Clinical whole-genome sequencing identi�es NSD3 as the correct fusion partner of NUP98 in a patient with acute myeloid leukemia and t(8;11) (p11.2;p15): a case report

Background: Only rare cases of acute myeloid leukemia (AML) have been shown to harbor a t(8;11)(p11.2;p15.4). This translocation is believed to involve the fusion of NSD3 or FGFR1 with NUP98; however, apart from targeted mRNA quantitative PCR analysis, no molecular approaches have been utilized to de�ne the chimeric fusions present in these rare cases. Case Presentation: Here we present the case of a 51-year-old female with AML with myelodysplastic-related morphologic changes, 13q deletion and t(8;11), where initial �uorescence in situ hybridization (FISH) assays were consistent with the presence of NUP98 and FGFR1 rearrangements, and suggestive of NUP98/FGFR1 fusion. Using a streamlined clinical whole-genome sequencing approach, we resolved the breakpoints of this translocation to intron 4 of NSD3 and intron 12 of NUP98, indicating NUP98/NSD3 rearrangement as the underlying aberration. Furthermore, our approach identi�ed small variants in WT1 and STAG2, as well as an interstitial deletion on the short arm of chromosome 12, which was cryptic in G-banded chromosomes. Conclusions: NUP98 fusions in acute leukemia are predictive of poor prognosis. The associated fusion partner and the presence of co-occurring mutations, such as WT1, further re�ne this prognosis with potential clinical implications. Using a clinical whole-genome sequencing analysis, we resolved t(8;11) breakpoints to NSD3 and NUP98, ruling out the involvement of FGFR1 suggested by FISH while also identifying multiple chromosomal and sequence level aberrations.


Background
The presence of t(8;11)(p11.2;p15.4) has been described in rare cases of acute myeloid leukemia (AML) as an isolated aberration [1,2], with an additional chromosomal aberration [3][4][5], or in the setting of a complex karyotype [6]. Previously reported FISH studies had suggested the involvement of either FGFR1 [6] or NSD3 [1] in the formation of a chimeric protein with NUP98. However, when performed, targeted RT-PCR [1,4] and Southern hybridization [5] analyses have revealed expression of NUP98/NSD3 chimeric fusion transcripts. Here we present a case of AML with t(8;11)(p11.2;p15), which initially appeared to be an isolated NUP98/FGFR1 rearrangement as detected by uorescence in situ hybridization (FISH). However, a subsequent clinical whole-genome sequencing assay [7] identi ed NSD3 as the correct fusion partner, while also revealing the presence of sequence variants in WT1 and STAG2, as well as an interstitial deletion on the short arm of chromosome 12. To our knowledge, this represents only the seventh reported case of t(8;11), a translocation which has not been documented in the literature or publicly accessible databases since 2009 [5]. Furthermore, this report illustrates the potential clinical utility of whole-genome sequencing in AML and myelodysplastic syndrome (MDS).

Case Presentation
This is a case of a 51 year-old female who had received a heart transplant with basiliximab induction therapy 6 years prior to nonischemic cardiomyopathy. She presented with 1-week of exertional chest pain associated with nausea, shortness of breath and dizziness. She had no prior history of cardiac graft rejection and was on stable doses of tacrolimus and mycophenolate mofetil, as well as infectious prophylaxis with acyclovir and trimethoprim/sulfamethoxazole. Initial work-up for a primary cardiopulmonary etiology, including electrocardiogram (ECG), troponins, transthoracic echocardiogram (TTE) and chest radiograph (CXR), was unremarkable.
Sequential metaphase FISH studies were performed to determine the possible involvement of FGFR1 (8p11) and NUP98 (11p15) loci. Results showed 23% with 3'FGFR1 on derivative 11p and 36.5% with 5'NUP98 on derivative 8p, indicative of FGFR1 and NUP98 rearrangements ( Table 1; Fig. 1b-c). Subsequent clinical genome-wide sequencing was performed via ChromoSeq (Methods S1) [7], a next generation sequencing (NGS)-based assay which detects SNVs and indels (≥73bp) in 40 genes or gene hotspots, and 612 pre-de ned structural variants including 624 genes, and genome-wide copy number alterations (≥5Mb). The ChromoSeq assay identi ed NSD3 (Nuclear Receptor Binding SET Domain Protein 3), a gene 27 kb downstream of FGFR1, as the fusion partner of NUP98 (Table 1; Fig. 1d), speci cally between intron 4 of NSD3 (Fig. 1e) and intron 12 of NUP98. Consistent with karyotyping, subsequent FISH assay con rmed the deletion of the long arm of chromosome 13 (13q14.3) in 51% of the cells (Table 1; Fig. S1g-h). Although deletion of 13q is relatively more common in lymphoid malignancies, it is also found in various types of myeloid neoplasms. The chromosome interval that is deleted in all reported myeloid cases was found to be 13q13-21, which includes RB1 [8]. Studies have shown that the 13q12-q22 deletion interval is more recurrent in MDS, while 13q21 and 13q12-q32 deletions are more common in AML and myeloproliferative neoplasm (MPN), respectively. By ChromoSeq, the deletion in our patient was precisely determined to be 11.5 Mb in size involving a more distal interval, 13q14.11-q14.3 instead of 13q12-q14 (Table 1). Furthermore, ChromoSeq also identi ed a ~9.5Mb interstitial deletion on the short arm of chromosome 12 (p12.3p13.31) ( Table 1) that is not detectable in G-banded chromosomes at a band level of 400. This deletion involves 19 OMIM genes, including ETV6. Deletions and structural alterations involving 12p that includes ETV6 have been described to exhibit tumor-suppressor characteristics and that loss of ETV6 plays a role in leukemogenesis and possibly prognosis [9].
In addition to structural and copy number variations, both MyeloSeq (Methods S2) [10] and ChromoSeq assays (i.e., next generation targeted and wholegenome sequencing, respectively) also identi ed two premature stop codons in WT1 and STAG2 (Table 1); two genes with known implications in AML [11,12]. Both were classi ed as variants of uncertain clinical signi cance following standard guidelines provided by Association for Molecular Pathology, American Society of Clinical Oncology, and College of American Pathologists [13].
The overall combined genomic studies in this case revealed a complex karyotype with aberrations detected by karyotype, FISH, MyeloSeq; and ChromoSeq. Importantly, the use of a single clinical genome-wide sequencing assay along with a streamlined analytical approach (ChromoSeq) not only identi ed the aberrations initially identi ed by conventional genomic studies, but further resolved the correct gene fusion partners involved in t(8; 11), and uncovered another large deletion (12p-).
Based on the presence of MDS-related changes in the bone-marrow along with a complex karyotype (>3 abnormalities) and the identi cation of the previously discussed rearrangements/variants, the patient began induction therapy for high-risk AML with MDS-related changes using liposomal daunorubicin and cytarabine (VYXEOS ® ). The patient achieved a complete remission with the day-30 post-induction bone marrow biopsy demonstrating a mildly hypocellular marrow with multilineage dysplasia, but no diagnostic features of acute leukemia, and no increase in blasts by ow cytometry. In agreement with these ndings, her repeat MyeloSeq assay did not reveal any of the previously identi ed variants. In addition, the karyotype was normal and there was only a low level NUP98 rearrangement (0.5%) detected by FISH, with no evidence of deletion in the long arm of chromosome 13. The patient was subsequently treated with 2 consolidation cycles of VYXEOS ® with a sustained remission and the ultimate plan to pursue consolidation allogeneic hematopoietic cell transplantation, if possible.

Discussion And Conclusions
In our review of the literature, we identi ed only six documented cases of t(8;11) prior to this report occurring in AML and MDS [1][2][3][4][5][6]. One of these cases showed FGFR1 and NUP98 rearrangements by FISH [6], ndings similar to what we initially observed in this AML patient. The suspected NUP98/FGFR1 fusion, however, was not veri ed by other techniques, leaving the possibility of a different partnership with a different gene at an adjacent locus. Of note, the WHO (2016) [14] has classi ed any FGFR1 rearrangements, regardless of partner, into "myeloid and lymphoid neoplasm with FGFR1 abnormalities" which is also known as 8p11 myeloproliferative syndrome. This diagnosis represents a rare, generally aggressive, and clinically heterogeneous class of hematologic malignancies uni ed by the presence of the FGFR1 rearrangement. Neither our patient nor the patient reported by Sohal et al.
[6] showed features of an MPN. Based on combined FISH and RT-PCR results, the ve remaining documented cases were ultimately determined to have NUP98/NSD3 chimeric fusions, similar to our current case. However, the breakpoints reported by previous studies were intron 3 and intron 11 [1], while in our patient the breakpoint occurred in intron 4 of NSD3. Therefore, it is likely that none of the reported cases of AML and MDS contained an FGFR1 rearrangement. Moreover, it is possible that additional unreported cases classi ed as "myeloid and lymphoid neoplasms with FGFR1 abnormalities" may, in fact, represent a genetically heterogenous cohort that does not have an FGFR1 abnormality and in which at least some of the clinical variability and poor prognosis may be attributable to chromosomal aberrations which go unidenti ed and uncharacterized in the absence of broader, more precise genomic pro ling.
This case highlights the biologically relevant implications of identifying and resolving such aberrations. The gene NUP98 (nucleoporin 98 kDa) encodes a central component of the multi-peptide nuclear pore complex, which regulates RNA tra cking between the nucleus and cytoplasm, as well as playing a role in transcriptional regulation and cell-cycle progression [15][16][17]. To date, >30 NUP98-partner fusion genes have been described in association with a number of myeloid neoplasms, including AML and MDS [18]. Additional studies have con rmed the presence of a NUP98 gene fusion de nes a high-risk subset of leukemia with a poor prognosis [19]. Variation within the fusion partner may further modulate this risk, as does the co-occurrence of additional mutations including FLT3-internal tandem duplication (ITD) and WT1. Interestingly, the frequency and types of NUP98 fusion partners also varies by disease subset. For example, NUP98/NSD1 fusions have been associated with myelomonocytic leukemias while NUP98/KDM5A are more commonly observed in monocytic, erythroid and megakaryoblastic leukemias [20,21]. Meanwhile, NSD3, the fusion partner identi ed in this case, encodes a histone methyltransferase which has been found to be transcriptionally activated with NUP98, and has been associated with MDS-related AML as well as leukemogenic transformation of de novo and treatment-related AML [1] (reviewed by Alexander et al, 2016 [22]). Thus, identi cation of the speci c NUP98/NSD3 fusion in this case reveals key biologic underpinnings of the disease and may allow greater ability to risk stratify and manage patients such as this in the future.
Furthermore, identifying the presence of a NUP98-fusion AML may also allow for the development of rational, targeted treatment approaches that leverage our growing understanding of the drivers of disease biology. Given this patient's MDS-related changes and complex karyotype, in addition to the poor prognosis associated with NUP98 gene fusions, an induction and consolidation treatment approach for adverse-risk AML was undertaken with ultimate plans for an allogeneic hematopoietic cell transplantation in rst remission, if possible. However, there are a number of potential agents which may be relevant in targeting the disease biology of NUP98-fusion AML. Based on the role of NUP98 in regulating RNA and peptide tra cking to and from the nucleus, as well as the importance of transcriptional regulation of additional genes by the NUP98 gene-fusion product, the use of selinexor, an XPO1 inhibitor, may disrupt the nuclear tra cking and impair expression of gene products relevant to the leukemogenesis driven by the NUP98 gene-fusion product [23]. Additional pre-clinical studies using various sequencing techniques have also suggested the therapeutic potential for BCL2 inhibition, and JAK-STAT inhibition, in targeting key signaling pathways operative in the disease biology of NUP98-fusion AML [24,25]. Meanwhile, given the observation of increased co-occurrence of FLT3-ITD alterations in the context of NUP98-fusion AML, the use of any number of FLT3 inhibitors may be a consideration [25,26]. In addition, in the setting of NSD3 rearrangements, the development of histone methyltransferase inhibitors or the use of currently available histone deacetylase inhibitors (HDACi) may represent another potential treatment approach [27].
In summary, although NUP98-NSD3 chimeric mRNA had been previously reported in rare cases of AML, to the authors' knowledge, this case represents the rst report of the detection of this fusion using a genome sequencing approach. As such, this case highlights the ability of clinical genome sequencing to collectively provide a comprehensive understanding of the AML genetic architecture consistent with conventional approaches while also allowing base-pair level resolution of the t(8;11) with precise identi cation of the translocation partners. Notably, this observation may have particular relevance in cases of AML or MDS when FISH is consistent with the presence of an FGFR1 rearrangement, which may in fact be an NSD3 rearrangement. More broadly, the application of genome sequencing may result in a more complete understanding of genetic underpinnings of each patient's disease, and thus allow clinicians the opportunity to more accurately risk stratify and tailor treatment strategies [10]. The patient provided written informed consent for the use of her health information in this publication, which is available upon request.

Availability of data and materials
Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study. Any patient speci c information or data is not publicly available due to patient privacy and protection of health information.

Competing interests
The authors declare that they have no competing interests. JS participated in the clinical care of the patient, analyzed/interpreted patient data, and assisted in the conceptual planning, edited, and approved the nal manuscript.
JF participated in the clinical care of the patient, edited, and approved the nal manuscript.
ED analyzed/interpreted patient data, read and approved the nal manuscript.
DS analyzed/interpreted patient data, read and approved the nal manuscript.
JN analyzed and interpreted data, edited, and approved the nal manuscript.
KL participated in the clinical care of the patient, read and approved the nal manuscript.
IA conceptualized and designed the study, analyzed and interpreted data, participated in drafting and nal editing of the manuscript.

Figure 1
Results from karyotype, interphase and sequential metaphase FISH, and ChromoSeq analyses.