeted resequencing to investigate the 4 hotspots of 3 spliceosome genes in 371 pediatric and 30 adult cases.

Pediatric patients were enrolled in study 98 or 2006 (www.clinicaltrials.gov; NCT00047268 and NCT00662090) of the European Working Group of MDS in Childhood. Adult MDS samples were enrolled in multicenter treatment trials that investigated the use of all-trans retinoic acid,⁵  antithymocyte globulin (NCT00004208),^6,7  deferasirox,⁸  lenalidomide, or thalidomide for treatment of MDS. Informed consent to use biologic material had been obtained from patients or patients' guardians after approval by the institutional review board of each participating institution in accordance with the Declaration of Helsinki.

Targeted resequencing was performed using genomic DNA as detailed in supplemental Table 1 (see the Supplemental Materials link at the top of the article). Mutations were confirmed in at least 2 independent runs.

TA subcloning and sequencing of bacterial colonies were performed as previously described.⁹  On average, 26 colonies were sequenced per single ligated amplicon of U2AF35 or KIT.

Longitudinal denaturing high-performance liquid chromatography analysis of NRAS exon 1 and exon 2 was performed using primers listed in supplemental Table 1.

We analyzed 4 spliceosome gene hotspots in 401 patients with MDS and JMML; the subtypes are listed in Table 1. All 371 children were enrolled in European Working Group of MDS in Childhood studies and 30 adult MDS patients participating in various clinical trials^5-8  were included. Because mutations in spliceosome genes occur somatically in leukemic cells and are not found in healthy controls (dbSNP and the 1000 Genomes databases), an additional control cohort was not included in this study. As expected, results in the adult MDS population were comparable with those reported previously¹ : 10 of 30 samples had a spliceosome mutation. The SRSF2 gene was affected most frequently. Of 7 patients with SRSF2 mutation, 4 carried the known mutations c.284C > T/A/G(p.P95L/H/R) and in 3 patients a novel deletion SRSF2 c.284_307del(p.P95_R102del) was detected. The U2AF35 and SF3B1 mutations were discovered in 2 patients and 1 patient, respectively (Table 1).

By contrast, in the cohort of 371 children, spliceosome mutations were found in only 3 cases (Figure 1A; supplemental Figure 1). A heterozygous SRSF2(p.P95L) mutation, being the most prevalent mutation in adult CMML,¹  was identified in 2 JMML cases (patients SC092 and D361). Analysis of DNA from buccal epithelial cells confirmed the somatic nature of this mutation in SC092 (supplemental Figure 1). Hematopoietic cells of both children had a normal karyotype and concomitant mutations of the RAS pathway: PTPN11(p.E76Q) in SC092 and NRAS(p.G13D) in D361 (Figure 1A). Patient SC092 was successfully transplanted and is currently disease-free. Patient D361 carried a SRSF2 and a singular NRAS mutation at diagnosis (Figure 1B; supplemental Figure 1D). He relapsed after hematopoietic stem cell transplantation (HSCT). At relapse, the SRSF2(p.P95L) mutation was no longer detectable, whereas the initial NRAS(p.G13D) mutation persisted. After a second HSCT, which failed to induce remission, splenectomy was performed because of acceleration of the disease. In spleen cells, the SRSF2 mutation remained absent, but strikingly, a new NRAS(p.Q61K) mutation had appeared in addition to the preexisting NRAS(p.G13D) mutation. The hypothetical model of clonal evolution for this patient is shown in Figure 1B. At diagnosis, 2 distinct clones are postulated to exist: an NRAS(p.G13D)–bearing primary clone, which was outcompeted by a faster-growing but less refractory offspring clone carrying a SRSF2(p.P95L) mutation as an additional lesion. Although HSCT eradicated this secondary clone, the primary clone persisted and gave rise to disease recurrence. Under selective pressure exerted through second HSCT, a novel secondary clone with NRAS p.G13D/Q61K double mutation emerged.

The third patient with a spliceosome mutation (D717) was diagnosed at the age of 17 years with systemic mastocytosis with a KIT(p.D816V) mutation, normal cytogenetics, and associated clonal hematologic nonmast cell lineage disease, which resembled MDS, subtype refractory cytopenia (Figure 1C). Although the mastocytosis component of the disease responded to interferon-alpha2b, HSCT was performed for persistent thrombocytopenia and anemia. Two years after HSCT, the patient relapsed with 35% autologous cells. Allele-specific clonal sequencing of affected U2AF35 and KIT exons revealed the presence of both mutations in the sample at diagnosis and at time of disease recurrence. Total allelic burdens of the 2 mutations were similar both at diagnosis and relapse (Figure 1C), indicating that the KIT(p.D816V) and the U2AF35(p.S34Y) mutations were likely present in the same clone. As previously postulated, it is probable that the systemic mastocytosis and associated clonal hematologic nonmast cell lineage disease compartments in this patient originated from a common neoplastic precursor.^14,15 

The drastically reduced frequency of spliceosome mutations in pediatric compared with adult MDS suggests a different pathogenetic mechanism in childhood disorders; the pathogenesis in children is probably more related to genetic determinants acquired prenatally and less to accumulation of somatic events during life. This view fits well with previous reports that somatic mutations of genes, such as DNMT3A, TET2, IDH, and ASXL1, are also much less prevalent in pediatric MDS.^16-19 

In conclusion, the data reported here suggest that alterations of the spliceosome complex do not constitute a fundamental etiologic factor in pediatric MDS and JMML. First, they are exceedingly rare in a group of diseases that is infrequent in itself. Second, in the positive cases identified here, spliceosome mutations did not occur independently of oncogenic mutations and in 1 case even disappeared during disease evolution. It is reasonable to assume that in children these mutations coexist as passengers in an oncogenic environment alongside primary mutations of signal transduction pathways, known to act as drivers in the malignant process.

The authors thank Sandra Urbaniak for excellent technical assistance, Alexandra Fischer and Wilfried Truckenmüller for expert data management, and Annamaria Cseh for case review.

This work was supported by Deutsche Krebshilfe (109005; M.W.W.) and Deutsche Forschungsgemeinschaft (FL345/4-1; C.F.).

Contribution: S.H. performed research and analyzed data; C.F. interpreted data and wrote the manuscript; J.M. performed research; M.H., H.H., B.G., T.K., and F.T. recruited patients; B. Schlegelberger reviewed cytogenetics; I.B. reviewed pathology; B. Strahm, J.S., F.L., M.Z., E.B., M.D., M.M.v.d.H.-E., and B.D.M. recruited patients; S.O. designed research; C.M.N. recruited patients and designed research; and M.W.W. designed and performed research and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Marcin W. Wlodarski, Pediatric Hematology and Oncology, University of Freiburg, Mathildenstr 1, 79106 Freiburg, Germany; e-mail: marcin.wlodarski@uniklinik-freiburg.de.