A Novel KMT2A-USO1 Fusion Gene-Induced De Novo Secondary Acute Myeloid Leukaemia in a Patient Initially Diagnosed with Acute Promyelocytic Leukaemia

Wen Jin,Li Chen,Yabin Liu,Qiusheng Chen,Ming Zhao,Yun Tan,Wei Zhang,Huan Song,Xiangqin Weng,Jianqing Mi,Saijuan Chen,Zhu Chen,Junmin Li,Kankan Wang

DOI: https://doi.org/10.1111/bjh.17183

2021-01-01

British Journal of Haematology

Abstract:The question whether secondary acute myeloid leukaemia (sAML) originates from a newly acquired or pre-existing clone has not been resolved for decades. Most cases of sAML are generally considered as therapy-related leukaemia,1 especially those with chromosomal translocations at 11q23 that involve the KMT2A gene (also known as MLL).2, 3 Here, we describe a 62-year-old male who was initially diagnosed with acute promyelocytic leukaemia (APL) carrying t(15;17)(q24;q21)/PML–RARα. He achieved complete remission after induction therapy with all-trans retinoic acid (ATRA), arsenic trioxide (ATO) and idarubicin (IDA). Subsequently, consolidation therapy with ATRA, IDA and cytarabine was administered for two cycles, and ATRA plus a intermediate dose of cytarabine for one cycle, followed by maintenance therapy with ATRA, ATO and methotrexate (MTX). Seventeen months after complete remission, the patient developed sAML with a t(4;11)(q21;q23) chromosomal translocation but no evidence of a PML–RARα transcript (Fig 1A and Figure S1, Table S1). During the consolidation and maintenance courses, the percentage and absolute count of peripheral monocytes in this patient increased repeatedly (Table S1). We first conducted whole genome sequencing (WGS) and RNA sequencing (RNA-seq) using bone marrow (BM) samples of APL and sAML to identify genomic and transcriptomic abnormalities, respectively. In the primary APL sample, we identified a typical PML–RARα transcript (Fig 1B, left panel). In the sAML sample, we found exon 9 of KMT2A to be fused with exon 2 of USO1 (Fig 1B, right panel), a novel KMT2A fusion partner, consistent with the t(4;11)(q21;q23) translocation detected in cytogenetic analysis (Fig 1A). USO1, also known as vesicle transport factor or P115, is a membrane tethering factor and has been reported to promote tumourigenesis through interaction with macrophage migration inhibitory factor and activation of the Erk pathway.4, 5 In addition to the distinct fusion genes, somatic mutation analysis revealed no shared mutations between the APL and sAML samples (Fig 1C,D, Tables S2 and S3), suggesting that the APL and sAML clones most likely arose from distinct founding clones. Notably, two distinct forms of FLT3 mutations, i.e. FLT3–ITD and FLT3–TKD, existed in the APL and sAML samples, respectively (Fig 1C,D). To characterize the phenotype and the potential mechanisms of sAML induced by this novel fusion KMT2A–USO1, we performed single-cell RNA sequencing (scRNA-seq) using the sAML sample enriched for CD11b/CD15 double-negative progenitor cells to explore the mechanisms underlying KMT2A–USO1-induced leukaemogenesis. High-quality gene expression data were obtained from 5 088 single cells, and unsupervised clustering identified eight cell-type clusters (Fig 1E). These scRNA-seq data further revealed that KMT2A–USO1 showed similar but distinct molecular features to other KMT2A rearrangements. We found that KMT2A–USO1 also harboured the common characteristics of KMT2A fusions in significantly upregulating HOXA cluster genes (HOXA9 and HOXA10) and the HOX coregulator gene PBX3 (Fig 1F). More importantly, KMT2A–USO1 possessed several distinctive features in promoting leukaemogenesis. First, MPO, the gene generally expressing lower counts in KMT2A-rearranged AML6 but higher counts in APL,7 was observed to be highly expressed in sAML (Fig 1G); this might explain the similar flow cytometric immunophenotyping between initial APL and KMT2A–USO1-induced sAML (Figure S1B). Second, genes involved in the fusion partner USO1-related pathways, such as the ERK1 and ERK2 cascade, were observed to be activated (MAPK3 and CD74) (Fig 1G). Third, based on the cell identity of each cluster, KMT2A–USO1-positive progenitors displayed a dramatic intratumoural heterogeneity (Figure S2A,B). KMT2A–USO1 interfered with the differentiation of multiple progenitor types along the myeloid lineage, including early-stage progenitor-like leukaemic cells (C0), immature myeloid progenitor-like leukaemic cells (C1), and typical granulocyte-macrophage progenitor-like leukaemic cells (C2). We also observed a cluster of proliferative progenitors (C3) that strongly expressed genes involved in the cell cycle and cell division. In addition, we found that C4 contained a relatively small population of immature and slowly dividing leukaemic stem cells (LSC), which expressed LSC markers (CD9, CD99 and CD44). The existence of LSCs and heterogeneous progenitors indicated that the development of sAML induced by KMT2A–USO1 might be a multi-lineage leukaemogenesis. Next, we performed a droplet digital polymerase chain reaction (ddPCR) to investigate the clonal evolution from initial APL to sAML and trace the origin of sAML (Fig 2A). First, APL-specific PML–RARα and FLT3–ITD decreased rapidly and could not be detected after the consolidation course (Fig 2B), suggesting that the ATRA/ATO-based treatment efficiently eradicated APL cells (Fig 2C, upper panel). Second, we found that a sAML-specific KMT2A–USO1-positive clone was present in only 0·02% of BM cells at APL onset and repeatedly increased during the consolidation and maintenance courses (Fig 2D). This suggested that KMT2A–USO1 was a pre-existing event and that sAML in our patient was de novo rather than a treatment-induced condition generally associated with KMT2A rearrangements. Also, the genomic breakpoint in KMT2A–USO1, located at intron 9 of KMT2A, also supported that our patient developed de novo sAML after complete remission of APL, since breaks in intron 9 are often found in de novo AML2 but those in intron 11 are highly sensitive against TOP2 inhibitors .2, 3, 8 Third, FLT3–TKD, another sAML-specific mutation, was detected only at the sAML stage [variant allele frequency (VAF) = 32%, ≈64% FLT3–TKD-positive cells] (Fig 2D). This suggested that sAML developed after KMT2A–USO1-positive pre-leukaemic monocytes acquired FLT3–TKD, supporting the multi-hit model of leukaemogenesis. Fourth, we found a strong correlation between KMT2A–USO1-positive cells and the progressively outgrowing monocytes before these cells transformed to sAML (Fig 2D,E). This observation suggests that KMT2A–USO1 promotes the abnormal proliferation of monocytes in BM. According to the 2016 WHO classification,9, 10 the stage before sAML was most likely to be re-diagnosed as chronic myelomonocytic leukaemia (CMML), based on the findings of the persistent peripheral monocytosis for more than three months combined with KMT2A–USO1, an acquired cytogenetic abnormality. This situation led us to propose the model for progression from primary APL to sAML presented in Fig 2F. Our detailed analysis of this rare case of sAML has led to a model of clonal evolution in which a novel KMT2A fusion gene, KMT2A–USO1, causes a pre-leukaemic clonal expansion of monocytes, which then acquire the proliferation mutation FLT3–TKD to give rise to AML. Our study opens up new lines of investigation to elucidate the origin and evolution of AML. The authors thank the patient and his family for their contribution to this study, and the clinicians and research staff at the Shanghai Institute of Hematology for making this study possible. This work was supported in part by the National Natural Science Foundation of China (No. 81890994, 81530003, 81770153, and 81800141) and the National Key R&D Program of China (No. 2019YFA0905900). KW designed the research; WJ, HS and XW performed experiments; LC, JL, QC and JM followed up the patient; YL, MZ, YT and WZ analysed results; KW and WJ wrote the paper; KW, LC, ZC and SC revised the manuscript; and all authors approved the manuscript. The authors declare no competing financial interests. Fig S1. Cytomorphologic and immunophenotypic findings of initial APL and sAML. (A) Morphological analysis of bone marrow smears (Wright–Giemsa stain; ×1000) in initial APL (left panel) and sAML (right panel), respectively. Red arrows point to the abnormal promyelocytes in the primary APL sample (left panel); green arrows to blast cells in the sAML sample (right panel). (B) Multiparameter flow cytometry (FCM) analysis of bone marrow cells in initial APL (left panel) and sAML (right panel). Blast cells were first gated by the side scatter (SSC) and the expression of CD45, then myeloid lineage-associated antigens and myeloperoxidase (MPO) were successively retrieved in the analyses. Abnormal leukaemic cells are shown in purple Fig S2. Single-cell RNA sequencing analysis of KMT2A–USO1-positive blasts in sAML samples. (A) Hierarchical clustering heatmap showing differentially expressed genes (rows) clustered in the eight cell-type clusters (columns). Prog, progenitor; MP, myeloid progenitor; GMP, granulocyte-macrophage progenitor, PPs, proliferative progenitors; LSC, leukaemic stem cell; Promono, promonocyte; T, T cell; Ery, erythroid progenitor. (B) Expression of two representative genes per cell-type cluster. Table SI. Clinical features of the patient during progression from initial APL to sAML. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.