Comprehensive Mutational Analysis of Juvenile Myelomonocytic Leukemia Using Whole-Genome Sequencing

Background

Juvenile myelomonocytic leukemia (JMML) is a rare and exclusively pediatric myelodysplastic/myeloproliferative neoplasm. This disease is genetically characterized by an extremely small number of somatic mutations (an average of 0.8 mutations/exome/patient). It has been shown that causative somatic and/or germline mutations activating the RAS pathway are located in PTPN11, NF1, NRAS, KRAS, and CBL in 85% of patients with JMML. Furthermore, up to 20% of the patients have additional secondary mutations including SETBP1, and JAK3 mutations. In 2% of the patients, we identified, by RNA sequencing, activating kinase lesions affecting ALK or ROS1. Such findings suggest that other kinase fusions are present in JMML. There is an exceptional scarcity of somatic passenger mutations on the exome, suggesting that a small number of driver mutations drive JMML. However, to date, this hypothesis has not been investigated by whole-genome sequencing.

Patients and Methods

We performed a whole-genome sequencing (WGS) study in 48 patients with JMML. Bone marrow specimens and in vitro-cultured T cells were used as tumor and germline samples, respectively. Next-generation sequencing was performed using a HiSeq X platform (Illumina). Data analysis was performed by our in-house pipeline. Specifically, the pipeline detects single nucleotide variants (SNVs), copy number variants, somatic loss of heterozygosity (LOH), and chromosomal structural variations (SVs). The study was approved by the institutional review board of Nagoya University Graduate School of Medicine.

Results

In each patient we detected an average of 28 somatic mutations. These were primarily C-to-T transition in the CpG context, indicating that the mutations occurred by cell division. Besides RAS pathway and known secondary mutations, we observed no significant accumulation of somatic mutations in either coding or non-coding regions. Although we detected RAS pathway mutations in 90% of the patients, all mutations were on exome. However, we identified germline microdeletions affecting CBL and NF1, which had not been identified by exome sequencing. Additionally, we found two LOH events that affected NF1. Bi-allelic inactivation of NF1 is generally observed in patients with JMML; however, no pathogenic SNVs were identified in these two patients.

We identified two chromosomal translocations that caused activating kinase lesions (i.e., RANBP2-ALK and TBL1XR1-ROS1). These had been pointed out in our previous RNA sequencing study. Another patient carried a complex SV that affected XPO1 (encoding exportin 1 or chromosome region maintenance 1 protein homolog). Although fusion genes involving XPO1 are reported to be present in lymphoid malignancies, the role of this SV in JMML remains unclear.

Conclusions

JMML is characterized by driver mutations that are largely present within the exome. However, WGS can still play a role in identifying both coding and non-coding mutations. LOH events without pathogenic SNVs suggest the presence of novel regulatory mechanisms of NF1.

Conclusively, JMML is characterized by a paucity of somatic alterations and driver mutations. Hence, current research efforts should focus on RAS pathway mutations and known secondary mutations, many of which can be targeted.