Nucleophosmin mutations in childhood acute myelogenous leukemia with normal karyotype

Childhood acute myelogenous leukemia (AML) is a clinically and molecularly heterogeneous disease.^1,2  The identification of recurrent chromosomal abnormalities allows different prognostic subgroups to be defined.¹  Unfortunately, this is not yet feasible in a large proportion of cases (20%-25%) in which no chromosomal abnormalities are visible by conventional karyotyping and the underlying genetic lesion is still unknown.

Progress has been recently made in the molecular characterization of adult AML with normal karyotype. Falini et al³  reported that nucleophosmin 1 (NPM1), a nucleus-cytoplasm shuttling protein^4-7  involved in rearrangements in leukemias and lymphomas,^8-10  showed mutations at its C-terminal region, causing an aberrant cytoplasmic expression in the leukemic cells of about 35% of primary adult AML. NPM1 is a multifunctional protein that prevents protein aggregation in the nucleolus and regulates the assembly and transport of preribosomal particles through the nuclear membrane.⁴  Since NPM1 is a multifunctional protein involved in the regulation of the ARF-p53 pathway,^11-14  it is likely that the mutation and/or ectopic location of the protein may play a leukemogenic role.^3,15 

This finding prompted us to investigate the prevalence of NPM1 mutations in a large group of childhood AML patients and to correlate this finding with the major biologic and clinical features.

From September 2002 to December 2004, 111 childhood patients (0-18 years of age) with primary AML (other than French-American-British [FAB] classification M3) were enrolled in the ongoing AML protocol of the Associazione Italiana di Ematologia e Oncologia Pediatrica (AIEOP-AML02). Of them, 107 patients were successfully analyzed for the NPM1 mutations. Leukemia samples were obtained from bone marrow (BM) aspirates at diagnosis. The karyotype and the FAB subtypes were available for 96 (89.7%) of 107 patients. The main biologic and clinical features of the patients are indicated in Table 1.

DNA from one NPM1-mutated patient at diagnosis was also investigated at remission. Informed consent has been obtained at each participating center. This study was approved by the AIEOP-AML Scientific Committee.

Cytogenetic investigations were performed by standard procedures. Reverse-transcriptase-polymerase chain reaction (RT-PCR) analysis for promyelocytic leukemia-retinoic acid receptor alpha (PML-RARalpha), acute myelogenous leukemia 1-eight twenty-one (AML1-ETO), and core binding factor beta-myosin heavy chain 11 (CBFB-MYH11); analysis of MLL gene status by fluorescence in situ hybridization (FISH); and mutational analysis of the FLT3 gene (internal tandem duplication [ITD] and 835-836 amino acidic residues) were performed as previously described.^16-19 

Genomic DNA was extracted from BM mononuclear cells at diagnosis by standard methods. The exon 12 of the NPM1 gene was amplified from genomic DNA, using a forward primer (NPM1-F) in intron 11 and a reverse primer (NPM1-R) in the 3′ gene flanking region, as previously described.³  Purified PCR products were directly sequenced using primer NPM1_1112R.³ 

RT-PCR and direct sequencing of the NPM1 coding region revealed mutations affecting exon 12 in 7 (6.5%) of 107 childhood AML cases (Table 2). Six sequence variants were observed among our cases. Four cases corresponded to nucleotide deletions and insertion observed in adult AML (1 type A, 1 type B, and 2 type D)³ ; in one additional case, the protein sequence was identical to type E of adult series, with a different nucleotide sequence; 2 new variants were observed (here provisionally called mutations G and H). As for the adult cases, all mutations consisted of either the insertion of 4 nucleotides at position 960 (types A to D) or, alternatively, the deletion of 5 nucleotides (positions 965-969) and the insertion at the same position of 9 new nucleotides (types E to H). Independent of the types, all mutations caused a frameshift in the region encoding the C-terminal of the NPM protein, resulting in the replacement of the last 7 amino acids (WQWRKSL) with 11 different residues. As for the adult cases, all NPM mutant proteins showed mutations in at least one of the tryptophan residues at positions 288 and 290, and shared the same last 5 amino acid residues (VSLRK). Thus, despite the heterogeneity at the DNA level, all NPM1 gene mutations resulted in the same sequence at the NPM protein C-terminus.

The mutations were heterozygous and were related only to the leukemic clone, since they were not present in a BM specimen tested at the time of complete remission (not shown).

Mutations of NPM1 were found in different FAB subtypes: M1, 1 of 21 cases; M2, 2 of 19 cases; M4 3 of 13 cases; and M6, 1 of 3 cases (Tables 1-2). This distribution over the FAB subtypes is similar to that observed in adult AML, with the exception of the M5 subtype, which was never found mutated in 18 pediatric cases. As in the adult series, NPM1- mutated patients were CD34^- at diagnosis.

There was no significant difference between NPM1-mutated and wild-type (wt) patients at presentation in terms of sex and white blood cell (WBC) count.

RT-PCR data on AML1-ETO and CBFB-MYH11 fusion genes, FISH data for MLL gene status, and standard cytogenetic data were available for 96 (89.7%) of 107 patients. Consistent with the report on adult AML,³  we found that all of the NPM1-mutated patients carried a normal karyotype, although they account for a lower number of childhood AML with normal karyotype (ie, 7 [26.9%] of 26 cases versus 60% in adults).³ 

Although the median ages of NPM1-mutated and wt patients were not significantly different (10.1 versus 7.6 years, P = .134), there is a clear gradient indicating a tendency to have a higher probability of NPM1 mutations for older AML children. This observation is in agreement with what was observed in adult AML, where the percentage of NPM1 mutations increased from 21.6% in the 15-to-30 age group to 44.1% in the 51-to-60 age group (B.F., personal oral communication, 2005). Moreover, 5 (35.7%) of 14 childhood AML patients older than 10 years and with normal karyotype were NPM1 mutated. This age-dependent distribution was already observed for other genetic abnormalities,²⁰  and together with the association with the absence of visible abnormalities may reflect a specific pathogenesis.

Overall, 13 (13.5%) of 96 cases carried a mutation in the FLT3 gene (11 ITDs and 2 mutations at residue 835). Of interest, one patient carried both FMS-like tyrosine kinase 3(FLT3)-ITD and NPM1 gene mutations, in a normal karyotype.

Although all mutated patients achieved complete remission, the time of observation is too short to derive a significant conclusion about the prognostic value of NPM1 mutation in childhood AML. None of the patients relapsed, 6 underwent bone marrow transplantation (BMT), and 5 of 7 are alive at last follow-up (Table 2). Two patients died of a BMT-related event.

In conclusion, we demonstrated that the NPM1 gene mutations are frequent in childhood AML patients with normal karyotype and older age, although the prevalence is lower with respect to adult AML.³  The children and adults with NPM1 mutations also share morphologic, phenotypical, and clinical features, such as wide morphologic spectrum, lack of CD34 expression, and a good response to induction chemotherapy.³  Our findings confirm in the childhood AML setting the importance of deregulated NPM1 in tumorigenesis and may have diagnostic and clinical relevance.²¹ 

The consistent deletion/insertion feature of the NPM1 mutations reconstitutes a patient- and leukemia-specific NPM1 sequence, which, similar to FLT3-ITD,²²  can be considered as a clonal marker for patients with normal karyotype, in which no alternative molecular markers are available. Thus, immunohistochemistry and mutational analysis of NPM1 may now enter into the routine diagnostic of AML, in order to identify cases in which minimal residual disease can potentially be monitored during therapy, to drive future interventions.

We are grateful to the work of Anna Leszl (cytogenetic analyses), Emanuela Frascella (RT-PCR analyses), Francesca Predieri and Monica Spinelli (samples and data collection), and the physicians from all AIEOP centers treating the children with AML included in the study.