The NPM1 mutant defines AML irrespective of blast count.

NPM1-mutated acute myeloid leukemia (AML) accounts for about 30%–35% of newly diagnosed cases in adults (50%–60% of all AML with normal karyotype1) and it is characterized by distinctive clinico-pathological and molecular features.2 For this reason, NPM1-mutated AML is now recognized as a distinct entity in both the International Consensus Classification (ICC)3 and the fifth revision of the World Health Organization (WHO) classification of myeloid neoplasms.4 The NPM1 gene encodes for a multifunctional protein,2 which continuously shuttles between the nuclear and the cytoplasmic compartments, but mainly resides in the nucleoli. Notably, NPM1-mutated AML cells show an aberrant cytoplasmic localization of mutant NPM1(NPM1c).1, 2 This is caused by the loss of nucleolar localization signal (tryptophans at positions 288 and 290 or 290 only) and the acquisition of a strong nuclear export signal (NES) at the C-terminus that, by interacting with the nuclear exporter Exportin-1 (XPO1), leads to the cytoplasmic accumulation of the NPM1 mutant.2 Unlike AML with recurrent genetic abnormalities, such as t(8;21), inv(16), t(16;16), or PML/RAR alpha, ≥20% bone marrow (BM) blast cells were required to diagnose NPM1-mutated AML, according to the 2016 WHO classification of myeloid neoplasms. This blast threshold has been abolished in the fifth revision of WHO4 while the ICC still requires at least 10% blasts for the diagnosis of NPM1-mutated AML,3 similar to other AMLs with recurrent genetic abnormalities. This raises the question of how NPM1-mutated myeloid neoplasms cases with <10% blast cells should be diagnosed, an issue that currently remains controversial. NPM1-mutated myeloid neoplasms showing <20% BM blasts have been reported as myelodysplastic syndrome (MDS) or myelodysplastic/myeloproliferative neoplasms (MDS/MPN, especially chronic myelomonocytic leukemia—CMML).5, 6 The frequency of NPM1 mutations in these disorders is low. As an example, in a total of 1900 patients with newly diagnosed MDS or CMML, only 31 (1.6%) carried NPM1 mutations.5 Diagnosis of MDS in these cases is often supported by the finding of multilineage dysplasia that usually occurs in about 25% of NPM1-mutated AML.2 However, NPM1-mutated MDS patients resemble more typical NPM1-mutated AML than MDS with NPM1 wild-type gene (Table 1). In fact, these cases are more frequently younger and females show normal karyotype as well as CD34 negativity.5, 6 Similarly, NPM1-mutated CMML cases (especially those with higher NPM1 mutation allelic burden5, 6), as compared with those carrying wild-type NPM1, show more aggressive clinical course, higher BM monocyte content, higher degree of anemia, normal karyotype, and higher probability for AML progression. These findings suggest that CMML cases with the above characteristics should be investigated for the presence of NPM1 mutations and that they represent early stage NPM1-mutated AML with extreme monocyte differentiation, mimicking CMML. Next-generation sequencing of NPM1-mutated MDS or CMML cases also reveals that they resemble more NPM1-mutated AML than MDS or CMML with wild-type NPM1. In fact, they usually lack the typical mutations (ASXL1, BCOR, EZH2, RUNX1, SF3B1, SRSF2, STAG2, U2AF1, and ZRSR2)5, 6 that define MDS or AML with myelodysplasia-related mutation.3 Conversely, in addition to mutated NPM1, they carry other mutations that are usually responsible for development of typical NPM1-mutated AML (e.g., DNMT3A, TET2, and NRASG12D). This is consistent with the experimental observation that knock-in npm1 mice develop a myeloproliferative syndrome and progress to AML only when additional mutations occur.2 However, it appears that these cases show a frequency of FLT3 mutations5 lower than typical NPM1-mutated AML that justify the presentation, of at least a percentage of these cases, with pancytopenia. Moreover, we found that BM biopsies from NPM1-mutated MDS and CMML often show clusters of NPM1 cytoplasmic-positive cells, suggestive of early stage AML. They are usually represented by islands of leukemic proerythroblasts, isolated megakaryocytes, and clusters of small sized, weakly positive myeloid blasts, which is in keeping with the multilineage involvement often detected in NPM1-mutated AML7 (Figure 1). As observed in typical NPM1-mutated AML,1, 2 these cells are consistently CD34 negative. The latter finding is in keeping with the results of immunohistochemistry and gene expression profiling (GEP) of typical NPM1-mutated AML.2 Unfortunately, there are no studies on NPM1-mutated MDS or CMML cases to confirm that they share with NPM1-mutated AML the typical upregulation of HOX genes. Immunohistochemistry helps identifying clusters of NPM1 cytoplasmic-positive cells (Figure 1) that usually outnumber blastic cells detectable by morphological criteria alone, reflecting the difficulty to reproduce blast counts among different observers. As an example, morphology cannot distinguish between normal and leukemic proerythroblasts that instead can be appropriately counted as leukemic according to the expression of cytoplasmic NPM1 mutant. Thus, a BM biopsy should be considered in all patients carrying an NPM1 mutation and showing a low percentage of blasts in the BM aspirate. Similarly, to other investigators,5, 6 we observed NPM1-mutated MDS cases that had high NPM1 variant allelic frequencies (VAFs) at the time of diagnosis, despite proportionally much lower BM blast percentages. In our experience, these cases are also characterized by the presence of a low number of NPM1 cytoplasmic-positive cells. The latter finding may reflect differentiation of the leukemic clone with associated loss of the NPM1 mutant (Falini B, unpublished observation), which could be due to increased degradation, although experimental work is required to verify this hypothesis. Clinically, NPM1-mutated MDS respond better to chemotherapy than hypomethylating agents,5 again resembling more NPM1-mutated AML (which responds poorly to hypomethylating agents)8 than higher risk MDS where standard care are hypomethylating agents.9 Similarly, in another study,6 NPM1-mutated MDS and CMML treated with chemotherapy had significantly higher CR rates (90% vs. 28%), longer median progression free survival (not reached vs. 7.5 months), and overall survival (not reached vs. 16 months) than patients treated with hypomethylating agents. This is in keeping with the known sensitivity of NPM1-mutated AML to chemotherapy1, 2 and the favorable prognosis of this leukemia entity in the absence of co-occurring FLT3-ITD mutations.1, 2 Interestingly, as observed in NPM1-mutated AML, clearance of the NPM1 mutation following chemotherapy and its reappearance at molecular relapse is also observed in patients with NPM1-mutated MDS or CMML patients.5, 6 These and other evidences10 (Table 1) strongly support the concept that NPM1-mutated MDS or CMML cases represent NPM1-mutated AML diagnosed at an early stage and that the NPM1 mutant defines AML irrespective of blast count. Therefore, the patients with NPM1-mutated MDS and CMML, even with <10% blast cells, should be treated with intensive chemotherapy +/− allogeneic stem cell transplantation, as typical NPM1-mutated AML.