Detection of Subclonal SETBP1 and JAK3 Mutations in Patients with Juvenile Myelomonocytic Leukemia Using Droplet Digital PCR

Background

Juvenile myelomonocytic leukemia (JMML) is a rare pediatric myelodysplastic/myeloproliferative disease. Approximately 85% of patients with JMML harbor germline and/or somatic mutations in RAS pathway genes, such as PTPN11, NF1, CBL, NRAS, and KRAS. In a subset of patients with JMML, SETBP1 and JAK3 mutations were identified as secondary mutations in addition to primary RAS mutations. These secondary mutations are associated with the disease progression and poor　clinical outcome. Recently, it has been reported that subclonal SETBP1 mutation also correlates with a dismal prognosis. Therefore, we hypothesized that subclonal JAK3 mutation is present in a higher than expected number of patients with JMML and associated with poor prognosis. The aim of this study is to identify patients with subclonal SETBP1 and/or JAK3 mutations at the diagnosis using droplet digital PCR (ddPCR) and to elucidate their clinical outcomes.

Patients and Methods

We enrolled 128 patients with JMML and 15 with Noonan syndrome-associated myeloproliferative disorder (NS/MPD). Using bone marrow (BM) or peripheral blood derived genomic DNA, ddPCR was performed in the 143 patients to detect SETBP1 p.D868N and JAK3 p.R657Q hotspot mutations with low variant allele frequencies (VAF). The study was approved by the institutional review board of Nagoya University Graduate School of Medicine.

Results

To assess the false-positive rate of the ddPCR assay for each mutation, the assay was also performed in 30 healthy volunteers. Among these, the false-positive rate (mean ± standard deviation) for SETBP1 and JAK3 mutations was 0.010% ± 0.010% and 0.013% ± 0.012%, respectively. Due to the presence of false-positive droplets, the sensitivity and the quantitative linearity was evaluated for >0.01% VAF. The significant correlation between the expected and the observed VAF in SETBP1 and JAK3 was observed (R-squared, 0.9923 and 0.9922, respectively). Therefore, 0.05% VAF was defined as the cut-off value in this assay.

Among the 143 patients, ddPCR detected SETBP1 and JAK3 mutations in nine (6.3%) and fifteen (10.5%), respectively. SETBP1 and JAK3 mutations, including variants with low allele frequencies, were not detected in NS/MPD. Among patients with SETBP1 and/or JAK3 mutations, two and six patients harbored less than 1.0% VAF. Patients with less than 1.0% VAF in SETBP1 or JAK3 mutation exhibited a significantly poorer 2-year transplantation-free survival than those without SETBP1 and JAK3 mutations (P = 3.05 × 10^-3).

JMML is genetically characterized by an extremely small number of somatic mutations (an average of 0.8 mutations/exome/patient). However, we demonstrated that among 19 patients with SETBP1 and/or JAK3 mutations, five patients (26.3%) harbored both the mutations. This finding suggested a statistically significant co-occurrence of SETBP1 and JAK3 mutations in JMML. In order to determine whether SETBP1 and JAK3 mutations were present in the same clone or not, we performed colony formation assays using BM cells in one of the five patients with both SETBP1 and JAK3 mutations. This case harbored 0.9% VAF in JAK3 and 41.9% in SETBP1 in addition to 46.0% in PTPN11 mutation (c.227A>C, p.E76A), respectively. In total, 93 colonies were collected and individually analyzed by Sanger sequencing, of which two colonies (2.1%) were identified with both SETBP1 and JAK3 mutations.

Conclusions

ddPCR is a useful tool to assess subclonal SETBP1 and JAK3 hotspot mutations and to estimate the prognosis. It would be better to start preparing for hematopoietic stem cell transplantation when patients with JMML harbored subclonal SETBP1 and/or JAK3 mutations at the diagnosis.

While JMML is characterized by a paucity of somatic mutations, clones harboring SETBP1 and JAK3 mutations were identified. This finding suggests that SETBP1 and JAK3 mutation are susceptible to each other. Furthermore, the serial acquisition of SETBP1 and JAK3 mutations might correlate with the disease progression.