NPM1-rearranged AML: clinical features, molecular pathogenesis, and therapeutic perspectives Authors: Yuko Shimosato, Susumu Goyama Published: 2026 ABSTRACT Nucleophosmin 1 (NPM1) gene aberrations are among the most common genetic alterations in acute myeloid leukemia (AML). NPM1 mutations (NPM1c), detected in 30–40% of adult AML cases, are well-characterized and associated with distinct clinical and molecular features. In contrast, NPM1 rearrangements (NPM1-r), involving rare fusion partners such as MLF1, CCDC28A, HAUS1, and RARA, are much less common and have historically been poorly understood. However, recent studies have begun to shed light on the molecular mechanisms and clinical characteristics of NPM1-r AML, revealing overlapping features with NPM1c-AML. Both types share key leukemogenic mechanisms, including cytoplasmic mislocalization of NPM1, interaction with XPO1, and sustained HOX gene expression, which drive leukemic transformation. NPM1-r AML may exhibit unique clinical characteristics, including a higher prevalence in younger patients and potentially a poorer prognosis, although further validation in larger cohorts is needed. Therapeutic strategies targeting XPO1 and Menin, which have shown promise in NPM1c-AML, may also hold potential for NPM1-r AML. Continued research is essential to further elucidate the biology of this rare AML subtype and to establish optimized treatment strategies. Introduction Aberrations in the nucleophosmin (NPM1) gene represent one of the most common genomic lesions in acute myeloid leukemia (AML). NPM1 mutations are detected in approximately 5% of pediatric AML cases and in 30–40% of adult AML cases, with the frequency rising to 40–50% in adult AML with a normal karyotype [1]. These mutations are almost invariably heterozygous, with the majority consisting of a four-base pair insertion in exon 12. This insertion causes a frameshift that generates a novel nuclear export sequence. While wild-type NPM1 predominantly localizes to the nucleolus, mutant NPM1 proteins are redistributed to the cytoplasm and are therefore referred to as cytoplasmic NPM1 (NPM1c) [2]. NPM1c frequently co-occurs with mutations in FLT3 (particularly FLT3-ITD), DNMT3A, IDH1, and IDH2, suggesting cooperative interactions between NPM1c and these mutations in driving leukemogenesis. Conversely, NPM1 mutations are largely mutually exclusive with mutations in RUNX1, CEBPA, and recurrent fusion oncogenes such as RUNX1-ETO, CBFB-MYH11, and MLL rearrangements [3]. Clinically, prior to the widespread use of FLT3 inhibitors, NPM1c-AML was associated with a relatively favorable prognosis in the absence of FLT3-ITD or when FLT3-ITD was present at a low allelic ratio. However, cases with a high FLT3-ITD allelic ratio were classified as intermediate risk. Furthermore, NPM1c-AML harboring mutations in WT1 or co-mutations involving FLT3-ITD and DNMT3A is known to confer an adverse prognosis [4]. In addition to the commonly detected frameshift mutations in exon 12, studies have identified balanced chromosome rearrangements involving NPM1 gene located at 5q35.1. Those include t(3;5)(q25.3;q35.1)/NPM1::MLF1, t(5;6)(q35.1;q23.1)/NPM1::CCDC28A, t(5;18)(q35.1;q21.2)/NPM1::HAUS1, and t(5;17)(q35.1;q21.1)/NPM1::RARA, which lead to the expression of NPM1 fusion proteins in AML [5–10]. These NPM1 rearrangements are considerably rarer, and their pathogenic roles remain less well understood compared to those of NPM1c. However, a series of recent studies has begun to shed light on the diverse functions of NPM1 fusion proteins and the clinical features associated with NPM1-rearranged (NPM1-r) AML. In this review, we summarize the clinical characteristics of NPM1-r AML and, incorporating our own data [11], provide an overview of current concepts regarding how NPM1 fusions contribute to leukemogenesis. Structure of NPM1 and NPM1 fusion genes Human NPM1 is located on chromosome 5q35 and comprises 12 exons. The protein consists of an N-terminal core domain, an acidic region, a nuclear localization signal (NLS), and a disordered C-terminal region (Fig. 1). The highly conserved N-terminal core domain mediates oligomerization and protein–protein interactions. The acidic region contains multiple acidic residues, forming a negatively charged stretch that contributes to histone binding. The C-terminal portion contains clusters of basic, positively charged residues followed by aromatic residues, which facilitate nucleic acid binding [12]. Wild-type NPM1 carries out a broad range of cellular functions, including nucleolar organization via liquid–liquid phase separation, regulation of ribosome biogenesis and export, DNA repair, centrosome duplication, and responses to nucleolar stress [13–15]. Nevertheless, the full spectrum of NPM1’s physiological roles has yet to be defined.Fig. 1Structures of NPM1, NPM1c, and NPM1 fusion genes A Schematic representation of the primary structure of NPM1 and its mutants. In all AML-associated NPM1 mutants, the nucleolar localization signal (NoLS) is disrupted, leading to altered subcellular localization. NES, nuclear export signal; AS, acidic stretch; NLS, nuclear localization signal; NoLS, nucleolar localization signal. B Three-dimensional structure of NPM1 Structures of NPM1, NPM1c, and NPM1 fusion genes A Schematic representation of the primary structure of NPM1 and its mutants. In all AML-associated NPM1 mutants, the nucleolar localization signal (NoLS) is disrupted, leading to altered subcellular localization. NES, nuclear export signal; AS, acidic stretch; NLS, nuclear localization signal; NoLS, nucleolar localization signal. B Three-dimensional structure of NPM1 NPM1 possesses multiple signals that regulate nucleocytoplasmic trafficking, functioning as a shuttling chaperone protein. The major signals include the nuclear export signal (NES), nuclear localization signal (NLS), and nucleolar localization signal (NoLS). The NES is recognized by the nuclear export receptor exportin 1 (XPO1), also known as chromosomal region maintenance 1 (CRM1), and facilitates the export of proteins from the nucleus to the cytoplasm. In contrast, the NLS is recognized by importin-α/β and mediates nuclear import from the cytoplasm. The NoLS directs proteins to the nucleolus and, in NPM1, is located within exon 12 [16]. Exon 12 is a mutation hotspot, where a typical four-base pair insertion results in a frameshift that leads to the loss of the NoLS and the acquisition of a novel C-terminal NES. Importantly, a similar principle applies to NPM1 fusion proteins (Fig. 1). In fusions such as NPM1::MLF1, NPM1::CCDC28A, NPM1::HAUS1, and NPM1::RARA, the exon 12 region of NPM1 is absent, resulting in the loss of the NoLS. Furthermore, fusion partners such as CCDC28A and HAUS1 possess their own C-terminal NES motifs [7, 8]. When fused to NPM1, these NES motifs provide an additional NES, thereby producing a fusion protein that mimics NPM1c, allowing it to be exported to the cytoplasm. Clinical features of NPM1-rearranged AML To elucidate the clinical features and outcomes of NPM1-r AML, we reviewed cases reported in the literature. Among the identified rearrangements, NPM1::MLF1 is the most prevalent, accounting for approximately 0.2–0.5% of AML cases, including secondary AML following myelodysplastic syndromes [5]. A previous study detailed 69 cases of NPM1::MLF1 AML [17], the key findings of which are summarized in Table 1. Table 2 summarizes the additional cases we incorporated into this study, comprising 8 cases of NPM1::MLF1, 3 of NPM1::CCDC28A, 2 of NPM1::HAUS1, and 9 of NPM1::RARA AML. This comprehensive analysis highlights critical observations regarding blast characteristics, co-occurring mutations, age distribution, and prognosis. Table 1Clinicopathological characteristics and outcomes of reported AML and MDS cases harboring NPM1::MLF1 [17]CategoryFindingsTotal reported AML and MDS cases with NPM1::MLF169Average age38.9 yearsFAB classification (AML)M2: 18 cases; M4: 11 cases;M6: 12 cases; M5: 2casesAML with myelodysplasia-related changes (AML-MRC)20• MDS only12• MDS → AML8Cases with available sex information51Sex distributionMale predominance(Male:Female = 34:17)Cases with available outcome data45Deaths (among cases with outcome data)20Survival after HSCT4/7Of 69 reported AML and MDS cases harboring NPM1::MLF1, outcome and sex were documented in 45 and 51 patients, respectively. The cohort showed a male predominance (Male:Female = 34:17) and a mean age of 38.9 years. AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; HSCT, hematopoietic stem cell transplantationTable 2Patient characteristics of NPM1-rearranged AMLAgeSexWBC(μl)PB Blast (%)CD34 expression in BlastFABTreatmentSCTOutcomeOS(month)RelapseOther somaticmutationRefsNPM1-MLF112.9M14,40035.6N.ASecondary AML from MDSAML-05DoneAlive45NoNone (Target genes)[6]3F80,40030NegativeM2LAME 91NoDead2NoNone (Target genes)[5]45F43008NegativeM2ALFA-9802DoneAlive84YesWT1 (Target genes)[5]49F35,50055NegativeM2ALFA-9802NoDead18YesFLT3-ITD(Target genes)[5]34F24,90085NegativeM2ALFA-0702NoDead14YesFLT3-ITD,IDH2 (Target genes)[5]15F43009NegativeM2ELAM02NoAlive16NoNone (Target genes)[5]2M28,1004NegativeM4ELAM02DoneAlive14NoNone (Target genes)[5]22F17,00031NegativeM2ALFA-0702DoneDead7NoIDH2 (Target genes)[5]NPM1::CCDC28A1.4F16,6004.6N.AM2AML-05DoneDead(Infection)6NoFLT3-ITD(Target genes)[6]68M32,000N.ANegativeN.AICN.ALoss FUN.AYesFLT3-ITD, KRAS DNMT3A(Whole exome)[7]61F84,900N.ANegativeN.AActinomycin DN.ADead(Relapse)N.AYesIDH2(Whole exome)[7]NPM1::HAUS163FN.AN.ANegativeN.AN.AN.ADead(Infection)1NoFLT3-ITD,DNMT3A(Whole exome)[8]78MN.AN.ANegativeN.AN.AN.ADead(Relapse)18YesFLT3-ITD(Whole exome)[8]NPM1::RARA5M49,300N.ANegativeM3ATRA + ICNoAliveN.ANoN.A[9]2.5FN.AN.AN.AM3ATRA + ICDoneDead7YesN.A[10]12MN.AN.AN.AM3ATRA + ICDoneDead8YesN.A[10]9F17,000N.AN.AM3ATRA + ICNoAlive29NoN.A[10]12M15,700N.AN.AM3ATRANoDead (DIC)5 daysNoN.A[10]29M2900N.AN.AM3ATRA + ICNoDead22YesN.A[10]64M15,0000NegativeM3ATRA + ICNoAlive18NoN.A[10]4M25,5600N.AM3ATRA + ICNoAliveN.ANoN.A[10]51M42000NegativeM3ATRA + ICNoAlive2NoN.A[10]M, male; F, female; WBC, white blood cell count; SCT, stem cell transplantation; OS, overall survival; FAB, French–American–British (FAB) classification; IC, standard intensive chemotherapy; FU, Follow-Up Table 1Clinicopathological characteristics and outcomes of reported AML and MDS cases harboring NPM1::MLF1 [17]CategoryFindingsTotal reported AML and MDS cases with NPM1::MLF169Average age38.9 yearsFAB classification (AML)M2: 18 cases; M4: 11 cases;M6: 12 cases; M5: 2casesAML with myelodysplasia-related changes (AML-MRC)20• MDS only12• MDS → AML8Cases with available sex information51Sex distributionMale predominance(Male:Female = 34:17)Cases with available outcome data45Deaths (among cases with outcome data)20Survival after HSCT4/7Of 69 reported AML and MDS cases harboring NPM1::MLF1, outcome and sex were documented in 45 and 51 patients, respectively. The cohort showed a male predominance (Male:Female = 34:17) and a mean age of 38.9 years. AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; HSCT, hematopoietic stem cell transplantation Clinicopathological characteristics and outcomes of reported AML and MDS cases harboring NPM1::MLF1 [17] M2: 18 cases; M4: 11 cases; M6: 12 cases; M5: 2cases Male predominance (Male:Female = 34:17) Of 69 reported AML and MDS cases harboring NPM1::MLF1, outcome and sex were documented in 45 and 51 patients, respectively. The cohort showed a male predominance (Male:Female = 34:17) and a mean age of 38.9 years. AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; HSCT, hematopoietic stem cell transplantation Table 2Patient characteristics of NPM1-rearranged AMLAgeSexWBC(μl)PB Blast (%)CD34 expression in BlastFABTreatmentSCTOutcomeOS(month)RelapseOther somaticmutationRefsNPM1-MLF112.9M14,40035.6N.ASecondary AML from MDSAML-05DoneAlive45NoNone (Target genes)[6]3F80,40030NegativeM2LAME 91NoDead2NoNone (Target genes)[5]45F43008NegativeM2ALFA-9802DoneAlive84YesWT1 (Target genes)[5]49F35,50055NegativeM2ALFA-9802NoDead18YesFLT3-ITD(Target genes)[5]34F24,90085NegativeM2ALFA-0702NoDead14YesFLT3-ITD,IDH2 (Target genes)[5]15F43009NegativeM2ELAM02NoAlive16NoNone (Target genes)[5]2M28,1004NegativeM4ELAM02DoneAlive14NoNone (Target genes)[5]22F17,00031NegativeM2ALFA-0702DoneDead7NoIDH2 (Target genes)[5]NPM1::CCDC28A1.4F16,6004.6N.AM2AML-05DoneDead(Infection)6NoFLT3-ITD(Target genes)[6]68M32,000N.ANegativeN.AICN.ALoss FUN.AYesFLT3-ITD, KRAS DNMT3A(Whole exome)[7]61F84,900N.ANegativeN.AActinomycin DN.ADead(Relapse)N.AYesIDH2(Whole exome)[7]NPM1::HAUS163FN.AN.ANegativeN.AN.AN.ADead(Infection)1NoFLT3-ITD,DNMT3A(Whole exome)[8]78MN.AN.ANegativeN.AN.AN.ADead(Relapse)18YesFLT3-ITD(Whole exome)[8]NPM1::RARA5M49,300N.ANegativeM3ATRA + ICNoAliveN.ANoN.A[9]2.5FN.AN.AN.AM3ATRA + ICDoneDead7YesN.A[10]12MN.AN.AN.AM3ATRA + ICDoneDead8YesN.A[10]9F17,000N.AN.AM3ATRA + ICNoAlive29NoN.A[10]12M15,700N.AN.AM3ATRANoDead (DIC)5 daysNoN.A[10]29M2900N.AN.AM3ATRA + ICNoDead22YesN.A[10]64M15,0000NegativeM3ATRA + ICNoAlive18NoN.A[10]4M25,5600N.AM3ATRA + ICNoAliveN.ANoN.A[10]51M42000NegativeM3ATRA + ICNoAlive2NoN.A[10]M, male; F, female; WBC, white blood cell count; SCT, stem cell transplantation; OS, overall survival; FAB, French–American–British (FAB) classification; IC, standard intensive chemotherapy; FU, Follow-Up Patient characteristics of NPM1-rearranged AML FLT3-ITD (Target genes) FLT3-ITD, IDH2 (Target genes) NPM1:: CCDC28A Dead (Infection) FLT3-ITD (Target genes) FLT3-ITD, KRAS DNMT3A (Whole exome) Dead (Relapse) IDH2 (Whole exome) Dead (Infection) FLT3-ITD, DNMT3A (Whole exome) Dead (Relapse) FLT3-ITD (Whole exome) M, male; F, female; WBC, white blood cell count; SCT, stem cell transplantation; OS, overall survival; FAB, French–American–British (FAB) classification; IC, standard intensive chemotherapy; FU, Follow-Up NPM1::RARA was exclusively identified in acute promyelocytic leukemia (APL) treated with all-trans retinoic acid (ATRA)-based regimens [9, 10]. These findings suggest that RARA is the primary driver in this AML subtype, whereas the leukemogenic contribution of NPM1 remains undefined. Conversely, other NPM1-r AMLs exhibit phenotypic similarities to NPM1c-AML. Specifically, CD34 expression on blasts was consistently negative where data were available. Furthermore, co-occurring mutations such as FLT3-ITD, DNMT3A, WT1, and IDH2—common in NPM1c-AML—were frequently observed. The lack of CD34 expression and the mutational landscape closely mirror NPM1c-AML, implying shared leukemogenic mechanisms [18, 19]. Despite these similarities, NPM1-r AML presents distinct clinical features. Of the 22 cases analyzed (Table 2), 12 were pediatric or young adult patients, suggesting a predilection for younger individuals. Clinical outcomes indicate a poorer prognosis for NPM1-r AML compared to NPM1c-AML, characterized by high mortality (57.1%) and relapse (40.9%) rates. These results underscore the need for therapeutic strategies specifically tailored to NPM1-r AML. However, given the limited sample size, further large-scale studies are warranted to validate these trends. Molecular pathogenesis of NPM1-r AML (1) Leukemogenic activities of NPM1 fusion proteins (1) Leukemogenic activities of NPM1 fusion proteins (1) Leukemogenic activities of NPM1 fusion proteins Although the functions of NPM1 fusion proteins have remained largely unknown, recent studies have gradually unveiled their leukemogenic activities. For instance, NPM1::MLF1 alone is insufficient to enhance the colony-forming ability of normal mouse bone marrow cells. However, it can augment colony-forming activity in NPM1± bone marrow progenitors, accompanied by increased expression of the Hoxa9, Hoxa10, and Meis1 [20]. Furthermore, NPM1::MLF1-transduced mouse bone marrow cells occasionally induce AML in a mouse retroviral gene transduction and transplantation model [11]. These findings suggest that NPM1::MLF1 possesses weak leukemogenic activity, which requires additional cooperating events—such as reduced activity of wild-type NPM1 and other mutations—for full leukemic transformation. In contrast, NPM1::CCDC28A exhibits significantly stronger leukemogenic activity. NPM1::CCDC28A-transduced mouse bone marrow progenitors acquire extensive serial replating capacity in vitro and consistently induce AML in nearly all recipient mice [11]. Given that CCDC28A also forms the NUP98::CCDC28A fusion gene in T-cell leukemia [21], it is likely that the oncogenic activity of CCDC28A further enhances the leukemic potential of NPM1::CCDC28A. NPM1::RARA has been shown to function comparably to PML::RARA, a well-established driver of APL. Ectopic expression of NPM1::RARA in mouse bone marrow cells induces APL-like phenotypes both in vitro and in vivo. Moreover, transgenic mice expressing NPM1::RARA develop an APL-like disease that responds to ATRA treatment similarly to PML::RARA-positive APL [22, 23]. Thus, NPM1::RARA appears to resemble PML::RARA more closely than other NPM1 fusion proteins.(2) Intracellular behavior of NPM1 fusion proteins (2) Intracellular behavior of NPM1 fusion proteins (2) Intracellular behavior of NPM1 fusion proteins The formation of NPM1c or NPM1 rearrangements invariably results in the loss of the nucleolar localization signal (NoLS) and, in many cases, the acquisition of a novel C-terminal nuclear export signal (NES) (Fig. 1). The newly generated C-terminal NES is preferentially recognized by XPO1 over the N-terminal NES, promoting the cytoplasmic localization of mutant NPM1 [24]. The NPM1 fusion proteins—NPM1::MLF1, NPM1::CCDC28A, and NPM1::HAUS1—all lack the NoLS and have been reported to localize to the cytoplasm [7, 8]. Because CCDC28A and HAUS1 carry NESs, NPM1::CCDC28A and NPM1::HAUS1 acquire an additional NES, similar to NPM1c. By contrast, MLF1 does not contain a C-terminal NES; therefore, NPM1::MLF1 does not gain a new NES. Our findings show that NPM1::CCDC28A exhibited more prominent cytoplasmic localization than NPM1::MLF1 [11], consistent with the notion that the acquisition of a new NES enhances cytoplasmic trafficking. Moreover, given that NPM1::CCDC28A demonstrates stronger leukemogenic activity than NPM1::MLF1[11], this observation suggests that the degree of cytoplasmic mislocalization may be linked to leukemogenic potential. Thus, the aberrant redistribution of NPM1—normally restricted to the nucleolus—to the cytoplasm is a unifying feature of AML cells with both NPM1c and NPM1 rearrangements. Wild-type NPM1 is essential for cell survival, and its complete loss is embryonically lethal. Consequently, NPM1 mutations and rearrangements are always heterozygous. NPM1c is known to form heterodimers with the wild-type protein, escorting it to the cytoplasm and significantly impairing nuclear NPM1 function. In addition, NPM1c aberrantly sequesters nuclear proteins, such as ARF, HEXIM1, FBW7, MIZ1, CTCF, PU.1, and caspases-6/8, to the cytoplasm [14, 25, 26]. As a result, both the nuclear function of NPM1 itself and the nuclear regulatory networks associated with NPM1 are disrupted in NPM1c-AML, likely contributing to leukemogenesis. A similar disruption of NPM1 function and its associated nuclear proteins is likely to occur in NPM1-r AML, although this hypothesis requires experimental confirmation in future studies.(3) Chromatin regulation by NPM1 fusion proteins (3) Chromatin regulation by NPM1 fusion proteins (3) Chromatin regulation by NPM1 fusion proteins A distinctive feature of AML with NPM1 abnormalities is the sustained high-level expression of HOX genes. Under physiological conditions, HOX genes such as HOXA9 are highly expressed in hematopoietic stem cells but are downregulated during differentiation. However, in NPM1-mutated AML, HOX expression remains at levels comparable to those in normal stem cells, driving leukemic transformation. Genome-editing studies have demonstrated that the upregulation of HOX genes is strictly dependent on the presence of NPM1c [27]. Pharmacologic inhibition of XPO1, which blocks the export of NPM1c, reduces HOX expression and promotes differentiation of NPM1-mutated AML cells [28]. These findings establish a direct causal relationship between NPM1 mutation and HOX activation. Recent studies have proposed that a small fraction of mutant NPM1 retained in the nucleus directly binds to the promoter regions of HOX clusters and MEIS1, thereby regulating transcription [29, 30]. XPO1, traditionally recognized as a nuclear export factor, has been identified as a central mediator of this process. XPO1 can bind directly to chromatin, and chromatin association of mutant NPM1 requires XPO1. It has been proposed that XPO1 first associates with chromatin and subsequently recruits NPM1c. NPM1c, via its acidic domain (amino acids 120–150), forms a complex with KMT2A/MLL1 and RNA polymerase II on chromatin to activate target-gene transcription. More recent work has shown that NPM1c phase-separates to form nuclear condensates, termed the "coordinating body" (C-body), which recruits XPO1, NUP98, KMT2A, and MENIN. The C-body plays a crucial role in regulating gene expression and promoting leukemic proliferation [31]. Importantly, XPO1 inhibition disrupts the association of mutant NPM1 with chromatin, leading to downregulation of HOX and MEIS1 expression. This highlights the XPO1-mediated pathway as a promising therapeutic target. Interestingly, NPM1 fusion proteins appear to function similarly to NPM1c. Both NPM1::MLF1 and NPM1::CCDC28A upregulate HOX genes. NPM1::CCDC28A binds to HOX promoters and activates their transcription, and this effect is abolished by XPO1 inhibition [11]. These findings suggest that NPM1 fusion proteins, like NPM1c, cooperate with XPO1 to sustain abnormal HOX expression and drive leukemogenesis. In conclusion, sustained HOX expression in NPM1-mutated AML is likely dependent on the interaction between mutant NPM1 and XPO1. XPO1 is now recognized to have dual roles: mediating the nuclear export of mutant NPM1 and anchoring them to chromatin (Fig. 2). It remains to be determined which of these functions is more critical for leukemogenesis. The prevailing model suggests that NPM1c and NPM1 fusion proteins bind to HOX gene promoters in cooperation with XPO1, maintaining constitutive HOX expression and driving leukemia.Fig. 2NPM1:CCDC28A predominantly localizes in the cytoplasm and induces HOX gene expression through XPO1. NPM1::CCDC28A binds to XPO1 and predominantly localizes in the cytoplasm. A portion of NPM1:CCDC28A in the nucleus is recruited to the promoters of HOX gene clusters by XPO1, where it promotes HOX gene transcription. When XPO1 is inhibited by selinexor, NPM1::CCDC28A dissociates from chromatin, resulting in HOX gene downregulation and reduced colony formation NPM1:CCDC28A predominantly localizes in the cytoplasm and induces HOX gene expression through XPO1. NPM1::CCDC28A binds to XPO1 and predominantly localizes in the cytoplasm. A portion of NPM1:CCDC28A in the nucleus is recruited to the promoters of HOX gene clusters by XPO1, where it promotes HOX gene transcription. When XPO1 is inhibited by selinexor, NPM1::CCDC28A dissociates from chromatin, resulting in HOX gene downregulation and reduced colony formation Therapeutic strategies for NPM1-r AML (1) XPO1 inhibitors (1) XPO1 inhibitors (1) XPO1 inhibitors XPO1 is involved in both nuclear export and chromatin recruitment of oncogenic complexes, making it an important therapeutic target in AML with NPM1c and NPM1 rearrangements. In a mouse leukemia model harboring NPM1c and FLT3-ITD, treatment with an XPO1 inhibitor significantly prolonged survival [28]. Selinexor is a selective inhibitor of nuclear export that specifically targets XPO1/CRM1, a critical mediator of nuclear protein export, and chromatin recruitment of oncogenic complexes. In our study, XPO1 inhibition suppressed colony-forming capacity, inhibited HOX gene expression, and restored the nuclear localization of NPM1::CCDC28A in mouse AML cells transformed by NPM1::CCDC28A [11]. Clinical trials of selinexor in AML are currently ongoing, with its safety already confirmed in a phase I trial for relapsed/refractory AML. Beyond its activity as a single agent, selinexor is expected to exert synergistic effects when combined with conventional chemotherapy [32]. The second-generation XPO1 inhibitor, eltanexor, offers advantages over selinexor, including lower toxicity and limited penetration of the blood–brain barrier. Eltanexor provides more sustained XPO1 inhibition and has demonstrated efficacy in patient-derived xenograft models of NPM1c-AML [33]. However, it is important to note that XPO1 inhibitors are not specific to NPM1; they also affect the nuclear export of other proteins, such as TP53 and CDKN1A [16]. Therefore, careful interpretation is required to assess the extent to which their antileukemic effects are directly attributable to NPM1.(2) Menin inhibitors (2) Menin inhibitors (2) Menin inhibitors Menin inhibitors are emerging as promising agents for the treatment of NPM1-mutated AML. Menin functions as a cofactor of MLL1 (KMT2A), promoting H3K4 methylation and transcriptional activation of target genes such as HOXA9 and MEIS1. Although these agents were initially developed to target leukemia with MLL1 rearrangements, Menin inhibition is also expected to be effective against NPM1c-AML due to the cooperative role of NPM1c with MLL1 in enhancing HOX/MEIS transcription. Indeed, novel Menin inhibitors, such as MI-3454 and VTP-50469, have demonstrated potent antileukemic activity in murine models of NPM1c-AML [34]. The oral Menin inhibitor revumenib (SNDX-5613) selectively disrupts the Menin–MLL interaction and induced complete remission (CR) or CR with partial hematologic recovery in approximately 30% of relapsed AML patients, with an acceptable safety profile in a first-in-human phase I trial (NCT04065399) [35]. In our study, AML cells expressing NPM1::CCDC28A were highly sensitive to Menin inhibition [11], suggesting that Menin inhibitors may be effective not only in NPM1c-AML but also in NPM1-r AML. Furthermore, the combined use of Menin and XPO1 inhibitors has shown synergistic effects in murine models of NPM1c-AML [36]. Similarly, the combination therapy outperformed monotherapy in human cord blood cells and murine bone marrow cells expressing NPM1::CCDC28A [11]. These findings highlight the potential of combination therapy with XPO1 and Menin inhibitors for NPM1-r AML, warranting further investigation in future clinical trials. Conclusions and future perspectives The leukemogenic mechanisms of NPM1 fusion proteins, excluding NPM1::RARA, are increasingly recognized to closely resemble those of NPM1c. In particular, key features such as the cytoplasmic localization of NPM1 fusion proteins, their interaction with XPO1, involvement in chromatin-level regulation, and induction of aberrant HOX gene expression—all of which contribute to leukemogenesis—are shared by both NPM1c and NPM1-r AML. This molecular similarity suggests that therapeutic approaches effective for NPM1c-AML, such as XPO1 inhibitors and Menin inhibitors, may also be beneficial for NPM1-r AML. At the same time, differences among fusion partners must be carefully considered. The biological behavior and therapeutic vulnerabilities of each NPM1 fusion appear to vary. For instance, NPM1::RARA exhibits functions similar to other RARA-rearranged genes, such as PML::RARA, and should therefore be considered separately from other NPM1 rearrangements. NPM1::CCDC28A appears to have stronger leukemogenic potential than other NPM1 fusions, which may necessitate more intensive treatment strategies, such as stem cell transplantation. The generally poorer prognosis associated with NPM1-r AML likely reflects the additive oncogenic activities of these specific fusion partners, which amplify the core leukemogenic drive of the NPM1 fusion and culminate in a more aggressive clinical phenotype. Several critical questions remain unresolved. Why does XPO1 possess two seemingly opposing functions—nuclear export and chromatin binding—and which of these functions is more critical in NPM1c and NPM1-r AML? Additionally, why do only mutant or fusion NPM1 proteins gain access to chromatin? Elucidating the protein interaction networks of wild-type NPM1, mutant NPM1, and NPM1 fusion proteins in both the nucleus and cytoplasm may uncover novel signaling pathways and potential therapeutic targets. Continued advances in both basic and clinical research are essential, and hold promise for improving outcomes in NPM1-r AML.