P15;P15) Is a Distinct Cytogenetic Entity with Poor Outcome and a Distinct Mutation Proﬁle: Comparative Analysis of 493 Adult Patients

Acute myeloid leukemia bearing t(7;11)(p15;p15) is a distinct cytogenetic entity with poor outcome and a distinct mutation proﬁle: comparative analysis of 493 adult patients

W-C Chou1,2,C-YChen2, H-A Hou2, L-I Lin3,J-LTang2,MYao2,WTsay2, B-S Ko2,S-JWu2, S-Y Huang2, S-C Hsu1, Y-C Chen1, Y-N Huang1, M-H Tseng2,C-FHuang2 and H-F Tien2

1Department of Laboratory Medicine, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan; 2Department of Internal Medicine, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan and 3Graduate Institute of Medical Biotechnology, College of Medicine, National Taiwan University, Taipei, Taiwan

Acute myeloid leukemia (AML) with t(7;11)(p15;p15), which marrow.4–8 Not only was this translocation seen in de novo results in a NUP98–HOXA9 fusion, is a distinct entity, but this AML but also, although rarely, in chronic myeloid leukemia and subtype has not been characterized in detail. In a comprehen- myelodysplastic syndrome.9–15 The translocation was found to sive study comparing 11 such patients with another 482 adult 16,17 patients, we found that those with t(7;11) were younger result in a fusion between NUP98 and HOXA9. Subsequent (P ¼ 0.0076) and female (P ¼ 0.0111), with almost all having the studies showed that several other HOX genes could be fused M2-subtype of AML (Po0.0001). Even when those with low-risk with NUP98 in myeloid leukemia.18–22 A number of animal karyotypes were excluded, patients with t(7;11) had poorer studies have described the pathogenetic role of NUP98–HOXA9 overall survival than the other AML group (median 13.5 and 20 in leukemogenesis.23–27 months, respectively, P ¼ 0.045) and poorer relapse-free survi- Reports concerning this translocation have been limited by val (median 6 and 12 months, respectively, P ¼ 0.003). The NUP98–HOXA9 fusion was strongly associated with KRAS and the fact that they generally included only a small number of WT1 mutations (P ¼ 0.015 and P ¼ 0.0018, respectively). We patients, did not compare this group of patients directly with characterized four varieties of this fusion, among which NUP98 other AML patients, and offered limited clinical description. exon 12/HOXA9 exon 1b was present in all 11 patients. We In the current analysis of 493 adult AML patients, we found 11 developed a highly sensitive and specific assay to quantify the patients bearing t(7;11)(p15;p15). To better characterize this abundance of leukemic cells, and found that the fusion remained detectable in morphological complete remission, subgroup of patients, we performed a comprehensive compar- even after allogeneic stem cell transplantation, suggesting that ison of clinical and biological characteristics between these and this disease was highly refractory to very intensive treatment. the remaining AML patients. We determined that AML with AML with NUP98–HOXA9 fusion therefore appears to have a t(7;11)(p15;p15) was a distinct cytogenetic and clinical entity distinct clinical and biological profile, and should be regarded that was closely associated with FAB M2-subtype, female sex, as a poor prognostic group. younger age, mutations of WT1 and KRAS, and poor outcome. Leukemia (2009) 23, 1303–1310; doi:10.1038/leu.2009.25; published online 19 February 2009 We also devised a highly sensitive and specific real-time PCR Keywords: NUP98–HOXA9; acute myeloid leukemia; assay to quantify and trace this subset of leukemic cells, and t(7;11)(p15;p15) found that this disease was highly refractory to intensive treatment, including allogeneic stem cell transplantation, as shown by the high copy number of NUP98–HOXA9 fusion transcripts in most samples obtained in complete remission. To the best of our knowledge, this is the most comprehensive study Introduction to date on de novo adult AML patients with t(7;11).

Acute myeloid leukemia (AML) is a heterogeneous group of diseases in terms of clinical presentation, genetic alterations and Materials and methods response to treatment. Accumulating evidence has identified numerous parameters that are able to stratify this complex Patients disease. Traditionally, the karyotype of leukemic cells is one of Between 1994 and 2007, we recruited 493 consecutive adult the most important prognostic factors. A recent focus on the patients newly diagnosed with de novo AML. The diagnosis genetic alterations in AML has further contributed to our ability was based on a complete examination of hematology, serum to distinguish subgroups of this heterogeneous entity. biochemistry, bone marrow cell morphology, cytochemical Acute myeloid leukemia with chromosomal translocation staining, immunophenotyping and chromosomal analysis. All 1–3 t(7;11)(p15;p15) was first reported more than 20 years ago. patients signed informed consent before their leukemic cells Subsequent studies indicated that this subtype of AML was a were collected for diagnostic and molecular studies. specific entity occurring preferentially in Asian countries, and usually presented with dyspoietic cell morphology and considerable numbers of mature myeloid cells in the bone Immunophenotyping To characterize the phenotype of the leukemic cells, we used a Correspondence: Dr H-F Tien, Department of Internal Medicine, panel of monoclonal antibodies to the following: myeloid- National Taiwan University Hospital, National Taiwan University, associated antigens, including CD13, CD33, CD11b, CD15, Taipei 100, Taiwan. E-mail: [email protected] CD14 and CD41a; lymphoid-associated antigens, including Received 27 October 2008; revised 2 January 2009; accepted 22 CD2, CD5, CD7, CD19, CD10 and CD20; and the lineage January 2009; published online 19 February 2009 nonspecific antigens HLA-DR, CD34 and CD56. Expression of AML with NUP98–HOXA9 W-C Chou et al 1304 surface antigens on leukemic cells was detected by an indirect Results immunoalkaline phosphatase method28 before 1998, and by flow cytometry thereafter.29 Clinical and genetic characteristics of patients with t(7;11)(p15;p15) Among 493 AML patients, 11 patients (2. 23%) were found to Cytogenetic analysis have t(7;11)(p15;p15). The detailed laboratory data and clinical Bone marrow samples were aspirated into heparinized syringes, outcome of these patients are listed in Table 1. The marrow of and chromosomal analyses by the G-banding method were patients in this group contained a mixture of blasts and more performed after 1–3 days of unstimulated culture as described mature myeloid cells. Most of the patients presented with 28 earlier. dyspoietic features and Auer rods in the leukemic cells. All but one patient had the FAB M2-subtype. Compared with the other AML patients, those bearing this translocation were younger Analysis of genetic mutations (P ¼ 0.0076) and predominantly female (P ¼ 0.0111), and Bone marrow mononuclear cells were isolated using Ficoll- tended to have a higher lactate dehydrogenase level Hypaque gradient centrifugation. The isolated cells were frozen (P ¼ 0.0553) (Table 2). The immunophenotyping profile of and stored at À80 1C for further genetic analysis. We analyzed patients in this group was not different from that of the others mutations in several genes, including NPM1, FLT3/ITD, (data not shown). Four of the 11 patients with t(7;11) had FLT3/TKD, AML1, KRAS, NRAS, CEBPA, MLL/PTD, JAK2, concurrent WT1 mutations. Some patients with t(7;11) also had PTPN11 and WT1. The detailed experimental protocol, with FLT3/ITD, and mutations in NRAS and KRAS, each of which is a primer sequences, has been described earlier.29–33 class I mutation that confers a proliferation or survival advantage to hematopoietic cells. In contrast, none of the 11 patients had Analysis of NUP98–HOXA9 break points MLL/PTD or mutations in AML1, NPM1 or CEBPA, which belong to class II gene mutations that affect transcription factors From patients with t(7;11), cDNA was amplified by NUP98-F 35,36 (50-ACTGGGACAGCATTTGGAAC-30) and HOXA9-R (50-CA and lead to impaired hematopoietic differentiation (Table 1). GTTCCAGGGTCTGGTGTT-30). A final PCR volume of 50 ml We note that we could not exclude the possibility that mutations of other class II genes which we did not analyze in this study contained cDNA, 200 nM dNTPs, 1.5 mM MgSO4, 200 nM each primer and 2 units of Platinum Taq polymerase (Invitrogen, were present in these patients. Further analysis showed that Carlsbad, CA, USA). The PCR program was carried out at 95 1C t(7;11)(p15;p15) was positively associated with mutations in for 10 min followed by 35 cycles of 95 1C for 30 s, 53 1C for 30 s WT1 and KRAS (Table 2). For patients who were younger than and 72 1C for 30 s, with a final step of 10 min at 72 1C. Then 1 ml 65 years and had received anthracycline and cytarabine- of the PCR product was amplified further by a pair of internal containing intensive chemotherapy (n ¼ 326), those carrying primers, NUP98-F-int (50-TGGGACAGCATTTGGAACAG-30) t(7;11) (n ¼ 10) had worse overall survival (median 13.5 vs 38 and HOXA9-R-int (50-GGGCACCGCTTTTTCCGAGT-30), using months, P ¼ 0.009), relapse-free survival (median 6 vs 15 the same protocol as the earlier amplification. The products months, Po0.0001) (Figures 1a and b respectively) and were resolved on 2% agarose gels, after which the products disease-free survival (median 5 vs 11.7 months, P ¼ 0.003, data were purified and sent out for direct sequencing. not shown). However, most patients with t(7;11)(p15;p15) had achieved a complete remission after intensive chemotherapy (Table 1). Even when we excluded 82 patients with good-risk Real-time PCR for NUP98–HOXA9 karyotypes (inv(16), t(15;17) and t(8;21)), patients with t(7;11) The sequences of the forward and reverse primers and probe, still had with poorer overall survival (median 13.5 vs 20 months, respectively, were as follows: 50-GATTTAATACTACGACAGCC P ¼ 0.045), relapse-free survival (6 vs 12 months, P ¼ 0.003) ACTTTGG-30;50- GCTGGGTTGTTTTTCTCTATCAACT-30; and (Figures 1c and d respectively) and disease-free survival (5 vs 9 50-TTTGGAGCCCCCCAGGCCC-30. The probe was labeled months, P ¼ 0.027, data not shown) compared with other with FAM and TAMRA (Applied Biosystems, Foster City, CA, patients. Multivariate analysis indicated that this translocation USA). In each tube, 20 ml of reaction mixture contained 500 nM was an independent factor predicting shorter overall survival, each primer, 250 nM probe, 10 ml master mix (FastStart TaqMan relapse-free survival and disease-free survival (P ¼ 0.015, 0.001 Probe Master, Roche Diagnostics, Mannheim, Germany) con- and 0.003, respectively). The survival differences were equally taining 4.5 mM magnesium ion, and 4 ml of cDNA. The PCR significant when we restricted our analysis to patients with the program was carried out at 95 1C for 10 min followed by 45 M2-subtype of AML (6 vs 15 months, P ¼ 0.001 for relapse-free cycles of 95 1C for 10 s and 60 1C for 1 min on an iQ5 machine survival; 13 vs 38 months, P ¼ 0.040 for overall survival; and (BioRad, Hercules, CA, USA). Each sample was tested indepen- 6 vs 12 months, P ¼ 0.007 for disease-free survival). dently at least twice. The human phosphoprotein (HUPO) gene (Applied Biosystems) was used as a loading control.34 The plasmid used to generate the standard curve was made by TA cloning of the nested PCR product (Yeastern Biotech, Taipei, Four types of fusion break points in NUP98–HOXA9 Taiwan). Earlier reports have described three types of fusion sequence, which appeared to result from alternative splicing of NUP98 and HOXA9.7,12,16,17 In this study, we found four types of fusion, Statistics each of which was in-frame (Figure 2). Leukemic cells from all The w2 test was used to calculate the significance of association patients with t(7;11) contained at least two fusion products of between t(7;11) and other discrete parameters, such as expres- different intensity (Figure 2). Sequencing showed the longest sion of antigens, cytogenetics or mutation of a specific gene. The transcript and major product in all patients to be a fusion Mann–Whitney test was used to compare continuous variables. between NUP98 exon 12 and HOXA9 exon 1b, with the other The Kaplan–Meier curve was calculated using SPSS software representing a fusion between NUP98 exon 11/12 and HOXA9 (Chicago, IL, USA). exon 1b/2 (Figure 2).

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1305 Use of real-time PCR to quantify and trace the leukemic cells PTD. b To quantify and trace leukemic cells in patients during their

MLL/ clinical course, we designed a pair of primers and a probe for real-time PCR. The ampliﬁed region, which covered the and CT and DLI Outcome junction between exon 12 of NUP98 and exon 1b of HOXA9, was clearly present in each of the 11 patients (Figure 2). We AML1

, tested the efﬁciency of ampliﬁcation by serially diluting test plasmids into a complex cDNA background derived from NB4 +5/+9 Currently in CR

(months) cells lacking the target fusion gene, and found the method to be NPM1 , highly sensitive, with the ability to detect as few as 10 copies with nearly 100% ampliﬁcation efﬁciency (Figure 3a). No CEBPA

À signals were seen when cDNA of NB4 cells alone was ampliﬁed /NA 0/15.5 Died with AML /NA 0/16 Died with AML /NA 0/16.5 Died with AML +/++/++/+ 7/29.5+/++/+ Died 5/13 with AML 9/11+/+ 3/5 Died 3.5/+5 with AML Died with AML Alive 7/10 Died by with salvage AML Died with AML +/+ 6/13.5 Died with AML +/ À À À (Figure 3a). We also serially diluted one patient’s cDNA (patient

CR/relapse RFS/OS no. 9) into a complex cDNA background from NB4 cells, and ydrogenase; M, male; NA, not applicable; OS, found a similarly robust and specific amplification (Figure 3b). We then quantified the abundance of leukemic cells from four patients who had serial follow-up samples taken along their WT1 WT1 WT1 WT1 treatment course (Figure 4). Interestingly, we found that neither intensive chemotherapy nor stem cell transplantation had appreciably reduced the tumor burden of any of the four a patients (Figure 4). Of the 30 samples tested, only 2 samples FF F F F FF yielded no detectable signal (Figure 4b, patient no. 11 and Figure 4b, patient no. 6). Most samples contained fewer than 5% Mutation blasts in the bone marrow, but had evident residual disease as

NRAS shown by real-time PCR (Figure 5). KRAS , ITD, ITD ITD Class IFLT3/ ClassNRAS IIKRAS Other FLT3/ FFF FF FFF FFF FLT3/ FFF FFF Discussion . Class II mutations studied in this report include

Although myeloid leukemia with t(7;11)(p15;p15) was ﬁrst

PTPN11 reported more than 20 years ago, comprehensive clinical and À À À À biological analyses necessary for its detailed characterization /+ /+ / and À À À rods are still lacking. To the best of our knowledge, this study Marrow includes the largest number of patients with t(7;11) to date. We JAK2 dyspoiesis/Auer

, carried out a comprehensive analysis of the clinicopathological

KIT characteristics, genetic alterations and fusion break points at , presentation in patients with t(7;11), and developed a highly sensitive and speciﬁc real-time PCR assay to quantify the KRAS

, number of residual leukemic cells. We also compared AML patients with and without this translocation. NRAS

blast/prom (%) Significantly more female than male patients were found to bear t(7;11) (Table 2), a finding which differs with the results of 3 TKD, an earlier study. However, grouping the patients in this study with all of those reported earlier confirms that female patients FLT3/ with t(7;11) are more numerous than males (18 vs 13).3,7–12,14,15 We cannot exclude the possibility that the discrepancy with one ITD, earlier reported was related to different definitions of the

FLT3/ molecular entities in question, or to a difference in the ethnic LDH (U/l) FAB Marrow origins of the respective patient populations. Other prominent features in this group of patients were younger age, more l)/

m frequent FAB M2-subtype and poor outcome. In an important study of NUP98 rearrangements in hematopoietic malignancies,37 the demographic analysis of patients with t(7;11) did not l)/PLT (/

m show a tendency toward female sex or young age. However, in that study the molecular level of the rearrangement was PB blast (%) incompletely deﬁned, in contrast to our study where only the Hb (g per 100 ml)/ WBC (/ NUP98–HOXA9 fusion was investigated. Thus, molecular heterogeneity, disease rarity and ethnic differences might contribute to the observed discrepancies in the reported

Clinical, pathological and genetic features of AML patients with t(7;11)(p15;p15) epidemiology of AML with t(7;11). gender Although most patients achieved a morphological complete remission after chemotherapy, all but one relapsed (Table 1). This high relapse rate indicated that a considerable number of 123 40/F5 31/F6 25/M7 6.2/213k/33k/63 7.3/24k/28k/52 59/F8 8.9/16k/51k/65 23/F9 2970 27/F 6.8/30k/14k/62 714 7.6/17k/65k/47 28/F 3630 M4 9.8/33k/82k/37 57/F 2860 M2 M2 8.8/8.7k/73k/61 1517 29.1/24 9.2/7.2k/10k/13 1516 M2 63.6/13.4 16.2/26 M2 545 M2 711 31/26.5 49.6/6.4 +/+ M2 M2 34/34 +/ 50.2/23.6 30.6/11.8 +/+ +/+ +/ 4 29/M 8.4/36k/16k/84 1500 M2 52/22.2 +/+ Patient nos. 6, 7, 8 and 11 had undergone hematopoietic stem cell transplantation. Class I mutations studied in this report include Table 1 Patient Age/ a b 10 37/F 6.0/15k/92k/18 2638 M2 23.5/27.4 +/ 11Abbreviations: AML, acute myeloidoverall leukemia; survival; CR, PB, complete peripheral remission; blood; CT, 55/F chemotherapy; prom, DLI, promyelocyte; donor RFS, lymphocyte relapse-free infusion; survival. F, female; LDH, lactate deh 5.0/5.8k/9k/6morphologically 714 M2 undetectable 55.6/16leukemic +/+ cells remained after

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1306 Table 2 Comparisons of clinical and biological data between AML patients with and without t(7;11)

Variant t(7;11)(p15;p15) (N ¼ 11, 2.23%) Others (N ¼ 482, 97.77%) P-value

N %(N /11) N %(N /482)

Age, median (range) 31 (23–59) 52.5 (16–90) 0.0076

Gender (N, % ¼ N/493) 0.0111 Male (N ¼ 283, 57.40%) 2 18.18 281 58.30 Female (N ¼ 210, 42.60%) 9 81.82 201 41.70

Laboratory data (median) LDH (U/l) 1516 838 0.0553 WBC (/ml) 16710 19440 0.7137 Blast (/ml) 10108 7239 0.8499 Platelet Â 1000/ml 33 42 0.4340 Hemoglobin (g per 100 ml) 7.6 8.1 0.5124

FAB (N, % ¼ N/493) M0 (N ¼ 9, 1.83%) 0 0.0 9 1.87 M1 (N ¼ 109, 22.11%) 0 0.0 109 22.61 M2 (N ¼ 166, 33.67%) 10 90.91 156 32.37 Po0.0001 M3 (N ¼ 39, 7.91%) 0 0.0 39 8.09 M4 (N ¼ 124, 25.15%) 1 9.09 123 25.52 M5 (N ¼ 24, 4.87%) 0 0.0 24 4.98 M6 (N ¼ 12, 2.43%) 0 0.0 12 2.49 M7 (N ¼ 0, 0.0%) 0 0.0 0 0.0 Undetermined (N ¼ 10, 2.03%) 0 0.0 10 2.07

Mutation (N, % ¼ N/493)a P-value CEBPA (N ¼ 66, 13.39%) 0 0.0 66 13.69 0.3743 FLT3/ITD (N ¼ 115, 23.33%) 3 27.27 112 23.24 0.7238 FLT3/TKD (N ¼ 40, 8.11%) 0 0.0 40 8.30 40.9999 AML1 (N ¼ 68, 13.79%) 0 0.0 68 14.11 0.3756 NPM1 (N ¼ 104, 21.10%) 0 0.0 104 21.58 0.1312 NRAS (N ¼ 39, 7.91%) 2 18.18 37 7.68 0.2135 KRAS (N ¼ 9, 1.83%) 2 18.18 7 1.45 0.015 KIT (N ¼ 11, 2.23%) 0 0.0 11 2.28 40.9999 MLL/PTD (N ¼ 24, 4.87%) 0 0.0 24 4.98 40.9999 JAK2 (N ¼ 4, 0.81%) 0 0.0 4 0.83 40.9999 PTPN11 (N ¼ 18, 3.65%) 0 0.0 18 3.73 40.9999 WT1 (N ¼ 27, 5.48%) 4 36.36 23 4.77 0.0018 Abbreviations: AML, acute myeloid leukemia; LDH, lactate dehydrogenase; WBC, white blood cell. aThe incidence of gene mutations in patients with a normal karyotype is 22.2% for CEBPA, 32.4% for FLT3/ITD, 38.2% for NPM1 and 5.8% for WT1.

chemotherapy, consistent with the results of our real-time PCR HOX family genes, to complete formation of the leukemia. In assay, which clearly showed the persistence of abundant disease this study, we used a novel method to identify genes that cells following intensive chemotherapy even when fewer than potentially cooperate with NUP98–HOXA9 in human leukemia. 5% of blast cells in the patients’ bone marrow remained. This Our genetic mutation analysis on human leukemic cells indicated that many leukemic cells bearing NUP98–HOXA9 indicated that the KRAS and WT1 mutations were significantly fusion presented with a misleadingly mature morphology. Thus, associated with this fusion gene, implying a critical role of these morphological assessment of marrow blast percentage alone is two genetic alterations in cooperation with NUP98–HOXA9 in insufficient to allow an accurate assessment of disease status the pathogenesis of this subtype of leukemia. Both KRAS and after chemotherapy. As NUP98–HOXA9 plays a critical role in WT1 are known to be important factors in leukemic patho- leukemogenesis in this group of patients, it is most likely that this genesis. Mutation of WT1, which normally induces cellular fusion persists as long as the leukemic cells survive. Our real- quiescence and differentiation,39 occurs in 10% of AML patients time PCR assay detected signals in all but two samples obtained with a normal karyotype, and may be associated with drug from patients in complete remission, suggesting that this resistance.40 KRAS mutations occur in about 5% of AML mutation confers a refractory character on the disease, and patients.41 Although the prognosis of AML patients bearing may be a useful marker for quantifying minimal residual disease. KRAS mutations appears not to differ from that of others,41 these NUP98–HOXA9 has been found to cooperate with BCR-ABL mutations result in constitutive activation of the protein, which to accelerate leukemogenesis in animals.24 Several studies using would be expected to promote tumor cell proliferation. The microarrays,26,27 identification of retroviral integration sites38 or strong association between NUP98–HOXA9 and WT1 and co-transfection23 have identified genes that potentially co- KRAS mutations suggests that the latter two gene alterations operate with the NUP98–HOXA9 fusion during leukemia could interact with NUP98–HOXA9 in acceleration of leukemo- development in both animals and humans. It has been proposed genesis and in conferring a poor response to chemotherapy. that NUP98–HOXA9 might induce a preleukemic state, requir- Although WT1 upregulation was seen in primary human ing the activation of other oncogenes, such as MEIS1 and other CD34 þ hematopoietic cells transduced by NUP98–HOXA9,27

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1307

Figure 1 Kaplan–Meier survival analysis. Patients with t(7;11)(p15;p15) (n ¼ 10) and others (n ¼ 316) were analyzed for (a) overall survival (OS) (median 13.5 vs 38 months, P ¼ 0.009) and (b) relapse-free survival (RFS) (median 6 vs 15 months, Po0.0001). When 82 patients with good-risk karyotypes (inv(16), t(15;17) and t(8;21)) were excluded, patients with t(7;11)(p15;p15) (n ¼ 10) continued to have poorer OS (median 13.5 vs 20 months, P ¼ 0.045) (c) and RFS (6 vs 12 months, P ¼ 0.003) (d) compared with other acute myeloid leukemia patients (n ¼ 234).

Patient No. M26 4 10 1 35 9 7 11 8 H2O NUP98 ex. 12-HOXA9 ex. 1b (537 bp), type I NUP98 ex. 12-HOXA9 ex. 2 (447 bp), type II NUP98 ex. 11-HOXA9 ex. 1b (396 bp), type III NUP98 ex. 11-HOXA9 ex. 2 (306 bp), type IV

Type I Type II NUP98 ex. 12 HOXA9 ex. 1b NUP98 ex. 12 HOXA9 ex. 2

A PVVDRE APVDNPA

Type III Type IV NUP98 ex. 11 HOXA9 ex. 1b NUP98 ex. 11 HOXA9 ex. 2

F G T V D R E FG TD N P A

Figure 2 DNA sequence analysis of fusion points. (a) The products of nested PCR of the fusion transcripts were resolved on 2% gels. Four types of bands were observed, the most predominant being the largest band (NUP98 exon 12/HOXA9 exon 1b). The exons involved in the fusion and the expected length are indicated. M: marker. (b) Fusion point sequence in each type of transcript. Each fusion was in-frame. The corresponding amino acids are listed below the nucleotide sequence.

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1308 40 35 30 25

Ct 20 y = -3.4906x + 38.501 105 104 103 102 101 15 2 10 R = 0.994 5 0 1015 20 25 30 35 40 45 0123456 Plasmid copies diluted in NB4 cDNA Log Copies of plasmids in NB4 cDNA

30 y = 3.4935x + 22.251

2 3 4 Ct 20 10x 10 x10x10x R2 = 0.9901 10

0 0 1234 5 10 15 20 25 30 35 40 45 Fold of dilution in NB4 cDNA (log scale) t(7;11) patient cDNA diluted in NB4 cDNA

Figure 3 High sensitivity and specificity of real-time PCR. (a) Plasmids serially diluted to yield copy numbers of 105,104,103,102, 10 and 0 were mixed with a fixed amount of cDNA from NB4 cells (200 ng RNA equivalent) and quantified by real-time PCR (left). There was no detectable signal when only NB4 cDNA was present, and the amplification curve showed nearly 100% amplification efficiency (slope À3.4906, right). (b)cDNA from patient no. 9 (200 ng RNA equivalent) was serially diluted into cDNA of NB4 cells (between 10- and 104-fold) and quantified by real-time PCR (left). The amplification efficiency was nearly 100% (slope À3.4935) and there was no signal when only NB4 cDNA was present in the reaction. 6 Patient No. 7 6 Patient No. 11 6 6 relapse 5 5 Sibling HSCT 4 4 (reduced intensity) 3 MUD HSCT 3 (myeloablative) HUPO 2 HUPO 2 1 1 0 0 012345678 Log NUP98-HOXA9/10 Log NUP98-HOXA9/10 024681012 Months from diagnosis Months from diagnosis 6 6 Sibling HSCT Patient No. 6 Patient No. 8 (myeloablative) 6 5 Relapse 5 4 Relapse 4 3 MUD HSCT 3 myeloablative

2 HUPO HUPO 2 1 DLI 1 AutoBMT 0 DLI 0 0 5 10 15 20

01020 30 40 50 Log NUP98-HOXA9/10 Log NUP98-HOXA9/10 Months from diagnosis Months from diagnosis

Figure 4 Fusion gene abundance over the course of treatment of four patients. All patients underwent stem cell transplantation (a–d). Only one sample in each of (b and c) lacked a detectable signal. In general, chemotherapy was only partially able to reduce the mutant signals. Arrows represent intensive chemotherapy. MUD, matched unrelated donor; HSCT, hematopoietic stem cell transplantation; DLI, donor lymphocyte infusion; AutoBMT, autologous bone marrow transplantation.

the mechanism for a potentially close association between the differentiation,42–44 suggesting that NUP98–HOXA9 might WT1 mutation and this fusion remain to be investigated. No belong to the class II group. Our finding that NUP98–HOXA9 earlier reports have shown a correlation between this fusion frequently occurred simultaneously with class I mutations such gene and KRAS mutations. as KRAS and FLT3/ITD, but not with the class II mutations, such The genetic events underlying AML pathogenesis appear to as AML1 and CEBPA that we analyzed (Table 1), supports the be classified into two broadly defined complementation hypothesis of multistep AML pathogenesis that may require the groups: class I, which includes FLT3 and RAS mutations; and cooperation of mutations in both classes.35 The role of WT1 class II, including the CEBPA and AML1 mutations.35,36 HOX mutation in leukemogenesis is still not completely defined. It genes are targets of polycomb genes that are important in cell may belong to another class of gene alteration.

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1309 6 transcript in myelodysplastic syndromes associated with t(7;11)(p15;p15). Br J Haematol 1999; 107: 600–604.

HUPO 5 10 Ahuja HG, Popplewell L, Tcheurekdjian L, Slovak ML. NUP98 6 gene rearrangements and the clonal evolution of chronic 4 myelogenous leukemia. Genes Chromosomes Cancer 2001; 30: 410–415. 3 11 Yamamoto K, Nakamura Y, Saito K, Furusawa S. Expression of the 2 NUP98/HOXA9 fusion transcript in the blast crisis of Philadelphia chromosome-positive chronic myelogenous leukaemia with 1 t(7;11)(p15;p15). Br J Haematol 2000; 109: 423–426. 12 Wong KF, So CC, Kwong YL. Chronic myelomonocytic leukemia

Log NUP98-HOXA9/10 0 with t(7;11)(p15;p15) and NUP98/HOXA9 fusion. Cancer Genet 0 102030405060 Cytogenet 1999; 115: 70–72. Blast percentage in bone marrow 13 Suzuki A, Ito Y, Sashida G, Honda S, Katagiri T, Fujino T et al. t(7;11)(p15;p15) Chronic myeloid leukaemia developed into Figure 5 Correlation between marrow blast percentages and blastic transformation showing a novel NUP98/HOXA11 fusion. quantified mutant load assessed by real-time PCR. The mutant signals Br J Haematol 2002; 116: 170–172. remained high even when the marrow blast percentage decreased 14 Kaneko Y, Maseki N, Takasaki N, Sakurai M, Kawai K, Sakurai M. below 5%. Possible association of a new translocation, t(7;11)(p15;p15), with Ph1 chromosome-negative chronic myelogenous leukemia. Int J Cancer 1985; 36: 657–659. 15 Takeda T, Ikebuchi K, Zaike Y, Mori M, Ohyashiki K, Ikeuchi T. From this comprehensive study, containing the largest Ph-negative chronic myelocytic leukemia with a complex translo- number of AML patients bearing t(7;11)(p15;p15) to date, we cation involving chromosomes 7 and 11. Cancer Genet Cytogenet conclude that this subtype of AML is a special entity with a 1986; 21: 123–127. number of distinct clinical and biological characteristics. 16 Nakamura T, Largaespada DA, Lee MP, Johnson LA, Ohyashiki K, Toyama K et al. Fusion of the nucleoporin gene NUP98 to HOXA9 Patients with this translocation should be suspected to have by the chromosome translocation t(7;11)(p15;p15) in human poor prognosis, which warrants the development of novel myeloid leukaemia. Nat Genet 1996; 12: 154–158. therapies directed specifically toward this population. In 17 Borrow J, Shearman AM, Stanton Jr VP, Becher R, Collins T, addition, their disease status should be monitored by mutation- Williams AJ et al. The t(7;11)(p15;p15) translocation in acute specific real-time PCR, as assessment based on morpho- myeloid leukaemia fuses the genes for nucleoporin NUP98 and logical observation appears to have the potential to be extremely class I homeoprotein HOXA9. Nat Genet 1996; 12: 159–167. 18 Nakamura T, Yamazaki Y, Hatano Y, Miura I. NUP98 is fused to misleading. PMX1 homeobox gene in human acute myelogenous leukemia with chromosome translocation t(1;11)(q23;p15). Blood 1999; 94: 741–747. Acknowledgements 19 Raza-Egilmez SZ, Jani-Sait SN, Grossi M, Higgins MJ, Shows TB, Aplan PD. NUP98-HOXD13 gene fusion in therapy-related This work was supported by Grants NSC 96-2628-B002-013-MY2 acute myelogenous leukemia. Cancer Res 1998; 58: 4269–4273. from National Science Council (Taiwan), NHRI-EX97-9731BI 20 Taketani T, Taki T, Ono R, Kobayashi Y, Ida K, Hayashi Y. The from National Health Research Institute (Taiwan) and YongLin chromosome translocation t(7;11)(p15;p15) in acute myeloid Healthcare Foundation. leukemia results in fusion of the NUP98 gene with a HOXA cluster gene, HOXA13, but not HOXA9. Genes Chromosomes Cancer 2002; 34: 437–443. References 21 Panagopoulos I, Isaksson M, Billstrom R, Strombeck B, Mitelman F, Johansson B. Fusion of the NUP98 gene and the homeobox gene 1 Ohyashiki K. Nonrandom cytogenetic changes in human acute HOXC13 in acute myeloid leukemia with t(11;12)(p15;q13). leukemia and their clinical implications. Cancer Genet Cytogenet Genes Chromosomes Cancer 2003; 36: 107–112. 1984; 11: 453–471. 22 Nishiyama M, Arai Y, Tsunematsu Y, Kobayashi H, Asami K, Yabe 2 Okada M, Mizoguchi H, Kubota K, Nomura Y. Quantitative M et al. 11p15 translocations involving the NUP98 gene in analysis of chromosomal G-bands in human hematopoietic childhood therapy-related acute myeloid leukemia/myelodysplas- disorders by methotrexate synchronization technique. Cancer tic syndrome. Genes Chromosomes Cancer 1999; 26: 215–220. Genet Cytogenet 1984; 13: 225–237. 23 Kroon E, Thorsteinsdottir U, Mayotte N, Nakamura T, Sauvageau 3 Sato Y, Abe S, Mise K, Sasaki M, Kamada N, Kouda K et al. G. NUP98-HOXA9 expression in hemopoietic stem cells induces Reciprocal translocation involving the short arms of chromosomes chronic and acute myeloid leukemias in mice. EMBO J 2001; 20: 7 and 11, t(7pÀ;11p+), associated with myeloid leukemia with 350–361. maturation. Blood 1987; 70: 1654–1658. 24 Mayotte N, Roy DC, Yao J, Kroon E, Sauvageau G. Oncogenic 4 Huang SY, Tang JL, Liang YJ, Wang CH, Chen YC, Tien HF. interaction between BCR-ABL and NUP98-HOXA9 demonstrated Clinical, haematological and molecular studies in patients with by the use of an in vitro purging culture system. Blood 2002; 100: chromosome translocation t(7;11): a study of four Chinese patients 4177–4184. in Taiwan. Br J Haematol 1997; 96: 682–687. 25 Dash AB, Williams IR, Kutok JL, Tomasson MH, Anastasiadou E, 5 Kwong YL, Liu HW, Chan LC. Racial predisposition to transloca- Lindahl K et al. A murine model of CML blast crisis induced by tion (7;11). Leukemia 1992; 6: 232. cooperation between BCR/ABL and NUP98/HOXA9. Proc Natl 6 Kwong YL, Chan TK. Translocation (7;11)(p15;p15) in acute Acad Sci USA 2002; 99: 7622–7627. myeloid leukemia M2: association with trilineage myelodysplasia 26 Chung KY, Morrone G, Schuringa JJ, Plasilova M, Shieh JH, Zhang and giant dysplastic myeloid cells. Am J Hematol 1994; 47: 62–64. Y et al. Enforced expression of NUP98-HOXA9 in human CD34(+) 7 Kwong YL, Pang A. Low frequency of rearrangements of the cells enhances stem cell proliferation. Cancer Res 2006; 66: homeobox gene HOXA9/t(7;11) in adult acute myeloid leukemia. 11781–11791. Genes Chromosomes Cancer 1999; 25: 70–74. 27 Takeda A, Goolsby C, Yaseen NR. NUP98-HOXA9 induces long- 8 Inaba T, Shimazaki C, Yoneyama S, Hirai H, Kikuta T, Sumikuma T term proliferation and blocks differentiation of primary human et al. t(7;11) and trilineage myelodysplasia in acute myelomono- CD34+ hematopoietic cells. Cancer Res 2006; 66: 6628–6637. cytic leukemia. Cancer Genet Cytogenet 1996; 86: 72–75. 28 Tien HF, Wang CH, Lin MT, Lee FY, Liu MC, Chuang SM et al. 9 Hatano Y, Miura I, Nakamura T, Yamazaki Y, Takahashi N, Miura Correlation of cytogenetic results with immunophenotype, geno- AB. Molecular heterogeneity of the NUP98/HOXA9 fusion type, clinical features, and ras mutation in acute myeloid

Leukemia AML with NUP98–HOXA9 W-C Chou et al 1310 leukemia. A study of 235 Chinese patients in Taiwan. Cancer 36 Frohling S, Scholl C, Gilliland DG, Levine RL. Genetics of myeloid Genet Cytogenet 1995; 84: 60–68. malignancies: pathogenetic and clinical implications. J Clin Oncol 29 Lin LI, Chen CY, Lin DT, Tsay W, Tang JL, Yeh YC et al. 2005; 23: 6285–6295. Characterization of CEBPA mutations in acute myeloid leukemia: 37 Romana SP, Radford-Weiss I, Ben Abdelali R, Schluth C, Petit A, most patients with CEBPA mutations have biallelic mutations and Dastugue N et al. NUP98 rearrangements in hematopoietic show a distinct immunophenotype of the leukemic cells. Clin malignancies: a study of the Groupe Francophone de Cyto- Cancer Res 2005; 11: 1372–1379. genetique Hematologique. Leukemia 2006; 20: 696–706. 30 Chou WC, Tang JL, Lin LI, Yao M, Tsay W, Chen CY et al. 38 Iwasaki M, Kuwata T, Yamazaki Y, Jenkins NA, Copeland NG, Nucleophosmin mutations in de novo acute myeloid leukemia: the Osato M et al. Identification of cooperative genes for NUP98- age-dependent incidences and the stability during disease evolu- HOXA9 in myeloid leukemogenesis using a mouse model. Blood tion. Cancer Res 2006; 66: 3310–3316. 2005; 105: 784–793. 31 Chen CY, Lin LI, Tang JL, Ko BS, Tsay W, Chou WC et al. RUNX1 39 Ellisen LW, Carlesso N, Cheng T, Scadden DT, Haber DA. The gene mutation in primary myelodysplastic syndromeFthe muta- Wilms tumor suppressor WT1 directs stage-specific quiescence tion can be detected early at diagnosis or acquired during disease and differentiation of human hematopoietic progenitor cells. progression and is associated with poor outcome. Br J Haematol EMBO J 2001; 20: 1897–1909. 2007; 139: 405–414. 40 King-Underwood L, Pritchard-Jones K. Wilms’ tumor (WT1) gene 32 Hou HA, Chou WC, Lin LI, Chen CY, Tang JL, Tseng MH et al. mutations occur mainly in acute myeloid leukemia and may confer Characterization of acute myeloid leukemia with PTPN11 drug resistance. Blood 1998; 91: 2961–2968. mutation: the mutation is closely associated with NPM1 41 Bowen DT, Frew ME, Hills R, Gale RE, Wheatley K, Groves MJ mutation but inversely related to FLT3/ITD. Leukemia 2008; 22: et al. RAS mutation in acute myeloid leukemia is associated with 1075–1078. distinct cytogenetic subgroups but does not influence outcome in 33 Summers K, Stevens J, Kakkas I, Smith M, Smith LL, Macdougall F patients younger than 60 years. Blood 2005; 106: 2113–2119. et al. Wilms’ tumour 1 mutations are associated with FLT3-ITD and 42 Hanson RD, Hess JL, Yu BD, Ernst P, van Lohuizen M, Berns A failure of standard induction chemotherapy in patients with normal et al. Mammalian Trithorax and polycomb-group homologues are karyotype AML. Leukemia 2007; 21: 550–551. antagonistic regulators of homeotic development. Proc Natl Acad 34 Chou WC, Tang JL, Wu SJ, Tsay W, Yao M, Huang SY et al. Sci USA 1999; 96: 14372–14377. Clinical implications of minimal residual disease monitoring by 43 Dellino GI, Schwartz YB, Farkas G, McCabe D, Elgin SC, quantitative polymerase chain reaction in acute myeloid leukemia Pirrotta V. Polycomb silencing blocks transcription initiation. patients bearing nucleophosmin (NPM1) mutations. Leukemia Mol Cell 2004; 13: 887–893. 2007; 21: 998–1004. 44 Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C et al. 35 Kelly LM, Gilliland DG. Genetics of myeloid leukemias. Annu Rev The polycomb group protein EZH2 directly controls DNA Genomics Hum Genet 2002; 3: 179–198. methylation. Nature 2006; 439: 871–874.

Leukemia