Nucleoporin 98 Rearrangements in Acute Leukemia: A Genomic Landscape Study Authors: Osama Batayneh, Natalie Danziger, Mahmoudreza Moein, Nour S Naji, Dean Pavlick, Jeffrey S Ross, Chelsea Marcus, Caleb Ho, Teresa Gentile, Zheng Zhou, Lincoln W Pasquina Published: 2026 ABSTRACT ABSTRACTNucleoporin 98 rearrangements (NUP98re) occur in a wide range of hematologic malignancies including acute leukemias with a variety of fusion partners. NUP98re is associated with adverse prognosis, especially in children. We aimed to better understand the genomics of acute leukemias with NUP98re including fusion partners and co‐occurring genomic alterations (GA). Results from 5905 patients with acute leukemia undergoing standard‐of‐care next generation sequencing on FoundationOneHeme were included for analysis. A total of 78 (1.3%) patient samples harbored NUP98re with a median age of 19 years compared with 62 years for the cases with no NUP98 rearrangement (NUP98wt) (p < 0.001). Patients with NUP98re were more frequently of admixed American ancestry (p < 0.001). Among patients with acute myeloid leukemia, individual genomic alterations more frequently identified in NUP98re cases included WT1 (77% vs. 11%, p < 0.001) and FLT3 (49% vs. 26%, p < 0.001); Alterations more frequent in the NUP98wt cohort included NPM1, KMT2A, TET2, DNMT3A, ASXL1, SRSF2, STAG2, and BCOR (all p < 0.05). NUP98re are rare in patients with acute leukemia, more frequent in pediatrics and younger adults but distributed across all age groups. NUP98re is associated with a unique genomic landscape featuring increased frequencies of mutations with proliferative features (FLT3 and KRAS) and growth suppression (WT1). Alterations in WT1 had the highest co‐occurrence rate in samples containing NUP98re. This work highlights the unique GA associated with NUP98 and emphasizes the need for clinical studies to reveal underlying biologic mechanisms and enhance optimal management in the presence of potentially targetable alterations such as NUP98re.Keywords: acute myeloid leukemia, genomic alterations, genomic study, nucleoporin, outcomes Abbreviations Acute lymphoblastic leukemia Acute Myeloid Leukemia Acute promyelocytic leukemia Genomic alterations Institutional Review Board Nucleoporin 98 rearrangements Introduction Nucleoporin 98 rearrangements (NUP98re) occur in a wide range of acute leukemias and have a potential for malignant transformation [1]. NUP98re are associated with specific disease phenotype with a unique cellular hierarchy [2]. The prognosis of patients with NUP98re is poor, especially in children [3]. More than 30 fusion partners with NUP98 have been identified across hematologic malignancies, including acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML). NUP98re are most frequently found in pediatric patients, occurring in approximately 5% of pediatric cases [4, 5, 6, 7]. NUP98re is a proposed predictive biomarker for response to menin inhibitor therapy in acute leukemia, under investigation in multiple prospective studies [8, 9]. Many potentially targetable alterations often co‐occur in acute leukemias, such as FLT3, IDH1/2, and/or NPM1 [10]. A therapeutic strategy of combining menin inhibitors and kinase inhibitors has been proposed for leukemia with NUP98re [11]. However, the prevalence of co‐occurring alterations in the context of NUP98re is not yet well described, perhaps due to the rarity of NUP98re. In this study we sought to describe an expanded range of NUP98 fusion partners across acute leukemias, as well as co‐occurring genomic alterations found in the context of NUP98re compared to NUP98wt acute myeloid leukemia. This comparison seeks to enhance our understanding of the molecular profile of AML and to identify potential therapeutic strategies. Methods and Materials Study Design Results were included from all patients with a diagnosis of acute leukemia, excluding acute promyelocytic leukemia (APL), undergoing next generation sequencing (NGS) by the laboratory developed test FoundationOneHeme (Foundation Medicine Inc., Boston, MA). All research was performed in accordance with a research protocol approved by an Institutional Review Board (WCG #20152817). Patients were assigned to one of the following age groups based on their age at the time of genomic profiling: pediatric, 0–14 years old; adolescent and young adult (AYA), 15–39 years old; or adult, ≥ 40 years old. Based on the genomic profiling results, study participants were divided into two cohorts: NUP98re cases and those with no rearrangements detected in NUP98 (NUP98wt). Genomic Profiling The FoundationOne Heme assay employs hybrid capture–based sequencing of the DNA coding sequence for 406 genes, including noncoding sequence for 31 genes, and RNA coding sequence for 265 genes to report base substitutions, insertions/deletions, copy number alterations, and rearrangements [12]. For NUP98 specifically, the exonic regions were evaluated by both DNA and RNA with overlap across splice sites into neighboring intronic regions [13]. Accepted sample types included peripheral whole blood, fresh bone marrow aspirate, and formalin‐fixed paraffin‐embedded (FFPE) blocks. Genetic ancestry for each patient was derived from the NGS assay using a single nucleotide polymorphism (SNP)‐based approach for research use only as previously described [14, 15]. In brief, a random forest classifier was trained using over > 40,000 germline SNPs with cross‐validation performed using the 1000 Genomes Project cohort [16]. Predominant genetic ancestry was classified into the five following groups: European (EUR), African (AFR), East Asian (EAS), South Asian (SAS), and admixed American (AMR). AMR ancestry has high overlap with Hispanic ethnicity [14]. Statistical Analyses Univariable analyses of continuous variables were conducted using the Wilcoxon rank sum test, while categorical variables were analyzed using the chi‐square or Fisher's exact test, as appropriate. Categorical variables are presented as frequencies and percentages, and continuous variables as medians and interquartile ranges (IQRs). For the comparison of the frequency of co‐occurring alterations in AML, Fisher's exact test was adjusted using the Benjamini–Hochberg procedure. A p value < 0.05 was considered statistically significant. Statistical analysis was performed using RStudio version 4.5.3. Results Clinical Characteristics—Overall Cohort A total of 5905 cases of acute leukemia underwent NGS between December 2013 and August 2025. 2844 samples (48.2%) were submitted as AML, 2284 (38.6%) as ALL, and the remaining 777 submitted as mixed phenotype or unspecified acute leukemia. 78 (1.3%) patients had samples harboring a rearrangement in NUP98. These patients had a median age of 19 years (IQR: 8–48) with 30 (36.1%) being pediatric, 22 (28.9%) AYA, 24 (32.5%) adult, and 2 with unknown age. 36 (46.2%) were female and 42 (53.8%) were male. NUP98re were detected in 69/2844 (2.4%) of patients with AML, 0/988 (0%) B‐ALL, 1/291 (0.3%) T‐ALL, 1/1005 (0.1%) ALL NOS, 6/730 (0.8%) acute leukemia NOS, and 1/47 (2.2%) of patients with mixed phenotype acute leukemia. NUP98 Rearrangements Of the 78 NUP98re, 2 (2.6%) were detected on DNA only, 7 (9.0%) were detected on both DNA and RNA, and 69 (88.5%) were detected on RNA only. 16 different rearrangement partner genes were identified in the study (Table 1). The most common partners were NSD1 (n = 40, 51.3%), KDM5A (also known as JARID1A, 15, 19.2%), DDX10 (4, 5.0%), HOXA9 (3, 3.8%), and PRRX2 (3, 3.8%). 9 (12%) rearrangements were with genes in the homeodomain (HD) protein family including class I homeotic genes (HOXA9, HOXD13, HOXB13, and HOXC13) and class II homeobox genes (PRRX2). Patients harboring those 9 rearrangements all had a diagnosis of AML. Of the non‐HD partner genes, several possessed recurrently described features associated with NUP98 partners, including SET domains (NSD1), PHD domains (KDM5A, PHF23, and TAF3), and coiled coil motifs (DDX10, RAP1GDS1, and SMCO4). TABLE 1Distribution of NUP98 rearrangements by subtype of acute leukemia, age within the AML subset, and partner gene.Partner genePartner gene locationAcute leukemia NOS/mixed phenotype (N = 7)All (N = 2)AMLTotal (N = 78)Pediatric (< 15 yrs) (N = 26)AYA (15 to < 40 yrs) (N = 21)Adult (≥ 40 yrs) (N = 20)Unknown (N = 2) NSD1 5q35.34 (57.1%)—11 (42.0%)16 (76.2%)7 (35.0%)2 (100.0%)40 (51.3%) KDM5A 12p13.332 (28.6%)1 (50.0%)8 (30.8%)3 (14.3%)1 (5.0%)—15 (19.2%) DDX10 11q22.3——2 (7.7%)1 (4.8%)1 (5.0%)—4 (5.0%) HOXA9 7p15.2——1 (3.8%)—2 (10.0%)—3 (3.8%) PRRX2 9q34.11————3 (15.0%)—3 (3.8%) TNRC18 7p22.1————2 (10.0%)—2 (2.6%) HOXD13 2q31.1——1 (3.8%)———1 (1.3%) HOXB13 17q21.32————1 (5.0%)—1 (1.3%) HOXC13 12q13.13————1 (5.0%)—1 (1.3%) MAML2 11q21————1 (5.0%)—1 (1.3%) PHF23 17p13.1——1 (3.8%)———1 (1.3%) RAP1GDS1 4q23—1 (50.0%)————1 (1.3%) SMCO4 11q211 (14.3%)—————1 (1.3%) TAF3 10p14——1 (3.8%)———1 (1.3%) TSPEAR 21q22.3————1 (5.0%)—1 (1.3%) ZNF711 Xq21.1——1 (3.8%)———1 (1.3%)Unknown————1 (4.8%)——1 (1.3%) Distribution of NUP98 rearrangements by subtype of acute leukemia, age within the AML subset, and partner gene. AML Subset—Clinical Characteristics Due to the majority of NUP98re being found in AML, and the differences between types of acute leukemias, further analysis was performed on patients with AML only. Within the subset of patients with AML (N = 2844), we compared the clinical characteristics of NUP98re and NUP98wt (Table 2). Patients with NUP98re had younger median age (19 vs. 62 years, p < 0.001) and many more patients in the pediatric and AYA age groups (37.7% vs. 13.8% and 30.4% vs. 13.0% respectively) than NUP98wt patients. Overall, among patients with AML, NUP98re were found in 6.3% of pediatric patients, 5.5% of AYA patients, and 1.0% of adults (Figure 1). The distribution of sexes was similar regardless of NUP98re status (43.5% vs. 44.1% female). Patients with NUP98re were enriched for admixed American ancestry compared to the NUP98wt cohort (21.7% vs. 10.7%). TABLE 2Clinical characteristics of the AML cohort by NUP98re status. p values were not corrected for multiple comparisons. NUP98re (N = 69) NUP98wt (N = 2775)Total (N = 2844) p valuePatient age< 0.001Median (Q1, Q3)19 (9, 51)62 (35, 74)62 (33, 74)Unknown22830Patient age group< 0.001Pediatric (< 15 yrs)26 (37.7%)384 (13.8%)410 (14.4%)Young adult (15 to < 40 yrs)21 (30.4%)361 (13.0%)382 (13.4%)Older adult (≥ 40 yrs)20 (29.0%)2002 (72.1%)2022 (71.1%)Unknown2 (2.9%)28 (1.0%)30 (1.1%)Sex0.912Female30 (43.5%)1225 (44.1%)1255 (44.1%)Male39 (56.5%)1550 (55.9%)1589 (55.9%)Predominant genetic ancestry< 0.001African6 (8.7%)319 (11.5%)325 (11.4%)Admixed American15 (21.7%)296 (10.7%)311 (10.9%)East Asian2 (2.9%)95 (3.4%)97 (3.4%)European36 (52.2%)1931 (69.6%)1967 (69.2%)South Asian2 (2.9%)51 (1.8%)53 (1.9%)Could not be determined8 (11.6%)83 (3.0%)91 (3.2%) Clinical characteristics of the AML cohort by NUP98re status. p values were not corrected for multiple comparisons. Prevalence of NUP98 rearrangements by age group within the AML cohort. AML Subset—Genomic Characteristics When comparing the genomic profiles of NUP98re versus NUP98wt AML, several differences in the frequency of co‐occurring gene alterations were identified (Figure 2). NUP98re samples were enriched compared to NUP98wt in alterations in FLT3 (49% vs. 26%, p < 0.001), WT1 (77% vs. 11%, p < 0.001), KRAS (20% vs. 10%, p = 0.04), and NOTCH1 (6% vs. 1%, p = 0.03). NUP98re samples less frequently harbored alterations in TET2 (4% vs. 23%, p < 0.001), DNMT3A (6% vs. 18%, p = 0.03), ASXL1 (3% vs. 17%, p = 0.004), NPM1 (0% vs. 16%, p < 0.001), KMT2A (0% vs. 14%, p < 0.001), SRSF2 (3% vs. 13%, p = 0.03), STAG2 (0% vs. 7%, p = 0.03), and BCOR (0% vs. 7%, p = 0.04). Table S1 in Supporting Information S1 appendix demonstrates genomic alterations frequencies comparison between NUP98re and NUP98wt. Comparison of co‐occurring alteration frequency by NUP98re status within the AML cohort. p values were calculated by fisher's exact test and corrected for multiple comparisons by the benjamini‐hochberg method. Genomic Alterations Comparison Amongst Pediatric, AYA, Adult Trends in the overall cohort were driven mostly by the adult population which represented ∼70% of the total AML patients. Across all age groups, WT1 remained strongly associated with NUP98re (88.4% in pediatric, 90% in AYA, 40% adult). Association with FLT3 and NUP98re in overall population was mostly driven by AYA and adult groups and was not seen in the pediatric population. Supporting Information S1: Figure S1 demonstrates comparison of co‐occurring alteration frequencies by NUP98re status within pediatric (Figure 1‐A) AYA (Figure 1‐B) and adult (Figure 1‐C). Discussion Acute leukemias encompass a wide diversity of genomic abnormalities [17]. Many key genes have been identified as essential drivers for leukemogenesis and significant prognostic markers for classification [18, 19]. NUP98 fusions are associated with a specific disease phenotype and have been shown to inhibit differentiation of hematopoietic precursors and to increase self‐renewal of hematopoietic progenitor cells, providing a potential mechanism for malignant transformation [3]. Historically, our understanding of NUP98 rearrangements was largely drawn from the pediatric population data with limited molecular information regarding older patients who harbor NUP98re [20, 21, 22, 23]. Here we found that rearrangements of the NUP98 gene were rarely detected in cases with acute leukemia (1.3%), consistent with the available literature [2, 23, 24, 25, 26]. NUP98 rearrangements are more frequently observed in pediatric patients, often with relapsed or treatment‐resistant AML in which they define a subgroup associated with markedly adverse prognosis [23, 27]. The median age of the NUP98r AML cases in our study was 19 years compared with 62 years for the NUP98wt cases (p < 0.001) with many more cases detected in the pediatric and AYA age groups than NUP98wt patients. Overall, among patients with AML, NUP98re were found more often in pediatric and AYA patients than adults, which is concordant with the literature [21]. The higher prevalence of NUP98re in pediatrics compared to adult AML reflects a convergence of developmental biology, mutagenic mechanisms, and fundamentally different leukemogenic pathways. The most fundamental theory behind higher NUP98re prevalence in pediatrics AML is that approximately 70% of young children AML is driven by structural variants (translocations, inversions) that produce fusion oncoproteins, whereas only ∼10% of AML in older adults carries structural variants, and instead is driven by single nucleotide variant and indel insertion/deletion mutations [23, 27, 28]. To the authors' knowledge, this is the first report of enrichment for admixed American ancestry for NUP98re among AML, suggesting a potential for a distinct ancestry distribution. Among Eastern Asian population in a different study, higher NUP98re was observed in adult AML, which further support an underlying distinct ancestorial distribution among NUP98re AML [29]. NUP98re are seen across acute leukemia subtypes with myeloid predominance; more commonly seen in AML and mixed phenotype acute leukemia compared to ALL broadly [30]. Within ALL cases, NUP98 is almost exclusively seen in T‐ALL, which is concordant to our study, demonstrating no rearrangements in B‐ALL. This is largely explained by the fundamental interaction between NUP98 and HOXA/MEIS1 program which ultimately have the potential to preferentially blocks myeloid rather than lymphoid differentiation [3, 31, 32, 33, 34]. The current emerging data suggest that clinical outcomes and prognosis in NUP98re AML vary significantly by fusion partner [23, 35, 36, 37]. Here, we report 5 novel fusion partners for NUP98, including HOXB13, consistent with previous reports of the amino‐terminal portion of NUP98 protein fused with a wide range of proteins involving chromatin regulators or gene transcriptions in addition to homeobox DNA‐binding domains like HOX genes [4, 5, 6]. Previous studies have highlighted NUP98 fusion with JARID1A, NSD1, HOXA9 and HOXD13 as potent promoters of leukemogenesis [2, 22, 33, 34, 38, 39, 40]. Meanwhile, data from Children's Oncology Group (COG) and European cohorts showed similar prevalence of 2% of particularly NUP98‐KDM5A rearrangements [23]. Clinical outcomes are significantly poorer when these fusion proteins are accompanied by additional g abnormalities such as WT1 or Fms‐like tyrosine kinase 3 internal tandem duplications (FLT3‐ITD) [41, 42]. Inhibition of menin leads to anti‐leukemic effect in NUP98re AML preclinical studies [43, 44]. A study by Heikamp et al. identified menin‐MLL1 axis as an essential component of NUP98 fusion‐driven leukemia and showed that pharmacologic blockade of this interaction suppresses leukemic progression while promoting cellular differentiation [1]. Moreover, treatment with the menin‐MLL1 inhibitor VTP50469 in NUP98‐JARID1A and NUP98‐HOXA9 leukemic models has led to extensive alteration in gene expression of ley leukemogenic target genes leading to global loss of menin and selective loss of both MLL1 and NUP98 fusion [1]. In NUP98 fusion leukemia models, combined loss of MLL1 and the NUP98 fusion results in reduced expression at key loci such as Meis1, and rescued Meis1 expression by retroviral overexpression or CRISPRa diminishes sensitivity to VTP50469 [1]. These results, together with the sensitivity of NUP98re patient‐derived xenograft models to VTP50469 in vivo, support targeted inhibition of menin‐MLL1 interaction as a promising therapeutic strategy for NUP98 fusion hematologic malignancies. In our current study, genomic alterations involving WT1, FLT3, KRAS, and NOTCH1 were highly frequent in patients with NUP98re. Compared to literature, the rate of WT1 alterations in NUP98re was higher in our study which was observed in 77% compared to 40%–50% listed in other studies [22, 45, 46] with one study showing a frequency of cooccurrence rate of 22.2% [42]. Across all age groups, WT1 remained strongly associated with NUP98re in our study. FLT3 mutations were observed more in NUP98re compared to NUP98wt (49% vs. 26%, p < 0.001). The most commonly reported mutation associated with NUP98re in literature was FLT3 ITD reaching up to 85% [22, 45, 46]. In our study cohort, the association between FLT3 with NUP98re in overall population was mostly driven by AYA and adult groups and was not statistically seen in the pediatric population, which is likely explained by higher incidence of FLT3 mutations and complex secondary abnormalities among adult AML overall compared to pediatric population [23, 27]. Across multiple studies, ∼70–91% of NUP98::NSD1‐positive patients harbor concurrent FLT3 ITD [26, 47, 48]. Literature suggests that cases with concomitant NUP98re and FLT3 ITD are associated with markedly worse clinical outcomes compared to those lacking FLT3 ITD, highlighting their adverse prognostic effect [24]. This was also confirmed in a large cohort of pediatric patients (n = 1421) with AML previously enrolled in consecutive Children's Oncology Group/Children's Cancer Group and SWOG trials, where patients with a dual FLT3/ITD and NUP98/NSD1 had a significantly inferior CR rate of 27% versus. 69% compared to FLT3/ITD without NUP98/NSD1 (p < 0.001), and a corresponding inferior 3‐year overall survival of 31% versus. 48% (p = 0.011), respectively, with sustained inferiority in a multivariate analysis [26]. However, these patients, with dual FLT3/ITD and NUP98/NSD1, seem to benefit from FLT3 targeted treatment and/or venetoclax therapy with hematopoietic stem cell transplantation (HSCT) [22, 49]. While both FLT3 and menin inhibitors have been shown to be effective targeted therapy, un unanswered question remains how to select therapy and best approach patients with concurrent NUP98re and FLT3 mutations. Genomic alterations more frequent in NUP98wt AML include TET2, DNMT3A, ASXL1, NPM1, KMT2A, SRSF2, STAG2, and BCOR. Concordant with literature, alterations in NPM1 and KMT2A were exclusively seen in NUP98wt [34]. Mutations observed in secondary related AML, namely ASXL1, SRSF2, STAG2, and BCOR, were less frequently observed with NUP98re, which is consistent with the reported literature [50]. Of note, mutations in TP53 were observed nearly equally in both groups. While NUP98re correlates with inferior prognosis in AML, studies have shown cure potential with HSCT. Pre‐transplant NUP98re status has lower prognostic value in comparison to persistent NUP98re post HSCT which has shown significant correlation with higher incidence of relapse and mortality, indicating necessity of early molecular surveillance and early intervention [51]. Conclusion In this cohort of real‐world samples from patients with acute leukemia, NUP98re were rare (1.3%), most frequent in pediatrics and younger adults but distributed across all age groups. Patients with AML had a greater frequency of NUP98re compared to other acute leukemias. NUP98re is associated with a unique genomic landscape featuring both increased frequencies of mutations with proliferative features (FLT3 and KRAS) and growth suppression (WT1). NUP98re are mutually exclusive with other genes that are known targets of menin inhibitors. This work highlights the unique GA associated with NUP98 and emphasizes the need for clinical studies to reveal underlying biologic mechanisms and associations between different mutations to optimize management in the presence of potentially targetable alterations such as NUP98re. Funding The authors have nothing to report. Ethics Statement The study was conducted in accordance with the Declaration of Helsinki and approved by the Western Institutional Review Board (Protocol No. WCG #20152817, 3/1/2025). Consent Waiver of informed consent and a HIPAA waiver of authorization were obtained from the Western Institutional Review Board. Conflicts of Interest N.D., D.P., J.S.R., C.M., C.H., and L.W.P. are employees of Foundation Medicine Inc., and own stock in Roche Holdings, A.G. O.B., M.M., and N.S.N. declare they do not have any personal or financial conflicts of interest. Supporting information Supporting Information S1 Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Supplementary Materials Supporting Information S1 Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.