Long‐term molecular remission in a patient with acute myeloid leukemia harboring a new NUP98‐LEDGF rearrangement

Abstract A large variety of molecular rearrangements of the NUP98 gene have been described in the past decades (n = 72), involving fusion partners coding for different transcription factors, chromatin modifying enzymes, as well as various cytosolic proteins. Here, we report the case of an AML‐M2 patient with a variant NUP98‐LEDGF/PSIP1 gene fusion (N9‐L10). In this patient, three different NUP98-LEDGF fusion mRNAs were characterized due to alternative splicing in LEDGF exon 11. Targeted high‐throughput sequencing revealed the presence of IDH1, SRSF2, and WT1 additional pathogenic mutations. To improve the therapeutic monitoring, quantification of NUP98‐LEDGF mRNA by real‐time PCR was developed. Because of poor response to conventional chemotherapy, allogeneic stem cell transplantation was performed, followed by 20 cycles of azacitidine‐based preemptive treatment of relapse. More than 31 months after diagnosis, corresponding to 25 months post SCT and 4 months after the last cycle of azacytidine, the patient is in complete molecular remission (undetectable NUP98‐LEDGF mRNA transcripts). This study highlights the considerable variability in breakpoint location within both NUP98 and LEDGF, associated with alternative splicing affecting LEDGF. It also emphasizes the need to fully characterize the breakpoints within the two genes and the identification of all fusion mRNAs, particularly for the development of a molecular monitoring assay. All these data seem critical for the optimal management of NUP98‐LEDGF + hematological malignancies commonly associated with a poor prognosis.

transporter for macromolecules such as proteins, mRNA, tRNA, and ribosome subunits. 4 It was recently shown that NUP98 fusion proteins could interact with MLL1 (mixed lineage leukemia 1) and NSL (nonspecific lethal) histone-modifiers within a complex that binds HoxA and HoxB promoter regions and drives leukemogenic transformation via the enhanced expression of HOX genes. 5 LEDGF (Lens Epithelium-derived Growth Factor) gene on 9p22, also named PSIP1 (PC4 and SFRS1 interacting protein 1) encodes two transcriptional coactivators (p75 and p52), generated by alternative splicing. 6 LEDGF/p75 and LEDGF/p52 proteins, localized in the nucleus, are known to bind chromatin. LEDGF/p75 interacts with HIV-1 integrase and promotes viral integration, 7 while LEDGF/p52 (lacking the C-terminal integrase-binding domain of LEDGF/p75) can play a role in the regulation of splicing. 8 Furthermore, it has been shown that LEDGF/p75 and MENIN1 bind to the N-terminus of MLL protein. This trimeric complex is quite important, because it targets MLL and MLL-fusion proteins to target gene promoters. In this respect, knockout mouse experiments have demonstrated that LEDGF/p75 has a critical role in the initiation of MLL-rearranged leukemia. 9 Rare t(9;11)(p22;p15) translocation with NUP98-LEDGF (NUP98-PSIP1) fusion has been described in adult patients with acute myeloid leukemia (M1/M2 AML), [10][11][12][13] transformed chronic myeloid leukemia, 14 or myelodysplastic syndrome with excess blasts (MDS-EB-2), 15 and in a pediatric case of transitional M2-M3 AML. 16 In this work, we have identified a novel NUP98-LEDGF rearrangement in acute myeloid leukemia (AML-M2). A comprehensive molecular analysis was conducted and three NUP98-LEDGF fusion transcripts were fully characterized. Using these data, a specific qRT-PCR assay was designed to evaluate the efficacy of the treatment. In addition to NUP98-LEDGF rearrangement, high-throughput sequencing has characterized additional IDH1, SRSF2, and WT1 pathogenic variants. Because of the nature of NUP98-related leukemia, the presence of additional genetic risk factors, and the absence of molecular remission after intensive induction chemotherapy, an allogeneic stem cell transplantation (SCT) was considered. Two and a half years after diagnosis and more than 2 years after SCT, the patient is still alive. He achieved complete remission without detectable minimal residual disease.

| Patient and IRB approval
The patient was admitted to our institution on February 2016, and the diagnosis of acute myeloid leukemia was made. He was included in a multicenter AML clinical trial (EudraCT Number: 2014-000699-24) and provided written informed consent in accordance with the declaration of Helsinki. For personal reasons, the patient withdrew from the trial 6 months later.

| Cytogenetic analyses
Whole chromosome painting FISH (fluorescent in situ hybridization) was performed using the XCP 9 green and XCP 11 orange DNA FISH probes (MetaSystems, Altlussheim, Germany) following the manufacturer's instructions. Conventional cytogenetic and FISH images were analyzed by means of the IKAROS and ISIS software, respectively (MetaSystems).

| Molecular experiments at diagnosis
Total RNA was extracted from both blood and bone marrow samples, and RT-PCR experiments were performed using the N988F forward primer (within NUP98 exon 8) associated with p75R (within LEDGF exon 12) or p52R (within LEDGF exon 11b) reverse primer as previously described. 12 Amplified fragments corresponding to NUP98-LEDGF rearrangements were then purified from agarose gel by the QIAquick gel extraction kit (Qiagen, Hilden, Germany) and sequenced using the BigDye Terminator v3.1 Cycle Sequencing kit (Life Technologies, Austin, TX). For Next Generation Sequencing (NGS) experiments, genomic DNA was extracted from bone marrow, and the amplicon library was generated using the TruSight myeloid sequencing panel (Illumina, San Diego, CA) that targets most genes mutated in AML. Paired-end sequencing was carried out on an Illumina MiSeq and data were analyzed by in-house bioinformatics pipeline.

| Characterization of the molecular rearrangement
Based on cytogenetic data, RT-PCR experiments were performed on blood and bone marrow samples using the N988F forward primer associated with p75R or p52R reverse primer. Following agarose gel electrophoresis, three different amplicons corresponding to one NUP98-LEDGF-p75 fusion  transcript and two NUP98-LEDGF-p52 fusion transcripts were observed ( Figure 1D). These amplified fragments were recovered from agarose gel and sequenced. In all cases, NUP98 exon 9 (N9) was fused in-frame with LEDGF exon 10 (L10), and the three NUP98-LEDGF fusion transcripts differed from an alternative exon 11 splicing within the LEDGF gene ( Figure 1E). More specifically, two splice isoforms corresponded to the usual LEDGF-p75 and LEDGF-p52 transcripts (NM_033222 and NM_001317900, respectively, Figure S1A,B). The third isoform was related to an unknown variant of LEDGF-p52 due to the presence of an unusual exon 11 ( Figure S1C). Using the Human Splicing Finder (www.umd.be/HSF3/) and Sroogle (http://sroogle.tau.ac.il) software, one splice acceptor site and several branch points were predicted upstream this atypical LEDGF exon 11 ( Figure S2). The potential impact of the detection in the blood and bone marrow samples from our patient of three NUP98-LEDGF fusion transcripts differing due to an alternative LEDGF exon 11, including one that was unknown, remains to be determined. Moreover, the two shorter versions of NUP98-LEDGF protein lack the IBD domain which would connect these fusion proteins to MLL biology ( Figure S3). Overall, we report a novel N9/L10 NUP98-LEDGF molecular rearrangement distinct from N8/ L3, 12 N9/L5, 16 N9/L7, 11,14 N11/L8, 15 and N12/L8 15 junctions previously reported in the literature (Table S1 and Figure S3).

| Targeted NGS analysis
Genomic DNA from bone marrow at diagnosis was then analyzed by NGS on a MiSeq sequencer using the TruSight Myeloid sequencing panel. Three genetic variants were characterized: IDH1 p.Arg132His, SRSF2 p.Pro95Thr, and WT1 p.Cys393Ter (Table S2). WT1 (tumor suppressor), SRSF2 (a component of the spliceosome machinery), and IDH1 (involved in DNA methylation) were found to be recurrently mutated in AML patients. 18 The pathogenic mutations of IDH1 (codon 132) and SRSF2 (codon 95) are commonly involved in MDS, AML associated with MDS and de novo AML. In the present case, IDH1 p.Arg132His and SRSF2 p.Pro95Thr mutations were detected at high variant allele frequency (VAF > 40%). The WT1 p.Cys393Ter variant (VAF 7%) has not yet been reported but appears to be probably pathogenic since it introduces a stop codon in the WT1 tumor suppressor protein.

| Treatment and molecular follow-up
Standard intensive induction chemotherapy treatment (idarubicin 9 mg/m 2 , days 1-5 and cytarabine 200 mg/m 2 , days 1-7) failed to induce a complete hematologic remission (46% bone marrow blast cells at day 44 with persistent aplasia). Complete remission with incomplete hematological recovery (CRi) was subsequently achieved using a cytarabine-based salvage treatment (1500 mg/m 2 /12 h, day 1, 3, 5). However, NUP98-LEDGF mRNA transcripts remained detectable, in both peripheral blood and bone marrow (Figure 2). Two consolidation cycles were then carried out with the same treatment protocol allowing CRi to be maintained. In this context, an allogeneic SCT was performed with an unrelated HLA-matched (10/10) donor. Peripheral blood stem cells were injected after a reduced intensity conditioning (fludarabine 30 mg/m 2 on days −6 to −2; busulfan 3.2 mg/kg on days −4 to −3 and thymoglobulin 5 mg/kg on days −3 to −2). Cyclosporine and mycophenolate mofetil were used as prophylaxis against Graft versus Host Disease (GvHD). The procedure was first complicated with acute cutaneous and gastrointestinal GvHD, and then successfully treated with F I G U R E 2 Molecular monitoring. Follow-up of the molecular disease by the quantification of NUP98-LEDGF mRNA transcripts in blood and bone marrow samples from the patient reported here. The x-axis represents the duration of the molecular monitoring (in months). Consecutive treatments are detailed below the curve. Ara-C, cytarabine; PBSCT, peripheral blood stem cell transplantation; aza C, azacitidine. Full circles and squares represent NUP98-LEDGF positive samples assessed by the NUP98-LEDGF/ABL1 ratio. Open circles and squares represent NUP98-LEDGF negative samples assessed by sensitivity of the test (1/number ABL1 copies) corticosteroids. At 7 months posttransplantation, the patient developed a moderate chronic mucocutaneous GvHD that resolved with corticosteroid and rapamycin therapy. Identified as a patient at high risk of relapse, it was decided that he would receive azacitidine as preemptive treatment (75 mg/ m 2 /day for five consecutive days every 28 days) beginning day 48 posttransplant. Concerning molecular monitoring, the NUP98-LEDGF/ABL1 ratio continuously decreased from the diagnosis to approximately 5 months posttransplant ( Figure  2). At this point, NUP98-LEDGF transcripts became undetectable, in both peripheral blood and bone marrow. Starting 5 months post allotransplant and up until now (more than 25 months post SCT and around 4 months after the last cycle of azacitidine), NUP98-LEDGF fusion transcripts have remained undetectable in peripheral blood (ABL1>60 000 copies in each experiment performed in triplicate). In addition, qRT-PCR performed at the time of the 20th and last cycle of azacitidine was negative in blood as well as in bone marrow samples. Overall, the patient achieved CR without minimal residual disease (CR MRD-).

| DISCUSSION
Very few cases of NUP98-LEDGF+hematological malignancies have been reported. [10][11][12][13][14][15][16] At diagnosis, the median age of patients was 52 (range 5-64) years and the median WBC was 50.5 (range 1.5-293) × 10 9 /L. However, at presentation, three patients had pancytopenia and two showed hyperleukocytosis. The median survival time was 8.3 (range 3-54) months. Overall, despite heterogeneous clinical and biological features, NUP98-LEDGF+ hematological malignancies usually present as very aggressive disorders. As shown by previously published data and by us in the present study, the high variability in breakpoint locations within both NUP98 (4) and LEDGF (5) is a peculiar characteristic of NUP98-LEDGF fusions. 11,12,[14][15][16] In all cases, NUP98-LEDGF-p75 and NUP98-LEDGF-p52 fusion transcripts arising via alternative splicing within LEDGF were detected. In addition to the present work, another group identified additional NUP98-LEDGF-type p52 fusions characterized by unusual splicing in LEDGF. 15 Therefore, NUP98-LEDGF fusions mostly differ from one patient to another. Theoretically, these rearrangements should encode fusion proteins harboring different structures and consequently displaying distinct functions. Theoretically, the NUP98-LEDGF fusions characterized in this work should encode three NUP98-LEDGF proteins (p75, p52, and type p52 variants). However, only variant p75 contains LEDGF integrasebinding domain which can interact with MLL to form the trimeric complex with MENIN1. This suggests that only the largest NUP98-LEDGF fusion protein is presumably able to bind to wild type MLL. However, from a prognostic point of view, the type of molecular rearrangement does not seem to correlate with clinical course. Nevertheless, complete characterization of breakpoints within NUP98 and LEDGF and identification of all fusion mRNA transcripts seem mandatory, particularly for implementation of a molecular monitoring system based on qRT-PCR.
Targeted NGS analysis performed at diagnosis and focused on acute myeloid leukemia driver genes is helpful since the presence of pathogenic mutants (IDH1 p.Arg132His, SRSF2 p.Pro95Thr, WT1 p.Cys393Ter in this study) could represent additional risk factors or opportunities for targeted therapy. Indeed, IDH1 mutations have an adverse prognostic effect in AML and constitute a drug-targetable gene alteration. 19,20 SRSF2 mutations are classically associated with poor outcome. 21,22 WT1 mutations are defined as a secondary event in AML and the unknown nonsense variant described here could have a pejorative impact. 23 In conclusion, we performed a comprehensive molecular analysis of a NUP98-LEDGF fusion detected in the blood and bone marrow samples from a patient with AML. A novel molecular rearrangement was characterized giving rise to three fusion mRNA transcripts. In addition, myeloid-focused NGS was performed, and a specific qRT-PCR system was developed for molecular follow-up. At the clinical level, an allogeneic SCT was performed because of (i) a rearrangement known to cause aggressive hemopathies, (ii) detection of additional IDH1, SRSF2, and WT1 point mutations, and (iii) the absence of molecular remission after intensive induction chemotherapy. More than 31 months after diagnosis, the patient achieved complete molecular remission. All in all, complete genetic characterization of the disease and implementation of personalized NUP98-LEDGF mRNA quantification facilitated therapeutic decisions and monitoring of treatment efficacy. More generally, in the case of NUP98-LEDGF fusion, allogeneic SCT should be considered as a reasonable therapeutic option, depending on the age of the patient, genetic factors and response to first-line conventional chemotherapy. In this regard, molecular monitoring based on the quantification of NUP98-LEDGF mRNA transcripts first in bone marrow (until the achievement of molecular remission) and then in peripheral blood appears critical for the evaluation of residual disease in malignancies classically associated with a poor prognosis.