Immunophenotypic and Genetic Overlap between JMML and CMML

Introduction:

Juvenile myelomonocytic leukemia (JMML) is a rare hematological malignancy of early childhood with characteristics of both myeloproliferative neoplasms and myelodysplastic syndromes. JMML shares pathological features and diagnostic criteria with chronic myelomonocytic leukemia (CMML), a malignancy predominantly affecting the elderly. While 85% of patients with JMML have somatic or germline mutations in RAS pathway genes (NF1, NRAS, KRAS, PTPN11, and CBL), the most frequently mutated genes in CMML include TET2, SRSF2, ASXL1, and RAS and are generally somatic-only. The extent to which histone modification genes (ASXL1, EZH2) or spliceosome machinery genes (SF3B1, SRSF2, U2AF1, ZRSR2) play a role in JMML pathogenesis is unclear. Despite mutational differences, both JMML and CMML manifest as myelomonocytic proliferation with varying amounts of dysplasia in the bone marrow. Clusters of clonally-related CD123+ plasmacytoid dendritic cells (PDCs) have been observed in the bone marrow of patients with CMML but have not been investigated in JMML. Here, we report the mutation profiles and immunophenotypic characteristics of JMML specimens from children treated at our institution.

Methods:

The pathology archives (1987-2017) at the Children's Hospital of Philadelphia (CHOP) were searched to identify JMML cases (n=21) and included formalin fixed paraffin-embedded diagnostic bone marrow biopsies and splenectomy tissue obtained prior to hematopoietic stem cell transplant. JMML diagnosis was confirmed in all cases by clinicopathological review. Cytogenetic analysis and whole genome SNP array were performed at initial clinical presentation. Genomic DNA and RNA were extracted from JMML patients' bone marrow (n=8) and spleen tissue (n=10) for next-generation sequencing analysis of 118 cancer genes for sequence and copy number variants and 110 genes for known and novel fusions via our custom CHOP Hematologic Cancer Panel.

CD123 immunohistochemical (IHC) staining was performed on bone marrow and spleen tissues from children with JMML. Presence of CD123+ PDC clusters was evaluated manually and by digital image analysis. CD123 staining was enumerated using the Aperio Image Scope quantitation of membranous staining v9 with the analysis parameters set such that normal endothelial staining was quantified as 1+, and true CD123 staining cells were quantified as 2+ or 3+. The percentage of CD123+ cells (out of total cellularity) was calculated. Bone marrow from patients with non-JMML myeloid malignancies (n=6) and splenectomy tissue from patients with sickle cell anemia (n=8) were used as controls for the CD123 IHC analysis.

Results:

We confirmed canonical JMML-associated somatic or germline NF1 (n=3), NRAS (n=4), KRAS (n=2), PTPN11 (n=6), or CBL (n=2) mutations in 16 of the 17 (94%) patients with sequencing data. Interestingly, both PTPN11^A72T and NF1^R2637* mutations were detected in one patient. In addition, we found potential variants in genes affecting histone modifications (ASXL1, DNMT3A, KDM6A, SETD2), spliceosomal processes (SF3B1, U2AF1), transcription (BCOR, RUNX1, ETV6), or cellular growth (SETBP1, BRAF) in 8/17 patients (47%). While mutations in these genes have been well-characterized in other myeloid disorders, many of these alterations have not been reported to date in children with JMML or are currently of unclear biologic and prognostic significance.

We also observed increased clustering of CD123+ PDCs in bone marrow and spleens from patients with JMML compared to IHC staining of control tissues. 2.2 ± 0.42% and 1.8 ± 0.74% of cells expressed CD123 in the spleen and bone marrow specimens, respectively. Control bone marrow and spleen samples did not show significant CD123+ staining.

Conclusions:

Our study demonstrates frequent variants in histone modification, splicing, and transcription-associated genes in JMML specimens in addition to known pathogenic RAS pathway mutations. We further report histopathologic CD123+ PDC clustering in JMML specimens analogous to that observed in CMML, which may aid in the workup of this often difficult-to-diagnose disease. Our findings of genetic and immunophenotypic overlap between JMML and CMML suggest similarities in pathogenesis despite typical presentation at extremes of age.