1. ¹Department of Pediatrics, Keio University School of Medicine, Tokyo, Japan
2. ²Molecular Oncology Division, National Cancer Center Research Institute, Tokyo, Japan

Correspondence: Dr T Mori, Department of Pediatrics, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan. E-mail: morite@sc.itc.keio.ac.jp

Received 13 December 2003; Accepted 11 March 2004; Published online 15 April 2004.

TO THE EDITOR

Juvenile myelomonocytic leukemia (JMML) is a clonal myeloproliferative disorder seen in young children and is characterized by leukocytosis with monocytosis, thrombocytopenia, hepatosplenomegaly, and elevated fetal hemoglobin caused by a reversion to fetal-like erythropoiesis.¹ The Ras signaling pathway is known to be deregulated by mutations in the RAS gene or the neurofibromatosis type 1 gene (NF1) in approximately 40% of JMML cases.

The PTPN11 gene, which encodes the nonreceptor protein tyrosine phosphatase SHP-2, has recently been shown to play a crucial role in the pathogenesis of JMML: SHP-2 acts as an upstream regulator of RAS, and somatic PTPN11 mutations were found in 34% of JMML patients without Noonan syndrome.² Germline PTPN11 mutations are known to lead to Noonan syndrome, a multiple malformation syndrome characterized by unique facial features, neck webbing, and pulmonic stenosis. A subset of patients with Noonan syndrome also develop JMML.

Currently, the level of hematopoietic differentiation at which somatic PTPN11 mutations arise in JMML patients remains unknown. In the present study, we performed a PTPN11 mutation analysis of leukemic marrow cells, Epstein–Barr virus (EBV)-transformed B lymphocytes, IL-2-stimulated peripheral blood (PB) cells, and buccal smear cells from two JMML patients without Noonan syndrome to document possible hematopoietic lineage-dependent genomic alterations at the PTPN11 locus.

Both patients reported here had neither the features nor a family history of neurofibromatosis type 1 or Noonan syndrome. The study protocol was approved by the Ethics Committee of Keio University's School of Medicine and the National Cancer Center. After obtaining informed consent from the patients' parents, the patients' leukemic marrow cells were screened for PTPN11 mutations. Genomic DNA was isolated using a desalting column (Qiagen, Chattsworth, CA, USA). Exon 3 and exon 8 in PTPN11, known to be mutation hot spots,² were analyzed by direct sequencing of the PCR products amplified from the genomic DNA using a previously described method.³ To obtain T lymphocytes for the mutation analysis, PB cells were cultured and expanded using a TLY Culture kit25 (Lymphotec, Tokyo, Japan), which comprises plastic flasks coated with anti-CD3 monoclonal antibody and RPMI-1640 medium containing 10% fetal calf serum and IL-2, as described previously.⁴ Almost all the IL-2-stimulated PB cells cultured by this method have been shown to be CD3-positive lymphocytes.⁴ To prevent contamination of DNA derived from myeloid cells, high molecular weight DNA was isolated from the IL-2-stimulated PB cells after culture for 3 weeks according to standard protocols.

Patient 1: A 2-year-old boy presented with marked hepatosplenomegaly, skin lesions, thrombocytopenia (25.0 10⁹/l), and leukocytosis (26.8 10⁹/l) with myeloid precursors and a monocyte percentage of 16.5%. A bone marrow aspirate revealed granulocytic hyperplasia consisting of less than 5% blasts and a normal karyotype (46,XY). An elevated fetal hemoglobin level (50%) and a decreased neutrophil alkaline phosphatase (NAP) score of 29 were consistent with a diagnosis of JMML. Analysis of the genomic DNA derived from the leukemic bone marrow aspirate revealed an A>C transition in exon 3 at nucleotide 227, leading to an E76A substitution. The E76A substitution mutation was not present in an ethnically matched healthy control population of 100 subjects. Furthermore, the E76A substitution mutation was not found in the patient's EBV-transformed B lymphocytes (Table 1). The patient was treated with 6-mercaptopurine (6-MP). At 4 months after the diagnosis of JMML, the patient underwent allogeneic bone marrow transplantation (allo-BMT) from an HLA-identical sibling.

Table 1 - Molecular findings in hematopoietic and buccal cells from two patients with PTPN11 mutation.

Full table

Patient 2: A 4-month-old boy presented with progressive hepatosplenomegaly, thrombocytopenia (71.0 10⁹/l), and leukocytosis (49.8 10⁹/l) with myeloid precursors and a monocyte percentage of 25.5%. A bone marrow aspirate revealed marked granulocytic hyperplasia consisting of less than 5% blasts and a normal karyotype (46,XY). A mildly elevated fetal hemoglobin level (3.6%) and a decreased NAP score of 75 were consistent with a diagnosis of JMML. Analysis of the genomic DNA derived from the leukemic bone marrow aspirate obtained at the time of diagnosis revealed an A>T transition in exon 3 at nucleotide 227, leading to an E76V substitution; this mutation was not present in an ethnically matched healthy control population of 100 subjects. In this patient, the E76V substitution was also present in the EBV-transformed B lymphocytes and the IL-2-stimulated PB cells, but not in the buccal cells (Table 1). The patient was treated with 6-MP and maintained in a chronic phase. A cytogenetic analysis of a bone marrow aspirate obtained 14 months after diagnosis revealed the advent of an abnormal karyotype: 46,XY,der(7)t(3;7)(q21;q22)[18]/46,XY,del(7)(q22)[1]/46,XY[1]. However, the EBV-transformed lymphocytes established from the patient's PB at the same time showed a normal karyotype (46,XY). The patient's fetal hemoglobin level had increased to 32.4% at that time (Table 2). The patient underwent an allo-BMT from a 5/6 HLA-matched unrelated donor 18 months after diagnosis.

Table 2 - Cytogenetic findings and HbF level in patient 2 with the PTPN11 mutation.

Full table

We identified mis-sense mutations E76A and E76V in the PTPN11 gene in leukemic myelomonocytic marrow cells from two JMML patients who did not exhibit any features of Noonan syndrome. These mutations have been previously identified in JMML patients without Noonan syndrome and are considered pathogenic, thus lending further support to the notion that somatic PTPN11 mutations can cause JMML. The absence of the E76A change in the EBV-transformed B lymphocytes in patient 1 suggests that the mutation occurred at the level of a myeloid-committed precursor cell. In patient 2, however, the presence of the E76V change in both the EBV-transformed B lymphocytes and the IL-2-stimulated PB cells, but not in the buccal cell germ line, indicates that the PTPN11 mutation occurred at the level of a pluripotential hematopoietic stem cell before it committed to a myeloid or lymphoid lineage. Somatic mutations with a heterogeneous clonal origin have also been observed in JMML patients with RAS or NF1 mutations.^5,6 Thus, we believe that causative mutations in RAS, NF1, or PTPN11, all of which result in the deregulation of the RAS pathway, can occur in both primitive myeloid precursors and pluripotential hematopoietic stem cells.

Patient 2, who harbored the PTPN11 mutation in the EBV-transformed B lymphocytes, acquired 7q monosomy in the bone marrow cells 14 months after diagnosis. The normal karyotype in the EBV-transformed B lymphocytes, established at the same time, demonstrates the serial acquisition of two 'hits': the first hit being the PTPN11 somatic mutation in a pluripotential hematopoietic stem cell and the second one being 7q monosomy in a subclone with a myeloid origin. Furthermore, the concurrent increase in fetal hemoglobin suggests clonal evolution and proliferation of JMML cells that acquired 7q monosomy at the level of a primitive myeloid precursor common to a granulocyte–macrophage and erythroid lineage. In this case, not only 7q monosomy but also a partial trisomy of 3q may have also contributed to the clonal evolution of JMML cells, considering that 3q trisomy has been occasionally reported in JMML.^7,8

In conclusion, PTPN11 mutant clones can originate from either a pluripotential hematopoietic cell or a cell with a primitive myeloid lineage.