SEARCH · Search PubMed

Results for “RUNX1”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 19 recordsLinked to original sources

PubMed2026

Preemptive hematopoietic stem cell transplantation in RUNX1 familial platelet disorder: a shared decision-making framework.

RUNX1 familial platelet disorder (RUNX1-FPD) is associated with a 35-50% lifetime risk of hematologic malignancy (HM). Like all germline HM predisposition syndromes, RUNX1-FPD can only be cured with allogeneic hematopoietic stem cell transplantation (HSCT). Current genetic screening techniques allow for early detection of germline predisposition and, consequently, the opportunity for HSCT before overt development of HM (i.e., preemptive HSCT). However, there is as yet no consensus on the use of preemptive HSCT for RUNX1-FPD. Described here is the case of an individual with RUNX1-FPD and a family history of HM who underwent preemptive HSCT. We introduce a shared decision-making framework designed to support individuals with RUNX1-FPD, their families, and their multidisciplinary clinical teams in evaluating whether and when to pursue preemptive HSCT versus continued surveillance. The framework reviews key medical factors that influence the decisions regarding timing of HSCT, including germline and somatic variants, clonal changes over time, familial history of HM, early morphologic or hematologic features, impacts on bleeding-related quality of life, and donor availability. The framework also summarizes the major risks and uncertainties potentially associated with preemptive HSCT while highlighting the associated ethical challenges. Together, the case and framework provide a structured, patient-centered approach for navigating the complex clinical decision of preemptive HSCT. Ongoing collaborative efforts to define cytogenetic and clonal changes preceding malignant transformation in RUNX1-FPD will refine the framework and bolster individualized treatment strategies aimed at preventing HM and improving the quality of life of individuals with RUNX1-FPD.

Humans↗

PubMed2026

Genomic Characterization of ETV6::RUNX1-Positive Childhood B-ALL in a Chinese Cohort: Novel Fusion Partners, Co-Occurring Mutations, and Risk-Stratifying Biomarkers.

BACKGROUND: ETV6::RUNX1 is the most common genetic abnormality in pediatric B-cell acute lymphoblastic leukemia (ALL; ∼25%), yet the comprehensive genetic architecture and molecular predictors of intermediate-risk (IR) stratification remain incompletely characterized. METHODS: We performed whole-transcriptome sequencing (Illumina NovaSeq 6000, rRNA depletion, 41.70 Gb/sample) on bone marrow samples from 93 pediatric ETV6::RUNX1-positive B-ALL patients. Bioinformatics analysis included STAR alignment, MuTect2 variant calling, FusionCatcher fusion detection, and VEP annotation. The Jaccard index with permutation testing assessed mutation co-occurrence; logistic regression identified independent predictors of IR classification. RESULTS: Beyond ETV6::RUNX1, we identified 51 distinct fusion genes across the cohort, including the reciprocal RUNX1-ETV6 (73.1%), chr8::KLF1210 (38.7%), and KLF12-chr8 (34.4%). Somatic mutations in 249 genes were detected; the most frequent were KIAA1715 (17.2%), KRAS (11.8%), and NSD2 (10.8%). Network analysis revealed significant chromatin modifier co-occurrence (KIAA1715-KMT2C: J = 0.136, p = 0.015) and KRAS-NRAS mutual exclusivity (J = 0.000, p = 0.042). PTCH1 (OR = 3.50, 95% CI 0.21-58.49, p = 0.41) and GNB1 (OR = 6.5, 95% CI 1.2-34.8, p = 0.029) mutations independently predicted IR classification. chr8::KLF1210 fusion correlated with higher Day-19 MRD levels (p = 0.038). CONCLUSIONS: GNB1 mutation represents a novel independent predictor of IR stratification in ETV6::RUNX1-positive B-ALL. The chromatin modifier co-occurrence module and extensive fusion architecture reveal biological heterogeneity within this favorable-risk subtype, with potential implications for risk-adapted therapeutic strategies.

B‐ALL↗

PubMed2026

A genetic signal at 8q12.3 modulates GGT levels via the Runx1-CYP7B1 axis in female ethnic minorities from Guizhou.

Gamma-glutamyl transferase (GGT) regarded as a biomarker of liver dysfunction or excessive alcohol consumption; however, existing genome-wide association studies (GWAS) have been conducted predominantly in European populations and East Asian populations from Japan and the Taiwan region, with limited investigation in ethnic minorities from Guizhou Province. Previous genetic studies have demonstrated that Guizhou ethnic minorities share an East Asian genetic background while exhibiting specific genetic structures, a pattern that is also confirmed by our principal component analysis (PCA) results. We therefore performed a GWAS in this population and identified a genome-wide significant signal at 8q12.3 in female ethnic minorities from Guizhou. Fine-mapping and functional annotation analyses suggest that a regulatory pathway involving Runt-related transcription factor 1 (Runx1)-Cytochrome P450 family 7 subfamily B member 1 (CYP7B1)-cholesterol-reactive oxygen species (ROS)-glutathione (GSH) may contribute to the regulation of GGT levels. Mendelian randomization (MR) analyses further supported a causal relationship between GGT levels and autoimmune hepatitis (AIH). These findings uncover a genetic mechanism underlying GGT variation at 8q12.3 in female ethnic minorities from Guizhou, implicating a pathway linked to cholesterol metabolism and oxidative stress, and providing potential targets and insights for precision prevention and treatment of related diseases.

Female↗

PubMed2026

Genetic Profile, Treatment Response, and Outcomes of BCR::ABL1-Positive Mixed-Phenotype Acute Leukemia: A Study From the BCR::ABL1 Pathology Group.

Mixed-phenotype acute leukemia (MPAL) with BCR::ABL1 fusion is rare, and its clinicopathological features, genetic landscape, therapeutic response, and patient outcomes remain incompletely defined, as does its relationship to blast-phase chronic myeloid leukemia. In this multicenter study of 44 patients, 86.4% had B/myeloid MPAL, 72.7% showed lymphoid predominance, 40.9% had complex karyotypes, and 68.3% harbored somatic mutations, most commonly RUNX1 mutations (46.3%). RUNX1 mutations frequently co-occurred with acute myeloid leukemia (AML)-associated alterations, whereas DNMT3A, TET2, and BCORL1 mutations were restricted to RUNX1-mutated cases. In contrast, acute lymphoblastic leukemia (ALL)-associated alterations (IKZF1 mutation/deletion and ETV6 mutations) were confined to RUNX1-wild-type patients. TP53 and signaling pathway mutations (NRAS, KRAS, PTPN11, and FLT3) were not detected. Forty-two patients received induction chemotherapy and/or immunotherapy combined with tyrosine kinase inhibitors: 74.2% of lymphoid-predominant patients and 63.6% of myeloid-predominant patients received ALL- and AML-type therapies, respectively. Ten patients relapsed, and 2 had primary refractory disease; some exhibited a dynamic shift in predominant lineage immunophenotype, chromosomal alterations, and somatic mutations at the relapse or refractory stage. The overall remission rate was 86.8%, with no significant differences across ALL-, AML-, or hybrid-type regimens. After a median follow-up of 24.2 months, the median overall survival was 52.5 months. Complex karyotype was associated with inferior overall survival compared with cases lacking additional chromosomal alterations (P = .02), whereas RUNX1 mutations were not. No significant differences in genetic profiles, treatment response, or outcomes were observed between patients with and without chronic myeloid leukemia-like features. This study provides a comprehensive genomic and clinical characterization of BCR::ABL1-positive MPAL, supporting improved risk stratification and future therapeutic strategies.

Adolescent↗

PubMed2026

Mutational Landscape and Clonal Dynamics in AML Undergoing PTCy Hematopoietic Cell Transplantation.

To improve risk stratification, we performed targeted NGS at diagnosis in 191 patients with AML undergoing myeloablative allogeneic HCT with PTCy-based prophylaxis. We also investigated clonal evolution using paired diagnostic and relapse samples from 39 individuals. A total of 610 mutations were detected in 184 patients (96%), most commonly in FLT3 (26%), DNMT3A (25%), RUNX1 (24%), and NPM1 (19%). Sixteen unique fusion genes were identified in 35 patients, with KMT2A (43%) and core binding factor rearrangements (23%) being the most frequent. TP53 and WT1 mutations were strongly associated with adverse outcomes, whereas NPM1 retained favorable significance. RUNX1 co-mutations with SF3B1 or NRAS were associated with inferior survival. In an exploratory allelic analysis, multi-hit TP53 alterations, but not single-hit mutations, were associated with distinctly poorer OS, EFS, and relapse risk. Relapse involved mutational shifts in ∼70% of cases, with significant enrichment of WT1 and more modest increases in TP53, KRAS, ASXL1, NF1, and MECOM, while DNMT3A, TET2, and ASXL1 persisted stably. Neither acute nor chronic graft-versus-host disease was associated with molecular remodeling at relapse. Incorporating TP53 and WT1 into risk models, recognizing context-dependent effects of DNMT3A and RUNX1, and applying longitudinal genomic monitoring may help guide personalized strategies to prevent relapse. Extended abstract BACKGROUND Relapse remains the leading cause of treatment failure after allogeneic hematopoietic cell transplantation (HCT) for acute myeloid leukemia (AML), yet the genetic mechanisms underlying post-transplant relapse remain poorly understood, particularly in the era of post-transplant cyclophosphamide (PTCy). Characterizing the mutational landscape at diagnosis and the clonal evolution leading to relapse may improve post-transplant risk stratification and identify opportunities for personalized surveillance and intervention. OBJECTIVES To characterize the diagnostic mutational landscape, evaluate its prognostic significance, and investigate clonal evolution from diagnosis to relapse in AML patients undergoing myeloablative HCT with PTCy-based graft-versus-host disease prophylaxis. STUDY DESIGN We performed targeted next-generation sequencing (NGS) at diagnosis in 191 consecutive AML patients undergoing myeloablative allogeneic HCT with PTCy-based prophylaxis. Paired diagnostic and relapse samples were available for 39 patients to evaluate clonal evolution. RESULTS A total of 610 mutations were detected in 184 patients (96%), most commonly in FLT3 (26%), DNMT3A (25%), RUNX1 (24%), and NPM1 (19%). Sixteen unique fusion genes were identified in 35 patients, with KMT2A (43%) and core binding factor rearrangements (23%) being the most frequent. TP53 and WT1 mutations were strongly associated with adverse outcomes, whereas NPM1 retained favorable significance. RUNX1 co-mutations with SF3B1 or NRAS were associated with inferior survival. In an exploratory allelic analysis, multi-hit TP53 alterations, but not single-hit mutations, were associated with distinctly poorer OS, EFS, and relapse risk. Relapse involved mutational shifts in ∼70% of cases, with significant enrichment of WT1 and more modest increases in TP53, KRAS, ASXL1, NF1, and MECOM, while DNMT3A, TET2, and ASXL1 persisted stably. Neither acute nor chronic graft-versus-host disease was associated with molecular remodeling at relapse. CONCLUSIONS This study provides a comprehensive characterization of the mutational landscape and clonal evolution of AML undergoing contemporary PTCy-based allogeneic HCT. TP53 and WT1 identify patients at particularly high risk of post-transplant relapse, whereas NPM1 retains favorable prognostic significance. The frequent acquisition of new genetic lesions at relapse underscores the dynamic nature of post-transplant clonal evolution and supports longitudinal molecular monitoring together with genomically informed post-transplant surveillance and relapse-prevention strategies.

Clonal Dynamics↗

PubMed2025

CDK12 inhibition reveals melanoma dependence on the RUNX1/CBFβ complex for genomic stability.

Cutaneous melanoma is the deadliest form of skin cancer, frequently driven by hyperactivation of the RAS/mitogen-activated protein kinase (MAPK) pathway. Cyclin-dependent kinase 12 (CDK12), a downstream effector of MAPK signaling, has emerged as a therapeutic target due to its essential role in transcriptional regulation and DNA damage repair. To identify vulnerabilities associated with CDK12 inhibition, we performed a genome-wide CRISPR-Cas9 screen and identified the Runt-related transcription factor RUNX1 and its cofactor CBFβ as synthetic lethal partners of CDK12. RUNX1 inhibition enhanced melanoma sensitivity to CDK12 inhibitors in a p53-independent manner, resulting in DNA damage accumulation and impaired repair capacity. Combined inhibition of CDK12 and RUNX1 suppressed melanoma growth in vivo. These findings identify RUNX1/CBFβ as a compensatory mechanism in CDK12-inhibited melanoma and define a synthetic lethal interaction with translational potential for combinatorial therapy.

Core Binding Factor Alpha 2 Subunit↗

PubMed2025

RUNX1A isoform is overexpressed in acute myeloid leukemia and is associated with FLT3 internal tandem duplications.

RUNX1A is the shortest and least expressed of the RUNX1 three main isoforms (A, B, C); despite this, the leukemogenic role of its overexpression has been clearly described. Several studies have shown RUNX1A involvement in different blood cancers and pilot observations in acute leukemia have been reported. In this context, we evaluated RUNX1 isoforms expression in a cohort of acute myeloid leukemia (AML) patients, finding overexpression of RUNX1A and RUNX1B, with higher median levels in thrombocytopenic cases. No difference was observed for RUNX1C. RUNX1A overexpression is higher in more immature AML phenotypes. According to the mutational profile, FLT3 internal tandem duplication (ITD) positive cases have the highest RUNX1A levels and the presence of FLT3-ITD was the only molecular variable able to influence RUNX1A expression. RUNX1A overexpression is disease-related, associated with a specific transcriptional profile, and reappears at relapse, with no clear kinetics except in FLT3-ITD cases. Overall, we demonstrate RUNX1A overexpression in AML and its association with the FLT3-ITD molecular subtype. Our data shed light on the dark side of RUNX1 deregulation, paving the way for further investigations.

Humans↗

PubMed2026

RAG-mediated structural variation and its impact on relapse risk in acute lymphoblastic leukemia.

Relapse during treatment of B-cell acute lymphoblastic leukemia (B-ALL) is a harbinger of poor outcomes. Identifying biomarkers for subsequent relapse risk which are detectable at B-ALL diagnosis remains a priority. Off-target recombination-activating gene (RAG)-mediated structural variants (SVs) generate genomic instability that drives leukemogenesis and may underlie treatment resistance. Leveraging sequencing data in 1,496 pediatric B-ALL patients enriched for relapse status (relapse n=532; non-relapse n=964), we characterized RAG-mediated SVs across B-ALL molecular subtypes and examined their association with patient characteristics and their impact on clinical outcomes. Off-target RAG-mediated SVs were overall frequent, particularly in ETV6::RUNX1, ETV6::RUNX1-like, and Ph-like B-ALL subtypes, while increasing age-at-diagnosis was positively associated with burden of off-target RAG-mediated SVs (P<.001). Off-target RAG-mediated SVs with a recombination signal sequence (RSS) at one breakpoint, a hallmark of off-target RAG activity, were significantly more frequent at diagnosis in patients who subsequently relapsed (P=.001). This association remained significant in multivariable regression analysis (per SV odds ratio [OR]:1.08, 95%CI:1.04-1.12), in minimal residual disease (MRD)-negative patients (OR:1.09, 95%CI:1.04-1.14) and across subtypes. Excluding deletions, MRD-negative ETV6::RUNX1 patients with ≥3 off-target RAG-mediated SVs had a >3-fold risk of relapse (hazard ratio:3.47, 95% CI:1.86-6.49). RAG-mediated SVs were also associated with relapse risk in T-cell ALL patients. Off-target RAG-mediated SV burden at diagnosis is a risk factor of relapse in pediatric ALL across molecular subtypes and independent of MRD status.

Journal Article↗

PubMed2026

Optical genome mapping enhanced by refined variant interpretation in pediatric acute lymphoblastic leukemia.

Reliable detection of structural variants (SVs) and copy number variations (CNVs) is crucial in the contemporary diagnostics of pediatric B-cell acute lymphoblastic leukemia (B-ALL). However, limitations of commonly used conventional and molecular cytogenetic methods may hinder the accurate genetic characterization of patients. Optical genome mapping (OGM) offers a reliable alternative by enabling high-resolution, genome-wide detection of CNVs and SVs. Chromosomal aberrations were screened using OGM in 51 children with B-ALL. The results were compared with those of karyotyping, fluorescence in situ hybridization (FISH), digital multiplex ligation-dependent probe amplification (digitalMLPA), and targeted RNA sequencing (RNA-seq). OGM data showed high congruency with karyotyping and FISH findings, detecting clinically relevant variants beyond G-banding results and unraveling a complex KMT2A fusion undetected by FISH. Gene fusions involved in complex ETV6::RUNX1 translocations, but not detected by RNA-seq, were confirmed using FISH. Normalization of OGM copy number values with DNA-index-improved concordance with FISH-derived copy numbers in near-tri/tetraploid cases. In the peripheral regions of OGM variants (fringe-zones), a novel evaluation strategy called 'FriZone' was applied, which significantly improved the concordance between OGM and digitalMLPA. In addition, a co-segregation analysis revealed strong associations between ETV6::RUNX1 fusion and deletions of ETV6, RAG2, and NR3C2. OGM uncovered complex rearrangements undetected by widely used methods in 15% of cases, improving genetic classification and risk stratification in 10% of the patients. The FriZone analysis and normalization by DNA-index provide a refined, more accurate approach to OGM variant interpretation, facilitating the efficient application of OGM in clinical diagnostics. © 2026 The Author(s). The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Humans↗

PubMed2026

Integrated analysis of amide proton transfer weighted MRI and proteomics uncovers altered protein dynamics in glioblastoma.

PURPOSE: Elevated amide proton transfer-weighted (APTw) MRI signals in glioblastoma (GBM) are often linked to increased intracellular mobile proteins, but the associated molecular patterns in human tissue remain unclear. We examined the relationship between regional APTw features and cellular protein composition and profiled proteomic differences between tumor and peritumoral tissue. METHODS: In this single-center prospective study, preoperative MRI data were integrated with intraoperative neuronavigation for 12 image-guided tissue samples (8 tumor and 4 peritumoral). Total, cytoplasmic, and nuclear proteins were quantified using bicinchoninic acid (BCA) assay. Data-independent acquisition (DIA) proteomics identified exploratory differentially expressed proteins (DEPs), followed by functional enrichment and protein-protein interaction (PPI) network analyses. Transcript-level expression patterns and survival associations were queried in The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) datasets to provide indirect external clinical context. RESULTS: Tumor regions showed higher APTw signals than peritumoral regions (p < 0.001) and increased cytoplasmic protein concentration (p < 0.05), without a corresponding increase in total or nuclear protein levels. DIA identified 654 DEPs. Further analysis highlighted 36 higher-significance DEPs, and prioritized 12 hub proteins in the PPI network. In public transcriptomic datasets, ERBB2, RUNX1, and SHC1 showed higher expression in GBM and were associated with poorer overall survival. CONCLUSION: These findings suggest that elevated APTw signal in GBM may be associated with increased cytoplasmic protein content and distinct proteomic alterations. This imaging-proteomic framework provides exploratory regional context for future mechanistic and follow-up studies, but larger, spatially matched and independently validated cohorts are required to confirm the molecular contributors to APTw contrast.

Humans↗

PubMed2026

Late acquisition of BCR::ABL1 during clonal evolution of SAMD9-associated MDS with phenotypic shift from AML to B-ALL.

We describe a unique case of SAMD9-associated myelodysplastic syndrome (MDS) with monosomy 7 that evolved over 16 years into BCR::ABL1-positive acute myeloid leukemia (AML) and subsequently manifested as B-cell acute lymphoblastic leukemia (B-ALL). Genomic analysis at AML diagnosis revealed a germline SAMD9 mutation together with somatic RUNX1 and PPM1D mutations, supporting stepwise clonal evolution, with BCR::ABL1 emerging as a late leukemogenic event. The dominant leukemic population at AML onset showed myeloid morphology and immunophenotype, whereas a minor CD19+CD10+ population was already detectable. Following venetoclax and azacitidine therapy, the dominant leukemic phenotype shifted to B-ALL while retaining BCR::ABL1 positivity. Detection of the Philadelphia chromosome in mature neutrophils at both AML onset and ALL relapse supported multilineage involvement of a multipotent BCR::ABL1-positive clone. Ponatinib achieved disease control. This case highlights late acquisition of BCR::ABL1 during SAMD9-associated clonal evolution and therapy-driven phenotypic shift within a shared Ph-positive leukemic stem-cell hierarchy.

Humans↗

PubMed2025

A quick guide to evaluating prime editing efficiency in mammalian cells.

According to the Clinvar database, modeling the diseases associated with pathogenic mutations requires the installation of base substitutions, small insertions or deletions. Prime editor (PE) was recently developed to precisely install any base substitutions and/or small insertions/deletions (indels) in mammalian cells and animals without requiring DSBs or donor DNA templates. PE also offers greater editing and targeting flexibility compared to other precision CRISPR editing methods because the versatile editing information is encoded in the reverse-transcription template of its prime editing guide RNA. However, optimal PE system selection and experimental design can be complex, and there are various factors that can affect PE efficiency. This chapter serves as a rapid entry-level guideline for the application of PE, providing an experimental framework for using PE at a specific genomic locus. RUNX1 was selected as a representative target site to illustrate the detailed methodology for constructing PE plasmids and the process of transfecting these plasmids into 293FT cells. We further examined the efficiency of PE-mediated genome editing in mammalian cells by using next-generation sequencing.

Gene Editing↗

PubMed2025

Genetic Analysis of Early Neoplasia in the Breast: Next-Generation Sequencing of Flat Epithelial Atypia and Associated Ductal and Lobular Lesions.

The molecular features of invasive breast cancers (IBC) have been well-characterized, but less is known about the earlier stages of neoplasia, including oncogenic drivers in early intraductal lesions. Flat epithelial atypia (FEA) is considered the earliest recognized precursor in the low-grade neoplasia pathway, but its mutational repertoire has not been studied, and drivers of the transition to morphologically more advanced lesions are unknown. Herein, we utilized next-generation sequencing to analyze 39 synchronous lesions from 13 patients, including FEA (n = 12) or predominantly FEA with early atypical ductal hyperplasia (FEA/early atypical ductal hyperplasia [ADH], n = 5) and associated ADH (n = 2), ductal carcinoma in situ (ductal carcinoma in situ [DCIS], n = 11), lobular carcinoma in situ (n = 3), and/or IBC with ductal and/or lobular differentiation (n = 6). Aside from 1 DCIS sample, all sequenced lesions in each patient were clonally related to one another. Recurrent alterations in FEA and FEA/early ADH included PIK3CA (69%), NCOR1 (31%), CBFB (31%), RUNX1 (15%), and GATA3 (23%). The mutational repertoire of FEA was similar to The Cancer Genome Atlas luminal IBC, except CBFB and NCOR1 mutations, which were more frequent in FEA and (along with PIK3CA, FOXA1, and CDKN1B) not always identified in paired morphologically advanced lesions. Compared with FEA, DCIS had more mutations and chromosomal copy number changes, including aberrations in PI-3 kinase pathway, transcription factors, chromatin remodeling genes, and TP53. CDH1 mutations identified in lobular carcinoma in situ were absent in paired FEA. Analysis of cases with ductal and lobular heterogeneity, including Rosen's triad, confirmed the shared clonality of the ductal and lobular components with features of genetic divergence. IBC of no special type were genetically similar to DCIS, and tubular carcinomas were similar to FEA. The results reveal the mutational repertoire of FEA and the genetics of early breast neoplasia, highlighting the clonal relationships of FEA to ductal and lobular carcinomas. Luminal breast cancer-associated genetic alterations are present at the earliest morphologically recognized stages of neoplasia.

Humans↗

PubMed2025

Deep learning-based cell-specific gene regulatory networks inferred from single-cell multiome data.

Gene regulatory networks (GRNs) provide a global representation of how genetic/genomic information is transferred in living systems and are a key component in understanding genome regulation. Single-cell multiome data provide unprecedented opportunities to reconstruct GRNs at fine-grained resolution. However, the inference of GRNs is hindered by insufficient single omic profiles due to the characteristic high loss rate of single-cell sequencing data. In this study, we developed scMultiomeGRN, a deep learning framework to infer transcription factor (TF) regulatory networks via unique integration of single-cell genomic (single-cell RNA sequencing) and epigenomic (single-cell ATAC sequencing) data. We create scMultiomeGRN to elucidate these networks by conceptualizing TF network graph structures. Specifically, we build modality-specific neighbor aggregators and cross-modal attention modules to learn latent representations of TFs from single-cell multi-omics. We demonstrate that scMultiomeGRN outperforms state-of-the-art models on multiple benchmark datasets involved in diseases and health. Via scMultiomeGRN, we identified Alzheimer's disease-relevant regulatory network of SPI1 and RUNX1 for microglia. In summary, scMultiomeGRN offers a deep learning framework to identify cell type-specific gene regulatory network from single-cell multiome data.

Deep Learning↗

PubMed2025

EZH2-driven immune evasion at disease presentation defines a targetable high-risk subset of acute leukemia exemplified by t(16;21) FUS::ERG AML.

The past 25 years of clinical trials have produced few improvements in pediatric AML (pAML) outcomes. This is acutely evident in patients with t(16;21)(p11;q22), yielding FUS::ERG. Patients with FUS::ERG-positive AML relapse quickly and do not respond to transplantation. Major histocompatibility complex (MHC) class I & II receptors and costimulatory molecules are absent at diagnosis in FUS::ERG-positive AML, mirroring the phenotype and outcomes of post-transplant relapse. We show that this is driven by overexpression of EZH2, in vitro and in multiple clinical cohorts. While FUS::ERG AML is the most extreme example, this phenotype is shared by lethal CBFA2T3::GLIS2-driven AML, and patients with RUNX1::RUNX1T1 have significantly worse outcomes when EZH2 overexpression co-occurs. The FDA-approved EZH2 inhibitor tazemetostat reverses this phenotype, re-establishes MHC presentation, and elicits immune effector cell-mediated elimination. EZH2 inhibitors may provide the first targeted therapeutic frontline option for AML patients with FUS::ERG, with the potential for broader frontline immunostimulatory benefits.

Journal Article↗

PubMed2026

Multimodal analysis of CD38 in T-cell Acute Lymphoblastic Leukemia Identifies Combinatorial Therapeutic Strategies.

Outcomes for pediatric patients with refractory or relapsed T-cell acute lymphoblastic leukemia (T-ALL) are poor, underscoring the need for improved therapeutic strategies. CD38, a type II transmembrane glycoprotein, is a promising target in T-ALL, with clinical trials evaluating CD38-targeting immunotherapies in frontline and relapsed settings. However, the biological role of CD38 in T-ALL has not been systematically defined. We interrogated CD38 biology through multimodal profiling of pediatric T-ALL samples. Bulk RNA sequencing of 1,335 primary tumors revealed that CD38 expression varies across genomic and immunophenotypic subtypes in T-ALL. Flow cytometry of 150 primary samples and CITE-sequencing of 40 cases demonstrated broad surface expression of CD38. A transcription factor CRISPR-screen identified RUNX1, RUNX3, and TP53 as candidate positive regulators of CD38. Metabolomic profiling of cell lines further revealed disruption of the polyamine pathway following CD38 perturbation. Supporting this finding, co-targeting CD38 with difluoromethylornithine (DFMO), a polyamine metabolism disruptor, improved survival in preclinical models. Across transcriptomic datasets, including primary tumors, cell lines, and patient-derived xenograft models, IL32 expression consistently decreased following CD38 loss or negativity, supporting an association between CD38 and inflammatory signaling pathways. Additionally, CD38 and LCK expression were positively correlated across majority of genomic subtypes, implicating SRC kinase signaling. Consistent with this, daratumumab in cell lines increased LCK phosphorylation, and combination therapy with dasatinib improved survival compared to monotherapy. Collectively, these findings define previously unrecognized interactions between CD38 and targetable pathways and genes in T-ALL and identify rational combinatorial strategies to enhance CD38-directed therapies and reduce relapse risk.

Journal Article↗

PubMed2026

Advances in the diagnosis and classification of B-ALL: comparative insights from updated guidelines.

Accurate molecular classification is essential for diagnosis, risk stratification, and treatment selection in B-cell lymphoblastic leukemia (B-ALL). In this study, we performed a comprehensive, real-world reclassification of 1015 consecutively diagnosed B-ALL patients using the fifth edition of the World Health Organization Classification of Haematolymphoid Tumours (WHO-HAEM5) and the International Consensus Classification (ICC). An integrative genomic strategy that combined whole transcriptome sequencing, fusion detection, mutational analysis, and cytogenetics enabled reclassification according to both the WHO-HAEM5 and ICC frameworks, thereby substantially reducing the proportion of unclassifiable B-ALL from 41.9% (2016 WHO revision [WHO-HAEM4R]) to 15.9% (WHO-HAEM5) and 11.9% (ICC). Distinct clinical and prognostic features were identified across newly defined subtypes. Multivariable analysis confirmed that this genomic classification is a robust, independent predictor of survival after adjusting for age, minimal residual disease status, and transplant intervention. Specifically, HLF-rearranged and MEF2D-rearranged B-ALL conferred a persistently poor prognosis across all age groups despite allogeneic hematopoietic stem cell transplantation, highlighting an urgent need for novel therapeutic strategies. Gene expression profiling resolved cryptic subtypes, including ETV6::RUNX1-like, ZNF384-rearranged-like, and BCR::ABL1-like B-ALL, and uncovered diagnostic ambiguity in patients with concurrent lesions. In addition, we report emerging high-risk groups, including IDH1/2- and ZEB2 Q1072-mutated B-ALL, that may warrant recognition as distinct molecular entities. Our findings demonstrate the clinical use of integrative transcriptomic profiling in refining B-ALL taxonomy in guiding risk-adapted therapies and informing future revisions of diagnostic standards. This study supports the incorporation of high-throughput molecular diagnostics into routine leukemia classification and precision treatment planning.

Humans↗

PubMed2026

Age- and sex-adjusted genomic differences between Korean and Beat AML cohorts.

Genomic profiling plays a central role in risk stratification and therapeutic decision-making in acute myeloid leukemia (AML), yet the clinical implications of population-specific genomic architectures remain incompletely defined. We conducted a prospective, multicenter study of 603 adults with newly diagnosed AML in Korea, integrating targeted sequencing of 83 recurrently mutated genes with comprehensive clinical annotation across treatment intensities, including allogeneic hematopoietic stem cell transplantation (allo-HSCT). For contextual comparison, genomic profiles were evaluated against the Beat AML cohort. The overall genomic landscape was broadly conserved, supporting shared core disease biology across populations. However, RUNX1::RUNX1T1, CEBPA, GATA2, KIT, and DDX41 mutations were more frequent in the Korean cohort, whereas FLT3 and NPM1 mutations were less common. These differences translated into a distinct distribution of European LeukemiaNet (ELN) 2022 risk categories, with implications for therapeutic stratification. Notably, most DDX41 alterations were germline (3.2%), highlighting the need for systematic germline evaluation with implications for genetic counseling and donor selection. Although unadjusted overall survival appeared longer in the Korean cohort, this difference was not significant after adjustment for key clinical variables. These findings indicate that population-specific genomic distributions reshape the clinical application of risk stratification and support population-aware precision medicine strategies in AML.

Journal Article↗