Acquisition of loss of the wild-type NRAS locus with aggressive disease progression in a patient with juvenile myelomonocytic leukemia and a heterozygous NRAS mutation

Abstract

In a patient with juvenile myelomonocytic leukemia, a NRAS mutation at codon 12 (GGT>TGT) was initially heterozygous, but became homozygous after blastic crisis (BC). According to microsatellite and FISH analyses, the post-BC homozygous mutation might result from the loss of the wild-type NRAS locus through mitotic recombination.

Somatic point mutations of the RAS genes are found in approximately 25% of patients with juvenile myelomonocytic leukemia (JMML).1 Although at present allogeneic hematopoietic stem cell transplantation is the only curative treatment for JMML, high relapse rate is a serious problem. Among variables analyzed, transformation to acute leukemia, often referred to as blastic crisis (BC), at the time of transplantation was associated with increased cumulative incidence of post-transplant relapse.2 The genetic mechanism underlying BC, however, remains to be determined.

A 2-month old boy was diagnosed as having JMML and treated with 6-mercaptopurine. After 2 years, the laboratory findings revealed a marked leukocytosis, thrombocytopenia, and 43% blasts in bone marrow. The blasts were positive for myeloperoxidase, alpha-naphthyl butylate esterase, and CD13/CD33, indicating progression toward myeloid BC. Three months later, chromosomal analysis of the bone marrow cells showed the advent of monosomy 7. The details of the clinical course have been previously reported.3

Figure 1.Mutation analyses of the NRAS gene (A). The direct sequencing traces showing codon 12 of the NRAS gene in PB cells before and after blastic crisis (BC) of the patient. The heterozygous mis-sense mutation was observed before BC, whereas the homozygous mutation was observed after BC. (B) Data show PCR products of the mutated and/or wild-type NRAS alleles obtained from PB-MNCs of 4 normal controls, 6 pre-BC GM colony-constituent cells and 6 post-BC GM colony-constituent cells of the present case with codon 12 GGT→TGT mutation, and 6 GM colony-constituent cells of another case with codon 12 GGT→AGT mutation, using the MASA method. This latter case was used to show the reliability of our mutant allele-specific amplification technique. Primers 5′-CTGGTGGTG-GTTGGAGCAT-3′ and NRAS-R (5′-TCCGACAAGTGAGAGACAGG-3′), were used to detect the mutant NRAS sequence (GGT→TGT), whereas the primers 5′-TGGTGGTG-GTTGGAGCAG-3′ and NRAS-R, were used to detect the wild-type NRAS sequence. For the other mutant NRAS sequence (GGT→AGT), we used the primers 5′-CTGGTGGTG-GTTGGAGCAA-3′ and NRAS-R. GGT, wild-type allele; TGT, GGT→TGT mutant allele; AGT, GGT→AGT mutant allele.

A single nucleotide substitution at codon 12 of the NRAS gene from GGT to TGT was detected with the concomitant normal sequence in peripheral blood mononuclear cells (PB MNCs) obtained before treatment with 6-mercaptopurine, as shown in Figure 1A. There were no mutations at codons 12, 13, and 61 of the KRAS2 gene or at exons 3, 8, and 13 of the PTPN11 gene. However, PB MNCs obtained 5 months after BC showed the homozygous NRAS mutation (GGT>TGT). Mutant allele-specific amplification (MASA) analysis was then performed on PB-derived granulocyte-macrophage (GM) colony-constituent cells generated with GM colony-stimulating factor (CSF) and stem cell factor (SCF) in a methylcellulose culture.4 As shown in Figure 1B, both the mutant PCR product and the wild-type product were detected in all of 11 GM colony-constituent cells before BC. By contrast, only the mutant product was amplified from all of 24 GM colony-constituent cells after BC. Thus, the patient’s malignant myeloid lineage cells might initially possess the heterozygous NRAS mutation and might subsequently acquire the homozygous NRAS mutation. Using microsatellite analyses, we attempted to verify whether the homozygous NRAS mutation after BC was due to loss of heterozygosity (LOH) related to the NRAS gene according to the procedure previously described.5 Since there were no polymorphic microsatellite markers inside the NRAS gene, we used three microsatellite markers (D1S250, D1S2687 and D1S189) near the gene on the short arm of chromosome 1p13 (Figure 2A). D1S468, 112 Mb downstream from the NRAS gene, was used as a control. We made 9 pairs of pre-BC and post-BC GM colony samples whose sequences were all determined by the MASA method, and then compared numbers of the dinucleotide repeats in the marker DNAs. As shown in Figure 2B, the ratio of P1:P2 in post-BC/P1:P2 in pre-BC of D1S250 was higher than 1.35 (27.25±1.57, mean±SD). Ratios of post-P1:P2/pre-P1:P2 in D1S2687 and D1S189 were less than 0.67 (0.26±0.05 and 0.37±0.04, respectively). By contrast, the ratio of post-P1:P2/pre-P1:P2 in D1S468 was 0.85±0.05. Thus, these results suggest that LOH related to the NRAS locus joined the heterozygous NRAS mutation after BC. Finally, we examined the mechanism of the LOH. FISH analysis (using probes for the region containing the NRAS gene on chromosome 1p13 and for the region containing D1S468 on chromosome 1p36)6,7 and MASA assay were simultaneously performed on post-BC GM colony-constituent cells generated with GM-CSF and SCF. Thirteen out of 30 GM colonies with the homozygous mutation contained 1–2 metaphase cells per colony. In each of the 23 metaphase cells examined, the signals for RP5-1000E10 encompassing the NRAS gene were found on both chromosome 1 homologs (Figure 2C). These results suggest that the LOH resulted from duplication of the mutated allele through mitotic recombination, rather than from deletion of the NRAS part of wild-type chromosome 1p.

Figure 2.LOH analyses of the NRAS gene. (A) Chromosome 1p showing the markers used for microsatellite PCR analysis and FISH. (B) Analyses of polymorphic microsatellite marker DNAs on the short arm of chromosome 1 We made 9 pairs of pre-blastic crisis (BC) and post-BC GM colony samples (sequences were all determined by the MASA method), and then compared numbers of the dinucleotide repeats in 4 marker DNAs: D1S250, 40kb downstream from the NRAS gene; D1S2687 and D1S189, 1 Mb and 1.5 Mb upstream to the gene, respectively; D1S468, 112 Mb downstream from the NRAS gene. The forward primers were labeled with a fluorescence dye marker (5-6-carboxyfluorescein). Alleles were defined as the two highest peaks (P1 and P2) within the expected size range. Numbers within the rectangle indicate the size of the short tandem repeat (STR), and numbers below the rectangles indicate the height of the STR peaks. According to the maufacurer’s instructions, when the ratio of P1:P2 in the post-BC sample/P1:P2 in the pre-BC sample was either less than 0.67 or greater than 1.35, a LOH was confirmed.

Braun et al.8 reported that somatic activation of a latent KRAS^G12D allele rapidly induces a fatal myeloproliferative disorder in mice. However, our recent study revealed hematologic improvement despite no chemotherapy during a 2- to 4-year follow-up in JMML patients with a NRAS or KRAS2 glycine to serine substitution,9 implying that all types of RAS mutations are not always sufficient to confer the disease progression. Given the results of a 7,12-dimethylbenz[a]anthracene-induced rat leukemia model,10 the two stages of genetic change in our case suggest that oncogenic heterozygous RAS mutations are an early event of leukemogenesis, but in JMML the aggressive development requires other genetic lesions.

Footnotes

• Funding: this work was supported by Grants-in-Aid No. 11390300 from the Ministry of Education of Japan.