To the Editor:
Genomic aberrations—gene fusions in the majority of cases—and corresponding transcriptional regulations define an increasingly complex landscape of molecular subtypes in B cell precursor acute lymphoblastic leukemia (BCP-ALL) . Up to 15% of patients cannot be allocated to established subtypes, suggesting the presence of unrecognized drivers—especially in adult patients who have been less studied so far.
We performed transcriptome sequencing (RNA-Seq) on n = 568 adult BCP-ALL patients prospectively treated according to pediatric-based protocols of the German Multicenter Acute Lymphoblastic Leukemia (GMALL) study group including risk stratification based on minimal residual disease (MRD) and treatment intensification for high-risk patients. To define molecular subtypes, we used our previous integrative analyses  to train a machine learning classifier to predict subtype allocation from gene expression profiles of subsequently sequenced samples. Feature selection (LASSO) was used to identify the most informative genes. Underlying genomic aberrations were analyzed (whole-genome sequencing (WGS), whole-exome sequencing (WES); SNP-arrays) to confirm subtype allocation in selected cases. With this approach, we were able to allocate n = 535/568 (94%) samples to 15 previously established  molecular subtypes (Fig. 1A–D), with confirmation of corresponding genomic alterations in 91% of analyzed cases (Fig. 1D). Unsupervised gene expression analysis of previously unassigned samples revealed a distinct patient subset (n = 12; Supplementary Fig. S1A) defined by a novel in-frame gene fusion of upstream binding transcription factor (UBTF) and ataxin-7-like protein 3 (ATXN7L3) occurring exclusively in this patient cluster (n = 12/12 vs. n = 0/556 in remaining cohort, p < 1E−10; Fig. 1D). Comparison of gene expression profiles revealed that UBTF::ATXN7L3 rearranged cases in our cohort match to a recently described BCP-ALL subtype, which so far was identified by increased expression of the homeobox transcription factor CDX2 (‘CDX2 high’ ALL)  (Supplementary Fig. S1B). UBTF::ATXN7L3 represents an 11.3 kbp in-frame read-through between UBTF exon 17/21 and a 5′ UTR splice site of ATXN7L3, with the same sanger sequencing confirmed break point in all samples (Fig. 2A, Supplementary Methods). WGS of 3 samples revealed a 10.08 kb genomic deletion involving UBTF 3′ exons (18-21) and most of the intergenic region between UBTF and ATXN7L3 as underlying mechanism (Fig. 2A, Supplementary Fig. S2). Break-point-specific PCR and Sanger sequencing confirmed presence of the deletion in n = 11/11 UBTF::ATXN7L3 patients with available material (Supplementary Fig. S3). The same ATXN7L3 transcript breakpoint has previously been identified in a patient with diffuse large B cell lymphoma (GPATCH8::ATXN7L3) , suggesting a shared driver function in different B-lymphoid malignancies. Both fusion partners were highly expressed across the entire cohort without significant differences in the novel subtype (Supplementary Fig. S4), suggesting either gain-of-function or a dominant-negative effect of the gene fusion.
Both, UBTF and ATNX7L3 are global epigenetic regulators involved in transcriptional control. UBTF is an essential co-activator of ribosomal RNA expression. Very recently, UBTF has been characterized as novel oncogene in acute myeloid leukemia (AML), where internal tandem duplications define a distinct molecular subtype with poor outcome and highest incidence in early adolescents [4, 5]. WGS and sanger sequencing ruled out UBTF internal tandem duplications in UBTF::ATXN7L3 patients (data not shown). ATXN7L3 is a global gene expression co-activator through the SAGA complex. It is essential for activation of the SAGA histone deubiquitinase module (DUBm) through USP22, which is part of the 11-gene signature “Death-from-cancer”  defining poor outcomes across entities. The SAGA DUBm competes for ATXN7L3-binding with other deubiquitinases suggesting global changes in gene expression upon imbalances in ATXN7L3-substrate binding . These findings align well with data on other molecular ALL subtypes driven by epigenetic perturbations [8, 9]. Analysis of subtype-specific gene expression by multi-comparison ANOVA revealed 332 genes with differential expression in UBTF::ATXN7L3 ALL when compared to each other subtype (Fig. 1B; Supplementary Tables S1 and 2). These differentially expressed genes included upregulation of 18 cancer-associated genes (COSMIC Cancer gene census, Supplementary Table S3), one of which was CDX2, which has been used to define ‘CDX2-high’ ALL  (Fig. 1C). However, few samples from other subtypes also showed increased CDX2 expression levels, limiting its applicability to identify this subtype. UBTF and ANTX7L3 are global epigenetic regulators without described functional interactions with CDX2. CDX2 is expressed in AML  and ALL , independently of the driver subtype. Conditional Cdx2 overexpression in hematopoietic progenitors resulted in myelodysplasia but required acquisition of secondary aberrations for leukemic transformation , suggesting a cooperative function during leukemogenesis. Although UBTF::ATXN7L3-specific gene expression showed little overlap with published CDX2 overexpression models (Supplementary Fig. S5A), we identified a functional module relating CDX2 to HOXA9 and MEIS1 overexpression in UBTF::ATXN7L3 ALL, in line with similar findings in AML  (Supplementary Figure S5B,C). HOXA9/MEIS1 are essential co-factors for KMT2A-driven leukemogenesis , making it possible that a CDX2-HOXA9/MEIS1 axis exerts a similar leukemia promoting role in UBTF::ATXN7L3 ALL. Further oncogenes related to hematologic malignancies were also upregulated in UBTF::ATXN7L3 patients (Fig. 1C, Supplementary Figure S6), including NTRK3 which might represent a therapeutic target for specific inhibitors (e.g. larotrectinib, entrectinib). To evaluate additional genomic driver aberrations, we performed WES (n = 7) and/or SNP-array analyses (n = 6) showing a described enrichment of chromosome 1q gains  (n = 5/7) and heterogeneous single chromosome aberrations. However, no subtype-specific recurrent driver events were identified (Supplementary Fig. S7A), supporting the functional relevance of UBTF::ATXN7L3 as recurrent hallmark of this subtype. UBTF::ATXN7L3 ALL was enriched for pro-B immunophenotypes (n = 5/12, 42% vs. n = 70/530, 13%; p = 0.016) and occurred predominantly in female patients (n = 10/12, 83% vs. 237/534, 44%; p = 0.008) and patients of advanced age (median: 48.5 years vs. 38 years; p = 0.05).
Outcome evaluable UBTF::ATXN7L3 patients (n = 11/12; Fig. 2B) received treatments on pediatric inspired GMALL protocols. Risk stratification identified 6 patients as high-risk due to pro-B immunophenotype (n = 4) or late response (n = 2). One patient died during induction therapy and another patient failed to achieve hematologic CR after consolidation I (overall cytologic CR rate: 82%). Only n = 3/10 patients cleared MRD after consolidation I (cytologic and molecular CR) compared to n = 271/402 (67%; p = 0.019; Fig. 2C) in the remaining cohort. Two out of these three good responders remained in molecular CR after conventional chemotherapy including allogenic stem cell transplantation (HSCT) due to high-risk criteria. One patient relapsed after discontinuation of standard chemotherapy due to poor performance status and achieved a second molecular CR after inotuzumab ozogamizin. Patients with intermediate MRD response (positive MRD < 10−4 or below quantifiable range, n = 3) experienced molecular relapses on standard therapy, received Blinatumomab followed by HSCT and remained in long-term remission (n = 2/3) or achieved sustained CR on standard therapy (n = 1/3). Among the remaining poor responders (n = 4), one cytologic non-responder achieved MRD-negativity after Blinatumomab, received HSCT and died due to transplant-related complications. Two patients received Blinatumomab, achieved a molecular CR, proceeded to HSCT, and remained in long-term remission. The fourth patient received HSCT without Blinatumomab, relapsed, and achieved intermediate MRD after 2nd HSCT. Together, we observed a median overall survival probability of UBTF::ATXN7L3 patients of 80% (±12%) compared to 73% (±2%; p = 0.07; Fig. 2D) in the remaining cohort, which is comparable to the ongoing GMALL08/2013 study . Yasuda et al . reported markedly lower survival rates (pOS: 26.7%, (4.8-56.3)) in ‘CDX2-high’ patients treated in historical cohorts without MRD-based risk stratification. Together, these data suggest that UBTF::ATXN7L3 ALL represents a less chemo-sensitive disease subtype, which can be successfully salvaged by current MRD-based concepts incorporating immunotherapies and stem cell transplantation .
Other subtypes with poor MRD response in our cohort included ZNF384 (48.2% MRD negative after consolidation I; p = 0.056), Ph-like ALL (54.0% MRD negative, p = 0.003) and KMT2A (55.8% MRD negative, p = 0.127), whereas high hyperdiploid (90.9% MRD negative, p = 0.01) and TCF3::PBX1 (94.1% MRD negative, p = 0.016) subtypes showed favorable MRD responses. This heterogeneity and recently published differences in treatment outcomes of ALL subtypes when treated with or without MRD-based risk-stratification highlight the importance of evaluating the clinical course of novel molecular subgroups in the context of current treatment strategies.
Yasuda et al . described a second novel BCP-ALL subtype defined by IDH1/2 hotspot mutations (1.9% of cohort). We screened RNA-Seq data of all remaining ‘unassigned’ samples of our cohort (n = 22) for the described gene expression signature or IDH1/2 mutations and identified one patient harboring IDH2 p.R140Q, which was confirmed by PCR on gDNA level, contributing to the heterogeneous frequency distribution of molecular subtypes in different BCP-ALL cohorts.
Our data identify UBTF::ATXN7L3 resulting from a 17q21.31 variant as novel subgroup defining candidate driver fusion for the recently described ‘CDX2-high ALL’ subtype. Poor MRD response indicates reduced chemosensitivity in these patients. Our data suggest MRD-based treatment intensification using salvage immunotherapies and allogenic stem cell transplantation as a promising strategy to rescue this high-risk phenotype.
Bastian L, Schroeder MP, Eckert C, Schlee C, Tanchez JO, Kämpf S, et al. PAX5 biallelic genomic alterations define a novel subgroup of B-cell precursor acute lymphoblastic leukemia. Leukemia. 2019;33:1895–909.
Yasuda T, Sanada M, Kawazu M, Kojima S, Tsuzuki S, Ueno H et al. Two novel high-risk adult B-cell acute lymphoblastic leukemia subtypes with high expression of CDX2 and IDH1/2 mutations. Blood. 2021: blood.2021011921.
Gao Q, Liang W-W, Foltz SM, Mutharasu G, Jayasinghe RG, Cao S, et al. Driver fusions and their implications in the development and treatment of human cancers. Cell Rep. 2018;23:227–.e3.
Stratmann S, Yones SA, Mayrhofer M, Norgren N, Skaftason A, Sun J, et al. Genomic characterization of relapsed acute myeloid leukemia reveals novel putative therapeutic targets. Blood Adv. 2021;5:900–12.
Umeda M, Ma J, Huang BJ, Hagiwara K, Westover T, Abdelhamed S et al. Integrated genomic analysis identifies UBTF tandem duplications as a recurrent lesion in pediatric acute myeloid leukemia. Blood Cancer Discov 2022: bloodcandisc.BCD-21-0160-A.2021.
Glinsky GV, Berezovska O, Glinskii AB. Microarray analysis identifies a death-from-cancer signature predicting therapy failure in patients with multiple types of cancer. J Clin Investig. 2005;115:1503–21.
Atanassov BS, Mohan RD, Lan X, Kuang X, Lu Y, Lin K, et al. ATXN7L3 and ENY2 coordinate activity of multiple H2B deubiquitinases important for cellular proliferation and tumor growth. Mol Cell. 2016;62:558–71.
Huang Y, Mouttet B, Warnatz H-J, Risch T, Rietmann F, Frommelt F, et al. The leukemogenic TCF3-HLF complex rewires enhancers driving cellular identity and self-renewal conferring EP300 vulnerability. Cancer Cell. 2019;36:630–644.e9.
Ayton PM, Cleary ML. Transformation of myeloid progenitors by MLL oncoproteins is dependent on Hoxa7 and Hoxa9. Genes Dev. 2003;17:2298–307.
Scholl C, Bansal D, Döhner K, Eiwen K, Huntly BJP, Lee BH, et al. The homeobox gene CDX2 is aberrantly expressed in most cases of acute myeloid leukemia and promotes leukemogenesis. J Clin Investig. 2007;117:1037–48.
Thoene S, Rawat VPS, Heilmeier B, Hoster E, Metzeler KH, Herold T, et al. The homeobox gene CDX2 is aberrantly expressed and associated with an inferior prognosis in patients with acute lymphoblastic leukemia. Leukemia. 2009;23:649–55.
Vu T, Straube J, Porter AH, Bywater M, Song A, Ling V, et al. Hematopoietic stem and progenitor cell-restricted Cdx2 expression induces transformation to myelodysplasia and acute leukemia. Nat Commun. 2020;11:3021.
Rawat VPS, Thoene S, Naidu VM, Arseni N, Heilmeier B, Metzeler K, et al. Overexpression of CDX2 perturbs HOX gene expression in murine progenitors depending on its N-terminal domain and is closely correlated with deregulated HOX gene expression in human acute myeloid leukemia. Blood. 2008;111:309–19.
Goekbuget N, Stelljes M, Viardot A, Nachtkamp K, Steffen B, Schneller F, et al. First results of the risk-adapted, MRD-stratified GMALL trial 08/2013 in 705 adults with newly diagnosed acute lymphoblastic leukemia/lymphoma (ALL/LBL). Blood. 2021;138:362–362.
Nicorici D, Satalan M, Edgren H, Kangaspeska S, Murumagi A, Kallioniemi O, et al. FusionCatcher - a tool for finding somatic fusion genes in paired-end RNA-sequencing data. 2014. https://www.biorxiv.org/content/10.1101/011650v1.
This study was in part funded by funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) – project number 444949889 (KFO 5010/1 Clinical Research Unit ‘CATCH ALL’ to LB, AH, MPH, MN, MBr, and CDB), Deutsche Jose Carreras Leukämie Stiftung (DJCLS 01R/2016 to LB and CDB).
Open Access funding enabled and organized by Projekt DEAL.
MBr received personal fees from Incyte (advisory board) and Roche Pharma AG, financial support for reference diagnostics from Affimed and Regeneron, grants and personal fees from Amgen (advisory board, speakers bureau, travel support), and personal fees from Janssen (speakers bureau), all outside the submitted work. The remaining authors declare no competing financial interests. None of the remaining authors has a relevant conflict of interest.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Bastian, L., Hartmann, A.M., Beder, T. et al. UBTF::ATXN7L3 gene fusion defines novel B cell precursor ALL subtype with CDX2 expression and need for intensified treatment. Leukemia 36, 1676–1680 (2022). https://doi.org/10.1038/s41375-022-01557-6
Die Onkologie (2022)