Letters to the Editor 1657 an 8 base-pair guanine mononucleotide repeat sequence Conflict of interest makes this reported variant suspicious for an artifact of PCR amplification rather than a true somatic mutation. This is a The authors declare no conflict of interest. well-known phenomenon that is commonly seen as an artifact 1,2 1 1 after PCR amplification of a region of DNA with homopolymer O Abdel-Wahab , O Kilpivaara , J Patel , runs.6,7 Moreover, slipped-strand mispairing of the PCR L Busque3 and RL Levine1,2 1 polymerase resulting in this in vitro frameshift does not Leukemia Service, Memorial Sloan Kettering Cancer Center, 6,7 New York, NY, USA; necessarily occur in every PCR amplification product, which 2 explains the variable presence of this allele in different patients Human Oncology and Pathogenesis Program, Memorial Sloan Kettering Cancer Center, New York, NY, USA and as assessed by Sanger resequencing, which has limited 3Research Centre, Maisonneuve-Rosemont Hospital, sensitivity. To improve our ability to detect this variant, Montreal, Quebec, Canada we perfomed amplification of this region of ASXL1 followed E-mail: [email protected] by sensitive mass spectrometry (Sequenom, San Diego, CA, USA) to distinguish between PCR products with 8 versus 9 guanine nucleotides in paired tumor and normal DNA (Figure 1b). When we performed PCR amplification followed References by (Seqeunom, San Diego, CA, USA) mass spectrometry for this variant in 10 paired samples from samples with myelodysplastic 1 Gelsi-Boyer V, Trouplin V, Adelaide J, Bonansea J, Cervera N, syndrome, myeloproliferative neoplasm and chronic myelomo- Carbuccia N et al. Mutations of polycomb-associated ASXL1 in nocytic leukemia, this variant was detected in tumor and normal myelodysplastic syndromes and chronic myelomonocytic leukaemia. Br J Haematol 2009; 145: 788–800. DNA in every instance. This again strongly suggests that this 2 Boultwood J, Perry J, Pellagatti A, Fernandez-Mercado M, alteration is not a somatic mutation. Finally, we performed Fernandez-Santamaria C, Calasanz MJ et al. Frequent mutation of Sanger sequencing of ASXL1 in granulocyte DNA extracted from the polycomb-associated gene ASXL1 in the myelodysplastic 96 individuals with no evidence of any hematologic disorder. syndromes and in . Leukemia 2010; 24: The c.1934dupG p.Gly646TrpsfsX12 variant was readily appar- 1062–1065. ent in 425% of samples from patients without hematologic 3 Boultwood J, Perry J, Zaman R, Fernandez-Santamaria C, Littlewood T, Kusec R et al. High-density single nucleotide polymorphism array disease (Figure 1c). Although mutations in ASXL1 have been analysis and ASXL1 gene mutation screening in chronic myeloid reported in individuals without clinical evidence of a hemato- leukemia during disease progression. Leukemia 2010; 24: 1139–1145. logic disorder at the time of DNA acquisition, the fact that this 4 Carbuccia N, Murati A, Trouplin V, Brecqueville M, Adelaide J, sample is found repeatedly in paired tumor and normal DNA Rey J et al. Mutations of ASXL1 gene in myeloproliferative makes this unlikely to be a somatic mutation.8 neoplasms. Leukemia 2009; 23: 2183–2186. The findings reported above indicate that the most commonly 5 Carbuccia N, Trouplin V, Gelsi-Boyer V, Murati A, Rocquain J, Adelaide J et al. Mutual exclusion of ASXL1 and NPM1 mutations in reported mutation in ASXL1, to date, is not a somatic mutation. a series of acute myeloid leukemias. Leukemia 2010; 24: 469–473. Much of the literature on the mutational frequency and clinical 6 Fazekas A, Steeves R, Newmaster SG. Improving sequencing quality correlates of ASXL1 mutations should be reanalyzed with this in from PCR products containing long mononucleotide repeats. mind. The 8 mononucleotide guanine repeat sequence in the BioTechniques 2010; 48: 351–355. reference sequence for ASXL1 in this region may confound 7 Clarke LA, Rebelo CS, Goncalves J, Boavida MG, Jordan P. PCR delimitation of the true repeat number in this region. These amplification introduces errors into mononucleotide and dinucleo- tide repeat sequences. Mol Pathol 2001; 54: 351–353. findings also further highlight the importance of the use of paired 8 Abdel-Wahab O, Manshouri T, Patel J, Harris K, Yao J, Hedvat C et al. normal tissue for accurate detection of true somatic mutations Genetic analysis of transforming events that convert chronic myelo- rather than polymorphisms of PCR/sequencing artifacts. proliferative neoplasms to leukemias. Cancer Res 2010; 70: 447–452.

Expression of cell–cell interacting distinguishes HLXB9/TEL from MLL-positive childhood acute myeloid leukemia

Leukemia (2010) 24, 1657–1660; doi:10.1038/leu.2010.146; a separate cohort of MLL-negative infant AML characterized by published online 1 July 2010 an early disease onset (o2 years) as well as t(7;12) HLXB9/TEL ( ¼ MNX1/ETV6) rearrangement and with concomitant high Molecular characterization of leukemic cells is a continuously expression of HLXB9 (MNX1). Surprisingly, all patients relapsed emerging field and has become fundamental for therapy having a 3-year EFS of 0%.2,3 The role of HLXB9, a transcription stratification and prediction of event-free survival (EFS). Infant factor of the family of is rarely studied in acute myeloid leukemia (AML) is in 460% cases characterized hematopoiesis and the data regarding its ability to cause by a genomic rearrangement involving the mixed lineage malignant transformation of hematopoietic stem cells (HSCs) is leukemia (MLL) (11q23) and the expression of a fusion not yet available. Interestingly, germline mutations of HLXB9 (450 fusion partners are described). Patients lead to annorectal malformations and in with primary diagnosis of MLL-positive leukemia are young children, but hematopoietic abnormalities are not described.4 (o2 years) and have generally an inferior outcome compared The poor clinical outcome in this HLXB9/TEL-positive leukemia with MLL-negative patients.1 We and others recently described subset prompted us to comprehensively characterize the two

Leukemia Letters to the Editor 1658 t(7;12)

CCGACTTCAACTGCT TGCAGC CA

Actin HLXB9 Exon1 TEL Exon3

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 breakpoint

t(7;12) vs 11q23

down-regulated up-regulated

PBX3 EDIL3 HOXA6 CNTNAP5 HOXA5 ANGPT1 HOXA10 DSG2 HOXA3 ITGA9 HOXA4 ITGAV HOXA7 KDR HOXA9 SIGLEC6 HOXA2

HOX genes cell-cell interacting genes 123456123456 11q23 t(7;12)

HLXB9HOXA9 MEIS1 C-MYB HLXB9/ 0.20 0.10 0.50 HLXB9/ neg. 1.40 MLL/ENL HLXB9 TEL+ 0.40 TEL contr. 1.20 0.15 0.08 HLXB9 1.00 0.06 0.30 0.80 0.10 0.60 0.04 0.20 to actin 0.05 0.40 0.02 0.10 0.20

relative expression expression relative 0.00 0.00 0.00 0.00 CFU after 3rd replating t(7;12) 11q23*t(7;12) 11q23* t(7;12) 11q23* t(7;12) 11q23*

Figure 1 profiling and transformation process. Bone marrow or peripheral blood mononuclear cells were obtained from the leukemia laboratory, Giessen, Germany and informed consent was obtained from all families. (a) RT–PCR amplifying the HLXB9/TEL-fusion transcript. Indicated is the in-frame fragment variant at 194 bp as well as the out-of-frame variant at 330 bp (nested PCR primer: HLXB9-F1 50-CTTCCAGCTGGACCAGTGGCTG-30; TEL-R1 50-CTGAAGGAGTTCATAGAGCACATC-30; HLXB9-F2 50-CACCGCGGGCATGATCCTGC-30; TEL-R2 50-ATCGATAGCGAAAGTCCTCTT-30). Shown are 18 representatives out of 42 screened samples. (b) Sequence of the HLXB9/TEL breakpoint fusing HLXB9 exon 1 and TEL exon 3. (c) Microarray analysis was carried out on RNA samples (Trizol; Invitrogen, Darmstadt, Germany) of six HLXB9/TEL- and six MLL/AF9-positive patients using Human Gene 1.0 ST Array (Affymetrix). In all, 1554 significant differentially regulated genes are illustrated. Red indicates relative upregulation and green downregulation. (d) Listed are selected significantly (Po0,05) altered genes in t(7;12) patients compared with t(11q23) patients representing a group of downregulated HOX genes and upregulated cell–cell interacting genes. (e) Relative expression of HLXB9, HOXA9, MEIS1 and C-MYB in t(7;12) and 11q23 patients. Quantitative real-time PCR was carried out in triplicates using TaqMan primer probe assays (Applied Biosystems, Darmstadt, Germany). Samples were normalized to b-ACTIN and DCt values were calculated. A significant difference was found for HLXB9 expression (Po0,05, Mann–Whitney U-test). t(7;12)-positive AML: n ¼ 5; t(11q23) AML/ALL: n ¼ 8, nMLL-positive ALL and AML patients. (f) Bone marrow of 5-FU-treated wild-type mice was collected, transduced with a g-retroviral pMSCV vector with the transgenes MLL/ENL, HLXB9, HLXB9/TEL, HLXB9/TEL þ HLXB9 or an empty vector and plated on methylcellulose. After three rounds of replating, cells were analyzed for their ability to form colonies.

entities of MLL- and HLXB9/TEL-positive AML regarding their polymerase chain reaction, French-American-British (FAB) cellular morphology, transcriptional profile and transformation subtype M5), who met the following criteria: diagnosis of process. AML, ageo2 years, blast content 460%, no trisomy 21. The MLL-positive cohort comprised six patients (t(9;11); Immunologic analysis revealed AML with coexpression of MLL-AF9, verified by fluorescence in situ hybridization and T-cell-associated antigens CD4 (mean: 81.25%), CD7 (mean: 26%)

Leukemia Letters to the Editor 1659 Table 1 Patient characteristics

Age (years) Karyotype Morphology Immunology V(D)J CD34 CD117 HLXB9/TEL 0.4 47, XX, t(7;16) (q36;q12), +mar M2 CD4/CD7 Neg 72 68 0.7 48, XY, t(7;12)(q36;p13), +8,+19 AMLa CD4/CD7 Neg 55 36 0.6 47, XX, t(7;12)(q36;p13), +19 M2 CD7/CD56 Neg NA NA 0.2 47, XX, t(7;12)(q36;p13), +19 M0 CD4/CD7 Neg 86 86 0.75 47, XX, del(7)(q11.2B21), del(12)(p13), +? M0 CD4/CD7 Neg 81 73 0.4 47, XY, del(12)(q12), +19 M2 CD4/CD7 Neg 61 59

MLL 0.65 46, XX, add(1)(p22), À10, der(17)?i(17)(p10), +mar M5 CD4/CD56 NA 1 8 0.84 NA M5 CD4/CD56 NA 0 49 0.76 46, XY, t(9;11)(p22;q23) M5 NA NA NA NA 1.4 46, XX, t(2;15)(p25;q13)[5]/46, idem, t(9;11)(p22;q23)[6] NA CD4/CD7 NA 0 65 0.32 NA M5 NA Neg NA NA 1.54 46, XY, ins(9;11)(p22;q12q23.3)[8]/46, XY[7] M5 CD4/CD7 NA 1 14 Abbreviations: AML, acute myeloid leukemia; MLL, mixed lineage leukemia; NA, not available; Neg, negative.Clinical characteristics of six HLXB9/TEL- and six MLL-positive patients including age (years), karyotype, morphology, immunology and T-cell rearrangement. aNot further characterized type of AML; V(D)J: V(D)J recombination. or CD56 (mean: 42.8%). In a cohort of 42 MLL-negative patients platform (Affymetrix, Santa Clara, CA, USA).6 The most frequent gathered in Germany and selected by the same clinical and differentially expressed HOX genes in human MLL-positive criteria, we detected a HLXB9/TEL transcript in six patients leukemia are HOXA5 and HOXA9. MLL fusion proteins can (incidence of 14%). All karyotypes were associated with trisomy induce transcription from the HOXA9 and MEIS1 locus and 19 or an additional, not further characterized marker chromo- probably promote transformation, extensively shown in murine some, a hallmark of this leukemia entity. The junction of transplantation models.1 In contrast, almost nothing is known the rearrangement fused HLXB9 exon 1 and TEL exon 3 about the gene expression pattern of t(7;12)-positive leukemia. (CAACT/GCTTG, in frame) in all positive patients analyzed by Comparative gene expression profiling of leukemic blasts, reverse transcriptase–PCR, subcloning and sequencing. A collected from the indicated HLXB9/TEL and MLL patients, second out-of-frame variant fusing HLXB9 exon 1 and TEL exon revealed significant downregulation of HOX genes such as 2 (mRNA fusion: CAACT/CAGGA, out of frame in TEL) was PBX3, HOXA9, HOXA6, HOXA5, HOXA10, HOXA3, HOXA4, detected in a subset of patients (Figure 1a). Clinically four HOXA7, HOXA2 in HLXB9/TEL patients using the Affymetrix patients had relapse, received allogeneic bone marrow trans- platform (Figures 1c and d). Especially genes characteristic for plantation and still remain in remission. One patient died after the MLL-induced transformation process such as MEIS1, HOXA9 the fourth relapse and allogeneic transplantation. Clinical and C-MYB were downregulated in HLXB9/TEL-positive patients information of one patient is missing. Morphology was defined whereas HLXB9 was upregulated compared with MLL-positive as FAB subtype M0 or M2 and the immunophenotype showed and other, not further characterized entities of infant AMLs, as the expression of myeloid antigens CD13/CD33/CD15 and shown by quantitative RT-PCR (Figure 1e; data not shown). coexpression of T-cell-associated antigens CD7/CD4/CD56 This strongly suggests another transformation mechanism in (mean expression of CD7: 72%, CD4: 50%, CD56: 16%; HLXB9/TEL-positive hematopoietic precursors compared with Table 1). CD34 expression was significantly higher in HLXB9/ MLL-positive cells. TEL compared with MLL-positive patients (71 vs 0.5%; Interestingly, we observed upregulation of genes, implicated P ¼ 0.014, Mann–Whitney U-test) (own data and von Bergh in cell–cell interactions and cell adhesion such as EDIL3, et al.3). Similarly CD117 was highly expressed in HLXB9/TEL- CNTNAP5, ANGPT1, DSG2, ITGA9, ITGAV, KDR and SIGLEC6 positive patient cells (mean 64 vs 34%; Table 1). The in HLXB9/TEL-positive patients (Figures 1c and d). Expression biphenotypic features of the HLXB9/TEL-positive blasts and the levels of ANGPT1 and KDR, the ligands of TIE2 and VEGFR, are fact that early T-cell progenitors in the thymus retain myeloid of special interest because the related signaling pathways are potential directed us to study V(D)J recombination in the blast important for the maintenance of quiescent HSCs and a cells.5 Neither in immunofluorescence nor by amplification of physiologic HSC niche.7 Consistently, the HLXB9/TEL gene DH-JH, Db-Jb or T-cell receptor-d segments using multiplex- expression profile revealed expression of early hematopoietic multifluorescent PCR and GeneScan analysis, we succeeded in marker genes such as KITL and CD34, corresponding with the detecting an immature T-cell receptor rearrangement in the immunotypic findings. To further investigate the ability for analyzed patients (data not shown). This is arguing against malignant transformation of HLXB9/TEL or HLXB9 in compar- the hypothesis that the HLXB9/TEL-positive blasts arise from ison with MLL fusion proteins, we collected murine wild-type T-cell precursors, which have already started V(D)J recombina- HSCs from 5-fluoruracil-treated mice and transduced the cells tion in the thymus. with a g-retroviral vector expressing MLL/ENL, HLXB9, In view of the clinical and immunophenotypic similarities of HLXB9/TEL, HLXB9/TEL þ HLXB9 or an empty g-retroviral the two leukemia entities, such as young age at diagnosis, vector and subsequently plated the cells on methylcellulose. coexpression of myeloid and T-cell antigens and bad clinical After three rounds of replating, MLL/ENL-transduced HSCs outcome, we aimed to characterize the molecular footprint continued to form colonies, whereas cells overexpressing representing the two leukemia entities. Gene expression profiles HLXB9 and HLXB9/TEL did not, indicating a high self-renewal of MLL-positive acute lymphoblastic leukemia and AML are very capacity of MLL but not of HLXB9-orHLXB9/TEL-transduced well studied and a recent report described specific MLL-associated cells (Figure 1f). This result suggests that HLXB9 does not and upregulated genes, such as MEIS1 and HOX genes, in function by a block of HSC differentiation as it was previously discrepancy to MLL-negative leukemia, using the Affymetrix described for MLL fusion proteins. Rather, the inability of

Leukemia Letters to the Editor 1660 HLXB9 to induce transformation in vitro may be related to S Wildenhain1, C Ruckert2,SRo¨ttgers3, J Harbott3, 4 1 5 1 6 the fact that HLXB9 impinges more on the communication of W-D Ludwig , FR Schuster , K Beldjord , V Binder , R Slany , J Hauer1,7 and A Borkhardt1,7 HSCs with their natural environment, a fact that is also 1 suggested by the respective gene expression pattern, represent- Department of Pediatric Oncology, Hematology and Clinical Immunology, Center for Child and Adolescent Health, ing maintenance of the hematopoietic stemness character and Heinrich Heine University, Du¨sseldorf, Germany; niche formation. Upregulation of cell–cell and cell–matrix 2Department of Medical Informatics and Biomathematics, interacting molecules through HLXB9 may alter the HSC– University of Mu¨nster, Mu¨nster, Germany; stroma cell interaction and thereby promote leukemia or 3Department of Pediatric Hematology and Oncology, alternatively protect the leukemic blasts from chemotherapy Justus-Liebig-University, Giessen, Germany; by altered niche conditions. Therefore, HLXB9/TEL-induced 4Department of Hematology, Oncology and Tumor transformation might not be reproducible in in vitro transforma- Immunology, HELIOS Clinic, Berlin-Buch, Germany; 5 tion assays. However, a precise mechanism explaining the Laboratoire d’hematologie, Hoˆpital Necker Enfants Malades, transformation mechanism and the unfavorable prognosis of Paris, France and 6Department of Genetics, Friedrich-Alexander University, HLXB9/TEL patients is subject of future studies. Erlangen, Germany This study shows that HLXB9/TEL-positive leukemia is a well- E-mail: [email protected] defined, homogeneous subgroup within the MLL-negative infant 7These authors contributed equally to this work. AMLs. Although this entity shares a short EFS and coexpression of T-cell-associated antigens with MLL-positive patients, expres- sion of either HOXA9 and MEIS1 or HLXB9 clearly distinguishes the two entities and suggests different mechanisms of HSC References transformation. Even though it is necessary to elucidate the molecular role of HLXB9 in the transformation process of 1 Krivtsov AV, Armstrong SA. MLL translocations, histone modifica- tions and leukaemia stem-cell development. Nat Rev Cancer 2007; hematopoietic cells in further experiments, this study provides 7: 823–833. the first step toward functional characterization of this new 2 Hauer J, Tosi S, Schuster FR, Harbott J, Kolb HJ, Borkhardt A. leukemia entity. Graft versus leukemia effect after haploidentical HSCT in a MLL-negative infant AML with HLXB9/ETV6 rearrangement. Pediatr Blood Cancer 2008; 50: 921–923. Conflict of interest 3 von Bergh AR, van Drunen E, van Wering ER, van Zutven LJ, Hainmann I, Lo¨nnerholm G et al. High incidence of The authors declare no conflict of interest. t(7;12)(q36;p13) in infant AML but not in infant ALL, with a dismal outcome and ectopic expression of HLXB9. Genes Cancer 2006; 45: 731–739. 4 Currarino G, Coln D, Votteler T. Triad of anorectal, sacral, and Acknowledgements presacral anomalies. Am J Roentgenol 1981; 137: 395–398. 5 Wada H, Masuda K, Satoh R, Kakugawa K, Ikawa T, Katsura Y et al. We thank Dr Wo¨bmann, the University Childrens Hospital, Adult T-cell progenitors retain myeloid potential. Nature 2008; Giessen; Prof Dr Niemeyer, University Chidren’s Hospital 452: 768–772. Freiburg; Prof Dr Holter, University Hospital Erlangen and Prof 6 Zangrando A, Dell’orto MC, Te Kronnie G, Basso G. MLL Dr Ju¨rgens, University Hospital Mu¨nster for providing clinical rearrangements in pediatric acute lymphoblastic and myeloblastic leukemias: MLL specific and lineage specific signatures. BMC Med patient data and supporting this work. We also thank Dr Sabrina Genomics 2009; 2: 36. Tosi, Brunel University for fruitful discussions and collaboration. 7 Renstrom J, Kroger M, Peschel C, Oostendorp RA. How the This work was supported by research funds of the Heinrich niche regulates hematopoietic stem cells. Chem Biol Interact Heine University, Du¨sseldorf, Germany. 2010; 19: 7–15.

Identification of HLA-An0201-presented T-cell epitopes derived from the tumor-associated antigen M-phase phosphoprotein II in patients with acute myeloid leukemia

Leukemia (2010) 24, 1660–1662; doi:10.1038/leu.2010.147; documented that MPP11 is capable of forming a heterodimeric published online 8 July 2010 complex that associates with ribosomes, acting as a molecular chaperone for nascent polypeptide chains as they exit the In the past decade, several tumor-associated antigens have been ribosome.5,6 By using serological screening of an expression reported to induce immune responses in acute myeloid library, Greiner et al. described MPP11 as a cancer testis leukemia (AML) patients. These antigens include human antigen, which is highly upregulated in malignant myeloblasts in telomerase reverse transcriptase, proteinase 3, Wilms’ tumor 1 patients with AML and chronic myeloid leukemia, but is absent protein, the preferentially expressed antigen in melanoma in normal differentiated tissues except for testis. In a small protein, survivin and the receptor for hyaluronic acid-mediated sample size, the authors were capable of showing significant motility (RHAMM).1–3 The M-phase phosphoprotein 11 (MPP11) titers of anti-MPP11 antibodies in some AML and chronic gene, a member of the MPP family, encodes a 66 kDa protein myeloid leukemia patients.2,3 In the current study, we identified and is located on the 7q22–q31.1.2–4 It has been T-cell epitopes capable of inducing specific cytotoxic T cell line

Leukemia