Mutational analysis of RNA splicing machinery components in 206 children with myeloid malignancies

Abstract

Myeloid malignancies involve a clonal proliferation of abnormal stem cells and show various phenotypes consisting of acute myeloid leukemia (AML), myeloproliferative neoplasms (MPN) and myelodysplastic syndromes (MDS) including various overlap and transition forms of these diseases1. Juvenile myelomonocytic leukemia (JMML) is an aggressive myeloid neoplasm of early childhood that is clinically characterized by overproduction of monocytic cells that can infiltrate organs, including the spleen, liver, gastrointestinal tract, and lung. JMML is categorized as an overlap MPN/MDS by the World Health Organization (WHO) and also shares some clinical and molecular features with chronic myelomonocytic leukemia (CMML), a pathomorphologically related disease in adults. Or both CMML and JMML have the following clinical characteristics: 1) persistent peripheral blood monocytosis >1 × 106/ml; 2) no Philadelphia chromosome; 3) no eosinophilia; 4) fewer than 20% blasts in the blood and bone marrow; and 5) dysplasia in one or more myeloid lineages. Recently, we have applied whole exome sequencing to a group of MDS patients followed by a targeted screen of specific genes for putatively recurrent mutations identified in the initial whole exome sequencing screen. In a systematic analysis of somatic events tested for recurrence and potential functional significance, we noted several somatic mutations affecting genes involved in the splicing machinery, consisting of SF3B1 (splicing factor 3b, subunit 1), U2AF1 (U2 small nuclear RNA auxiliary factor 1) and SRSF2 (serine/arginine-rich splicing factor 2) 2, 3, which Yoshida K and Papaemmanuil E reported to be frequently mutated in adult MDS, AML, and in particular chronic myelomonocytic leukemia (CMML) patients4, 5. These findings suggested that mutations of the genes involved in splicing machinery were a common pathogenetic pathway in adult myeloid malignancies. However, while it is likely that spliceosomal genes are also involved in the pathogenesis of childhood myeloid malignancies, including JMML, to date such a screen has been I performed by only one institute6. In the past, several similarities (e.g., CBL mutations in CMML and JMML) but also significant differences (e.g., TET2 mutations in CMML and MDS but not in JMML) were found in the mutational spectrum of patients with seemingly related disorders of adults and children. Consequently, we have designed this study to establish whether spliceosomal mutations contribute to the pathogenesis of pediatric myeloid malignancies, in particular JMML. We have studied 206 children, aged 0–18 years with myeloid malignancies, consisting of 102 with AML, 94 with JMML, and 10 with MDS. Patients with AML and MDS were classified according to the FAB classification. A total of 102 AML patients were studied [M0=1, M1=7, M2=16, M3 variant=1, M4=9, M5=14, M7=44 (including 21 patients with Down syndrome), and 10 with unknown subtypes]. All of the 10 children with MDS were classified as refractory anemia with excess blasts (RAEB) or RAEB in transformation (RAEB-T). JMML was defined as described in the 2008 WHO classification. Four of 94 JMML patients had clinical evidence of neurofibromatosis type 1. The results in children were contrasted with those obtained in an additional series of adult CMML tested for the presence of somatic mutations in the same set of genes. Previously we performed mutation analyses of PTPN11, NRAS, KRAS and CBL genes in 94 JMML patients, and detect genetic mutations in 29 (31%) for PTPN11, 12 (13%) for NRAS, 11 (12%) for KRAS and 7 (7%) for CBL genes7. Moreover, EZH2, RUNX1, ASXL1 and TET2 mutations were found in 0%, 0%, 4% and 0% of patients with JMML, respectively7,8. In contrast, our previous studies in adult CMML have shown the presence of TET2, ASXL1, CBL, EZH2, NRAS and KRAS, in 49%, 43%, 14%, 6%, 7% and 4% of patients, respectively6–8. In the current study, we analyzed genomic DNA derived from patients’ bone marrow or blood samples obtained before therapy according to appropriate protocol and consent approved by the intitutional review board of the two academic research centers. According to our initial screen in 50 adult patients with CMML, we have detected 2/50 mutations in exon 14 and 15 of SF3B1, 8/50 in exon 2 and 6 of U2AF1 and 18/50 in exon 1 of SRSF2 genes. The screening of mutations in these exons were performed by polymerase chain reaction (PCR) and direct sequencing. These exons were amplified from genomic DNA using the Platinum® Taq DNA Polymerase (Life Technologies, Carlsbad, CA, USA) and the following forward and reverse primers: SF3B1_ex13-14-F (5′-TCCCTTGATTAACAAAAGTCCTG-3′), SF3B1_ex13-14-R (5′-TGAGTCCAGTCTGGGCAAC-3′), SF3B1_ex15-16-F (5′-GTTGATATATTGAGAGAATCTGGATG-3′), SF3B1_ex15-16-R (5′-TTTAAAATTCTGTTAGAACCATGAAAC-3′), U2AF1_ex2-F (5′-GCTGCTGACATATTCCATGTG-3′), U2AF1_ex2-R (5′-AAGTCGATCACCTGCCTCAC-3′), U2AF1_ex6-F (5′-CATTTGGCAAAATCTTGGAC-3′), U2AF1_ex6-R (5′-TGGATAATGAAGTTACCTGTGTGC-3′), U2AF1_ex6-FS (5′-CGTGGATGGCAAGCACTTCTGTTT-3′), SRSF2_ex1-F (5′-AAGGCAACTGCCTGAGAGG-3′), SRSF2_ex1-R (5′-CGGACCTTTGTGAGGTCG-3′). In addition, we evaluated sequence analysis of spliceosome genes ZRSR2 (zinc finger, RNA-binding motif and serine/arginine rich 2) and LUC7L2 (LUC7-like 2) which we had identified as somatic mutations in adult MDS patients by whole exome sequencing3. The forward and reverse primers for ZRSR2 and LUC7L2 were summarized in supplemental table 1. The purified PCR products were directly sequenced using Big Dye Terminator V3.1 (Life Technologies) and analyzed using the ABI-3130 Genetic Analyzer. The most recurrent mutations in adult CMML cohort were S34F in U2AF1 and P95H in SRSF2. However, when children with myeloid malignancies including JMML were screened, no mutations in any of the genes were found in the 206 cases. Thus if present, spliceosomal mutations must occur at a frequency of <1% in pediatric leukemia. Our results suggested that there are distinct differences in the molecular pathogenesis of adult and childhood myeloid malignancies. This is further supported by the study of Yoshida, et al. which showed that the mutations of splicing machinary genes were mainly found in CMML patients4 and not in JMML, which was consistent with the report by Hirabayashi, et al6. JMML shows very similar clinical features as CMML and is considered to be the pediatric equivalent of CMML. However, the striking differences in the genetic profiles of JMML and CMML have recently been revealed; as shown in figure 1, alterations in PTPN11 and NF1 are specific to JMML, while alterations in TET2, JAK2, and RUNX1 are specific to CMML, and alterations in RAS and CBL are common to both7–11; while some CBL mutations are germline in JMML, they are acquired in CMML. The current study found that genetic mutations in RNA splicing machinary components were also specific to CMML but not to JMML, supported the notion of a distinct pathogenetic roots in JMML and CMML. Figure 1 The figure showed a scheme of difference of genetic mutations between CMML and JMML. Alterations in PTPN11 and NF1 are specific to JMML, while alterations in TET2, JAK2, RUNX1 and splicing machinery are specific to CMML, and alterations in RAS and CBL ...