Letters to the Editor 241 ACKNOWLEDGEMENTS 5 Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE et al. DNMT3A- We thank the patients and families who consented to the use of biologic specimens in Mut in acute myeloid leukemia. N Engl J Med 2010; 363: 2424–2433. these trials and the AML Reference Laboratories of the COG and SWOG for providing 6 Yan XJ, Xu J, Gu ZH, Pan CM, Lu G, Shen Y et al. Exome sequencing identifies diagnostic specimens. This investigation was supported in part by the following PHS somatic mutations of DNA methyltransferase DNMT3A in acute monocytic Cooperative Agreement grant numbers awarded by the National Cancer Institute, leukemia. Nat Genet 2011; 43: 309–315. Cancer Therapy Evaluation Program, DHHS: CA32102, CA38926, CA20319, CA12213. 7 Thol F, Damm F, Lu¨deking A, Winschel C, Wagner K, Morgan M et al. Incidence and prognostic influence of DNMT3A mutations in acute myeloid leukemia. J Clin Oncol 2011; 29: 2889–2896. AUTHOR CONTRIBUTION 8 Shen Y, Zhu YM, Fan X, Shi JY, Wang QR, Yan XJ et al. Gene mutation patterns and FO designed and performed the research, analyzed data and wrote the their prognostic impact in a cohort of 1185 patients with acute myeloid leukemia. manuscript; MO performed statistical analyses and edited the manuscript; PAH, Blood 2011; 118: 5593–5603. DEG, SHP, JEG, CLW, JPR and FRA performed the research and edited the 9 Markova´ J, Michkova´ P, Burcˇkova´ K, Brˇezinova´ J, Michalova´ K, Dohnalova´ A et al. Prognostic impact of DNMT3A mutations in patients with intermediate cytoge- manuscript; SM designed the research, analyzed data and edited the manuscript. netic risk profile acute myeloid leukemia. Eur J Haematol 2012; 88: 128–135. 10 Lin J, Yao DM, Qian J, Chen Q, Qian W, Li Y et al. Recurrent DNMT3A R882 1,2 3 1,4 5 1 F Ostronoff , M Othus ,PAHo , M Kutny , DE Geraghty , mutations in Chinese patients with acute myeloid leukemia and myelodysplastic 2 6 7 1,2 SH Petersdorf , JE Godwin , CL Willman , JP Radich , syndrome. PLoS One 2011; 6: e26906. FR Appelbaum1,2, DL Stirewalt1,2 and S Meshinchi1,4 11 Hou HA, Kuo YY, Liu CY, Chou WC, Lee MC, Chen CY et al. DNMT3A mutations in 1Clinical Research Division, Fred Hutchinson Cancer Research Center, acute myeloid leukemia: stability during disease evolution and clinical implica- Seattle, WA, USA; tions. Blood 2012; 119: 559–568. 0 2Division of Medical Oncology, Seattle Cancer Care Alliance/ 12 Anderson JE, Kopecky KJ, Willman CL, Head D, O Donnell MR, Luthardt FW et al. University of Washington School of Medicine, Seattle, WA, USA; Outcome after induction chemotherapy for older patients with acute myeloid 3 leukemia is not improved with mitoxantrone and etoposide compared to SWOG Group Statistical Center, Fred Hutchinson Cancer Research cytarabine and daunorubicin: a Southwest Oncology Group study. Blood 2002; Center, Seattle, WA, USA; 100: 3869–3876. 4 Division of Pediatric Hematology/Oncology, University of 13 Godwin JE, Kopecky KJ, Head DR, Willman CL, Leith CP, Hynes HE et al. A double- Washington School of Medicine, Seattle, WA, USA; blind placebo-controlled trial of granulocyte colony-stimulating factor in elderly 5Division of Pediatric Hematology/Oncology, University of Alabama patients with previously untreated acute myeloid leukemia: a Southwest onco- at Birmingham, Birmingham, AL, USA; logy group study (9031). Blood 1998; 91: 3607–3615. 6Southern Illinois University Cancer Institute, Springfield, IL, USA and 14 Slovak ML, Kopecky KJ, Cassileth PA, Harrington DH, Theil KS, Mohamad A et al. 7UNM Cancer Research Facility, University of New Mexico Cancer Karyotypic analysis predicts outcome of preremission and postremission therapy Research and Treatment Center, Albuquerque, NM, USA in adult acute myeloid leukemia: a Southwest Oncology Group/Eastern Co- operative Oncology Group study. Blood 2000; 96: 4075–4083. E-mail: [email protected] 15 Ho PA, Kutny MA, Alonzo TA, Gerbing RB, Joaquin J, Raimondi SC, Gamis AS, Meshinchi S. Leukemic mutations in the methylation-associated DNMT3A REFERENCES and IDH2 are rare events in pediatric AML: a report from the Children0s Oncology 1 Thiede C, Steudel C, Mohr B, Schaich M, Scha¨kel U, Platzbecker U et al. Analysis of Group. Pediatr Blood Cancer 2011; 57: 204–209. FLT3-activating mutations in 979 patients with acute myelogenous leukemia: 16 Ro¨llig C, Bornha¨user M, Thiede C, Taube F, Kramer M et al. Long-term prognosis of association with FAB subtypes and identification of subgroups with poor prog- acute myeloid leukemia according to the new genetic risk classification of the nosis. Blood 2002; 99: 4326–4335. European LeukemiaNet recommendations: evaluation of the proposed reporting 2 Falini B, Mecucci C, Tiacci E, Alcalay M, Rosati R, Pasqualucci L et al. GIMEMA Acute system. J Clin Oncol 2011; 29: 2758–2765. Leukemia Working Party. Cytoplasmic nucleophosmin in acute myelogenous 17 Marcucci G, Metzeler KH, Schwind S, Becker H, Maharry K et al. Age-related leukemia with a normal karyotype. N Engl J Med 2005; 352: 254–266. prognostic impact of different types of DNMT3A mutations in adults with 3 Ho PA, Alonzo TA, Gerbing RB, Pollard J, Stirewalt DL, Hurwitz C et al. Prevalence and primary cytogenetically normal acute myeloid leukemia. J Clin Oncol 2012; 30: prognostic implications of CEBPA mutations in pediatric acute myeloid leukemia 742–750. (AML): a report from the Children0s Oncology Group. Blood 2009; 113: 6558–6566. 18 Metzeler KH, Walker A, Geyer S, Garzon R, Klisovic RB et al. DNMT3A mutations 4 Yamashita Y, Yuan J, Suetake I, Suzuki H, Ishikawa Y, Choi YL et al. Array-based and response to the hypomethylating agent decitabine in acute myeloid genomic resequencing of human leukemia. Oncogene 2010; 29: 3723–3731. leukemia. Leukemia 2012; 26: 1106–1107.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Hairy cell leukemia cell lines expressing A1 and displaying B-cell receptor signals characteristic of primary tumor cells lack the signature BRAF mutation to reveal unrepresentative origins

Leukemia (2013) 27, 241–245; doi:10.1038/leu.2012.163 representation of parental disease, and here genetic imprints from recent advances in genomic sequencing of primary tumor cells may prove highly relevant. In the rare B-cell malignancy Tumor-derived cell lines provide valuable models to assess disease (HCL), a number of cell lines have been characteristics but require careful assessment with regard to their generated, but evidence linking cell lines to primary tumor

Accepted article preview online 18 June 2012; advance online publication, 13 July 2012

& 2013 Macmillan Publishers Limited Leukemia (2013) 233 – 258 Letters to the Editor 242 remains incomplete. These include Hair-M,1 HCLL-7876,2 EH,3 the HCLv-07 cell line (Figures 1a and d). CD103 expression in 3/4 of Eskol,4 HC-1(ref. 5) and HCLv-07,6 examined to date for ‘tumor’- the ANXA1 þ cell lines was at mid to strong levels ( þþ to þþ associated CD20 phosphorylation,7 interferon-a oncogene þþ), although in HCLL-7876 its expression was relatively low regulation,8 CD11c gene expression9 and HCL and (Figures 1a and d). Expression of CD103, CD25 and CD123 in HCLv-07 migration.10 Here we have assessed these HCL cell lines further by was low ( þ ), as expected (Figures 1a and d).16 EH cells, however, examining the phenotype in more detail, by comparing B-cell despite being ANXA1 À ve and showing low CD25 ( þ ), expressed receptor (BCR) function and by evaluating the BRAF mutation. The CD103 and CD123 at levels ( þþ) comparable to some of the four BRAF V600E mutation is present near universally in typical ANXA1 þ cell lines (Figures 1a and d), further raising the possibility HCL.11–13 However, a few cases lack this signature mutation, that it may differ from HCLv-07. associated with tumor use of the IGHV4-34 gene,14 and in a solitary TRAP expression has been previously reported in each of the four HCL case expressing CD5.15 In HCL variant (HCLv) disease, the ANXA1 þ HCL cell lines.1,2,4,5 The EH cell line, which is ANXA1 À ve,also absence of the BRAF V600E mutation is a consistent feature.12–14 expresses TRAP,3 whereas HCLv-07 is negative for TRAP activity.6 This We evaluated six HCL cell lines, each expressing a single pre- or again tends to suggest that EH may be closer to CD5 þ ‘atypical’ HCL. post-switched isotype for diagnostic markers, comparing them with The use of the IGHV gene was also evaluated in each of the primary tumors expressing a single isotype (s-HCL) (Figure 1). Of cell lines, revealing somatic mutations in each tumor-derived these cell lines, HCLv-07 is known to be derived from HCLv.6 Under gene (Figure 1a), resembling the main immunogenetic finding in the WHO (World Health Organization) (2008) guidelines, HCL HCL. None expressed the IGHV4-34 gene (Figure 1a).14 The dual immunophenotype includes bright CD11c expression with CD103, aspect of mutated IGHV genes yet absence of CD27 expression, CD25, CD123 and Annexin A1 (ANXA1) coexpression.16 ANXA1 is the classical marker for normal germinal center-derived memory B regarded as the most specific tumor marker for HCL, as it is absent on cells, has raised the question of an unusual cell of origin in HCL, other B-cell lymphomas16–18 and is negative in HCLv to provide a for which our recent report of normal IgM þ B cells that lack CD27 highly relevant marker to segregate these two forms of disease.16,17 and carry a low level of mutations in IGHV genes may be Immunostaining for tartrate-resistant acid phosphatase (TRAP) is also relevant.26 a diagnostic tool for HCL.16 CD5expressioninHCL,although We also examined BCR function in tumor cells selected infrequent (5/133; 4% of cases19), has led to CD5 þ HCL being by CD19, CD11cHi and CD103 expression, using anti-surface 15 described as ‘atypical’ disease. CD5 expression in HCLv is also rare immunoglobulin (sIg) isotype-specific stimuli with goat F(ab’)2- (1/72 cases20–22), and CD25 and CD123 expression is usually anti-human Ig antibodies (Southern Biotech, Birmingham, AL, negative,16 butthisisnotanabsolutefinding.20 USA). Release of intracellular Ca2 þ stores was measured using the Notably, four of the cell lines (HCLL-7876 ¼ Eskol4HC-14Hair-M) calcium-responsive fluorophore, fluo3-AM (Invitrogen, Paisley, UK). were positive for ANXA1 expression by western blot analysis for Briefly, cells were loaded with 7 mM Fluo3-AM at 37 1C for 30 min, the 38-kDa-size (Figures 1a–c), using the antibody ‘Purified and fluorescence was then measured by flow cytometry before Mouse anti-Annexin I’, clone 29/Annexin I17 (BD Transduction and after anti-sIg stimulation. Responses were calculated as % Laboratories, Oxford, UK). Positive controls for ANXA1 expression cells over threshold. Each cell line consistently induced Ca2 þ flux included four primary s-HCL tumor samples (s-HCL1/14/15/17) following anti-IgH or anti-IgL stimulation (Figure 1a). This mirrors and two healthy PBMNC (peripheral blood mononuclear cells) Ca2 þ flux signals in s-HCL tumor cells, although the intensity of preparations (non-B-cell positive). ANXA1 expression was also response was generally lower in the cell lines than in primary gauged by immunostaining of acetone-fixed cytospin prepara- tumors (Figure 1a). In 4/4 cell lines examined (EH, Eskol, Hair-M, tions, using the same antibody with anti-mouse immuno- HC-1), BCR signals induced phosphorylation of ERK1/2 (extra- globulin (Ig)-peroxidase and diaminobenzidine substrate for cellular signal-regulated 1/2) (Figure 1e), with cells detection, and hematoxylin staining for nuclei. The level of stimulated with anti-sIg antibody for 2–60 min, fixed, permeabi- ANXA1 expression in the HCLL-7876 HCL cell line was lower than lized and stained for phosphorylated ERK (Phosflow method; that seen in primary s-HCL tumor cells by both western blot BD Biosciences, Oxford, UK). In 2/2 cell lines (HCLL-7876, HCLv-07), analysis (Figure 1b) and immunostaining (Figure 1c). However, BCR-mediated stimulation for 48 h triggered a marked ANXA1 expression was a consistent feature of the bulk of cells in (Figure 1f), as measured using the irreversible pan-caspase inhibi- the selected HCLL-7876 cell line, with ANXA1 staining both tor ApoStat (R&D Systems, Abingdon, UK) or Apo2.7 (Beckman nuclear and cytoplasmic protein as found in primary s-HCL cells Coulter, High Wycombe, UK). In our recent study in a cohort (Figure 1c) and confirming western blot data. Separate studies also of 10 HCL cases, we showed that specific anti-BCR signals in report convergence of results for detection of ANXA1 with primary tumor cells initiate Ca2 þ flux and induce ERK1/2 western blots and immunostaining.23,24 It is noteworthy that in phosphorylation, leading to levels of apoptosis significantly extensive immunohistochemical staining for ANXA1 and other beyond that observed spontaneously ex vivo.27 Taken together, diagnostic markers in HCL cases showing weak or focal staining comparable BCR function in cell lines and primary HCL tumors were scored as positive.18 These criteria affirm ANXA1 expression were indicative of common behavior, but did not distinguish in the HCLL-7876 cell line, and indicate that its expression in the between the variant HCLv-07 and other cell lines. other three cell lines (Eskol, HC-1 and Hair-M) can also be To probe these cell lines further, we evaluated the BRAF considered positive. Two cell lines (EH and HCLv-07) were clearly mutation (V600E) that has recently been established as a signature negative for ANXA1 (Figures 1a and b) by western blot analysis, molecular lesion in typical HCL.11–13 We analyzed genomic DNA an expected phenotype for HCLv-07. ANXA1 immunostaining and mRNA from HCL cell lines and s-HCL cases by PCR in the EH cell line confirmed its absence (Figure 1c). As EH is amplification and DNA sequencing, with each sample analyzed CD5 þ (Figures 1a and d), it is unclear from these data whether in duplicate. Design of PCR primers amplified either a 256-bp DNA it associates with ‘atypical’ HCL15 or the rare CD5 þ HCLv fragment incorporating exon 15, where the BRAF (V600E) mutation cases.20–22 resides (BRAFDNA FWD 50-TACCTAAACTCTTCATAATGCTTGC-30, Other HCL phenotypic markers were assessed by flow cytometry, BRAFDNA REV 50-GTAACTCAGCAGCATCTCAGGG-30),11 or a 532- and relative levels of expression scored using well-defined mean bp fragment spanning exons 13–16 for mRNA analysis by RT-PCR fluorescence intensity ranges from replicate assays (Supplementary (BRAFmRNA FWD 50-GTGACAGCACCTACACCTCAGCAGTTAC-30, Table 1). All HCL cell lines expressed CD11c and were negative for BRAFmRNA REV 50-GGTCCCTGTTGTTGATGTTTGAATAAGG-30). CD27 (Figure 1a), a known feature in HCL.25 Of the four HCL cell lines Strikingly, the HCL-associated BRAF mutation was absent in shown to express ANXA1, each also expressed CD25 and CD123 at each of the four ANXA1 þ cell lines (Eskol, Hair-M, HC-1 and HCLL- mid to high levels ( þþ to þþþ), at least 2-fold higher than in 7876; Figures 2a and b), which differs from typical HCL. In contrast,

Leukemia (2013) 233 – 258 & 2013 Macmillan Publishers Limited Letters to the Editor 243

Figure 1. BCR characteristics and function in HCL cell lines and s-HCL. (a) Summary of phenotype data by CD19/11c/103/27/25/123/5 ( þ to þþþþ denotes relative expression level according to mean fluorescence intensity range determined by flow cytometry; Supplementary Table 1), ANXA1 expression ( þ to þþþþ by densitometry of western blots), sIg ( þ to þþþ denotes % cells positive) and Ca2 þ flux induced by anti-sIg stimuli ( þ to þþþþ denotes % cells responding over threshold), together with IGHV gene use (ND is not done). (b) Western blot for ANXA1 expression in HCL cell lines, primary s-HCL cases and controls (healthy PBMNC as positive (non-B cells), and lymphoma cell line, Ramos, as negative) (0.6 Â 106 cells loaded/lane). (c) Immunostaining for ANXA1 in HCLL-7876, EH and s-HCL1 using cytospin preparations, detected using ImmPRESS anti-mouse Ig-peroxidase (Vector Laboratories, Peterborough, UK) and DAB substrate (brown) with hematoxylin for nuclei (purple). (d) Flow cytometry profiles showing representative expression levels of CD25, CD123 and CD5 in selected cell lines and primary s-HCL and normal PBMNC. (e) A representative example of observed ERK1/2 phosphorylation in cell line HC-1 following isotype-specific stimulation with goat F(ab’)2-anti-human Ig antibodies (each was carried outX2 times). (f) The figure panel shows a representative example of BCR-induced apoptosis in the HCLL-7876 cell line following stimulation with goat F(ab’)2-anti-human Ig antibodies for 48 h and staining for Apo2.7-PE by flow cytometry (each assay was carried out X3 times). Statistical comparison was made using Student’s t-test (GraphPad Prism Software, La Jolla, CA, USA). the characteristic BRAF T1860A (V600E) mutation was readily Its absence in the CD5 þ EH cell line tends to further suggest an observed in 4/4 primary s-HCLs (Figures 2a and b). The lack of the association with CD5 þ ‘atypical’ HCL, although thus far only a BRAF mutation in HCLv-07 is to be expected of HCLv origins.14 single such case has been evaluated for this mutation.15

& 2013 Macmillan Publishers Limited Leukemia (2013) 233 – 258 Letters to the Editor 244 3Department of Hematology, University Medical Center, Groningen, The Netherlands E-mail: [email protected]

REFERENCES 1 Matsuo Y, Sagawa K, Morita M, Minowada J, Yokoyama MM. Establishment of a hairy cell leukemia cell line carrying Tac antigen and phagocytic activity with B-cell characteristics. J Natl Cancer Inst 1986; 76: 207–216. 2 Kanowith-Klein S, Saxon A, Uittenbogaart CH. Constitutive production of B cell differentiation factor-like activity by human T and B cell lines. Eur J Immunol 1987; 17: 593–598. 3 Faguet GB, Satya-Prakash KL, Agee JF. Cytochemical, cytogenetic, immuno- phenotypic and tumorigenic characterization of two hairy cell lines. Blood 1988; 71: 422–429. 4 Harvey W, Srour EF, Turner R, Carey R, Maze R, Starrett B et al. Characterization of a new cell line (ESKOL) resembling hairy-cell leukemia: a model for oncogene regulation and late B-cell differentiation. Leuk Res 1991; 15: 733–744. 5 Schiller JH, Bittner G, Meisner LF, Oberley TD, Norback D, Schwabe M et al. Establishment and characterization of an Epstein-Barr virus spontaneously transformed lymphocytic cell line derived from a hairy cell leukemia patient. Leukemia 1991; 5: 399–407. 6 Sasaki M, Aritaka N, Tsukune Y, Kawahara S, Masuda A, Tsutsui M et al. Establishment of a hairy cell leukemia variant cell line, HCLv-07. Acta Haematol Figure 2. Analyses of the BRAF (V600E) mutation in HCL cell lines and 2009; 121: 63–66. primary s-HCL. Both cell line and s-HCL sample mRNA or gDNA were 7 Genot EM, Meier KE, Licciardi KA, Ahn NG, Uittenbogaart CH, Wietzerbin J et al. analyzed for the presence of a T-A transversion at position 1860 of Phosphorylation of CD20 in cells from a hairy cell leukemia cell line. Evidence for BRAF mRNA (NM_004333) or position 171429 of BRAF gDNA involvement of calcium/-dependent protein II. J Immunol 1993; (NG_007873) by PCR amplification followed by Sanger DNA sequencing 151: 71–82. in duplicated assays. (a) The figure panel shows representative 8 Harvey WH, Harb OS, Kosak ST, Sheaffer JC, Lowe LR, Heerema NA. Interferon- examples of electropherograms from DNA sequencing of the cell line alpha-2b downregulation of oncogenes H-ras, c-raf-2, c-kit, c-myc, c-myb and Eskol revealing germline T and from a s-HCL patient sample (s-HCL15) c-fos in ESKOL, a hairy cell leukemic line, results in temporal perturbation of signal revealing a heterozygous T/A mutation. (b)Thefigurepanel transduction cascade. Leuk Res 1994; 18: 577–585. summarizes data on the absence of the BRAF (V600E) mutation in 6/6 9 Nicolaou F, Teodoridis JM, Park H, Georgakis A, Farokhzad OC, Bottinger EP et al. HCL lines and its presence in 4/4 s-HCL samples. CD11c in hairy cell leukemia is dependent upon activation of the proto-oncogenes ras and junD. Blood 2003; 101: 4033–4041. 10 Galiegue-Zouitina S, Delestre L, Dupont C, Troussard X, Shelley CS. These data reveal clear divergent origins of HCL cell lines. Underexpression of RhoH in Hairy Cell Leukemia. Cancer Res 2008; 68: 4531–4540. Although ANXA1 expression and other phenotypic markers 11 Tiacci E, Trifonov V, Schiavoni G, Holmes A, Kern W, Martelli MP et al. BRAF together with cellular behavior show some similarities with mutations in hairy-cell leukemia. N Engl J Med 2011; 364: 2305–2315. features of typical HCL in four of the cell lines examined (HCLL- 12 Tiacci E, Schiavoni G, Forconi F, Santi A, Trentin L, Ambrosetti A et al. Simple 7876, Eskol, HC-1, Hair-M), the absence of the BRAF mutation genetic diagnosis of hairy cell leukemia by sensitive detection of the BRAF-V600E clearly places into question their reliability as representative of this mutation. Blood 2012; 119: 192–195. form of the malignancy. The BRAF mutation-negative EH cell line 13 Arcaini L, Zibellini S, Boveri E, Riboni R, Rattotti S, Varettoni M et al. The BRAF on the other hand may mirror CD5 þ ‘atypical’ HCL disease,15 V600E mutation in hairy cell leukemia and other mature B-cell neoplasms. Blood although its origin from the rare CD5 þ HCLv disease cannot be 2012; 119: 188–191. ruled out, and finally the BRAF mutation negative HCLv-07 14 Xi L, Arons E, Navarro W, Calvo KR, Stetler-Stevenson M, Raffeld M et al. Both 12–14 variant and IGHV4-34-expressing hairy cell leukemia lack the BRAF V600E recapitulates the expected features of HCLv tumors. mutation. Blood 2012; 119: 3330–3332. Importantly, these observations exemplify the relevance of 15 Lennerz JK, Klaus BM, Marienfeld RB, Moller P. Pyrosequencing of BRAF V600E in molecular findings from next-generation sequencing of tumor routine samples of Hairy Cell Leukaemia identifies CD5 þ variant Hairy Cell genomes to integrate models of disease. Leukaemia that lacks V600E. Br J Haematol 2012; 157: 267–269. 16 Foucar K, Falini B, Catovsky D, Stein H. Hairy cell leukaemia. In: Swerdlow SH, Campo E, Harris NL et al. (eds). WHO Classification of Tumours of Haematopoietic CONFLICT OF INTEREST and Lymphoid Tissues. IARC Press: Lyon, 2008, pp 188–190. 17 Falini B, Tiacci E, Liso A, Basso K, Sabattini E, Pacini R et al. Simple diagnostic assay The authors declare no conflict of interest. for hairy cell leukaemia by immunocytochemical detection of annexin A1 (ANXA1). Lancet 2004; 363: 1869–1870. 18 Sherman MJ, Hanson CA, Hoyer JD. An assessment of the usefulness of ACKNOWLEDGEMENTS immunohistochemical stains in the diagnosis of hairy cell leukemia. Am J Clin This work was funded by Leukaemia and Lymphoma Research (UK). Cell lines were Pathol 2011; 136: 390–399. obtained from DMSZ (HC-1) or were kind gifts from The Danish Cancer Society/Bill 19 Robbins BA, Ellison DJ, Spinosa JC, Carey CA, Lukes RJ, Poppema S et al. Diag- Harvey (Hair-M and Eskol), Professor J Cawley at Liverpool (HCLL-7876), Dr Carl nostic application of two-color flow cytometry in 161 cases of hairy cell leukemia. Shelley at Harvard (EH) and Dr M Sasaki at Juntendo University, Tokyo (HCLv-07). Blood 1993; 82: 1277–1287. 20 Del Giudice I, Matutes E, Morilla R, Morilla A, Owusu-Ankomah K, Rafiq F et al. The NJ Weston-Bell1, D Hendriks1, G Sugiyarto1, NA Bos2, diagnostic value of CD123 in B-cell disorders with hairy or villous lymphocytes. HC Kluin-Nelemans3, F Forconi1 and SS Sahota1 Haematologica 2004; 89: 303–308. 1Cancer Sciences Unit, Faculty of Medicine, University of 21 Matutes E, Wotherspoon A, Catovsky D. The variant form of hairy-cell leukaemia. Best Pract Res Clin Haematol 2003; 16: 41–56. Southampton, Southampton, UK; 2 22 Sainati L, Matutes E, Mulligan S, de Oliveira MP, Rani S, Lampert IA et al. Department of Rheumatology and Clinical Immunology, A variant form of hairy cell leukemia resistant to alpha-interferon: Section Immunology, University Medical Center, clinical and phenotypic characteristics of 17 patients. Blood 1990; 76: Groningen, The Netherlands and 157–162.

Leukemia (2013) 233 – 258 & 2013 Macmillan Publishers Limited Letters to the Editor 245 23 Suo A, Zhang M, Yao Y, Zhang L, Huang C, Nan K et al. analysis of the 26 Weston-Bell N, Townsend M, Di Genova G, Forconi F, Sahota SS. Defining effects of sorafenib on human hepatocellular carcinoma cell line HepG2. origins of malignant B cells: a new circulating normal human IgM( þ )D( þ ) Med Oncol 2011, e-pub ahead of print 7 July 2011; doi:10.1007/s12032-011-0013-y. B-cell subset lacking CD27 expression and displaying somatically mutated 24 McArthur S, Cristante E, Paterno M, Christian H, Roncaroli F, Gillies GE et al. IGHV genes as a relevant memory population. Leukemia 2009; 23: Annexin A1: a central player in the anti-inflammatory and neuroprotective role of 2075–2080. microglia. J Immunol 2010; 185: 6317–6328. 27 Weston-Bell NJ, Babbage G, Forconi F, Kluin-Nelemans HC, Sahota SS. Hairy cell 25 Forconi F, Raspadori D, Lenoci M, Lauria F. Absence of surface CD27 distinguishes leukaemia displaying multiple surface immunoglobulin isotypes reveal a func- hairy cell leukemia from other leukemic B-cell malignancies. Haematologica 2005; tional B-cell receptor in which isotype roles differ. Blood (ASH Annual Meeting 90: 266–268. Abstracts) 2011; 118: 1567.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

A rare but specific subset of adult AML patients can be defined by the cytogenetically cryptic NUP98–NSD1 fusion gene

Leukemia (2013) 27, 245–248; doi:10.1038/leu.2012.230 of de novo NPM1 unmutated (NPM1wt) CN-AML. Furthermore, we screened 64 adult AML cases with cytogenetic aberrations involving 5 and 11 for the presence of NUP98– Acute myeloid leukemia (AML) is a heterogeneous group of NSD1 fusion gene. In addition, we aimed at evaluating this fusion diseases in terms of clinical presentation, genetic alterations transcript as a target for quantitative RT–PCR (qRT–PCR)-based and response to treatment. Cytogenetic abnormalities that are minimal residual disease (MRD) monitoring. The study design identified in B55% of all adult patients with AML enable adhered to the tenets of the Declaration of Helsinki and was subclassification and risk stratification. Fusion genes can be approved by our institutional review board before its initiation. detected in B20–25% of all AML cases, and usually are the result NUP98–NSD1 fusion transcript analysis was performed on either of a cytogenetically detectable chromosomal rearrangement. Most bone marrow or peripheral blood samples. Presence of recurrent chromosomal alterations are associated with specific morphological and clinical characteristics,1 such as the reciprocal t(8;21)/RUNX1-RUNX1T1, inv(16)/CBFB-MYH11 and t(15;17)/ Table 1. Demographics and clinical and molecular characteristics of PML-RARA rearrangements, all of which are associated with good AML patients according to NUP98–NSD1 fusion transcript prognosis. In contrast, complex aberrations with X3 chromosomal 2 abnormalities are associated with inferior outcome. However, Characteristics NUP98–NSD1 NUP98–NSD1 P even in high-quality cytogenetic preparations, G-banding analysis negative (n ¼ 370) positive (n ¼ 8) suffers from an inherent limit in resolution, such that rearrangements o5–10 megabases in size, particularly those No. of %of No. of %of involving uniformly pale G-banded regions, may be impossible to patients total patients total be detected. Molecular genetic approaches such as reverse Age (years) transcriptase–PCR (RT–PCR) have demonstrated that a small Median 62.4 42.3 o0.001 percentage of apparently normal karyotypes may harbor cryptic Range 15.7–89.6 20.9–71.4 versions of known recurrent translocations, which are generated Gender by submicroscopic insertions or more complex rearrangements.3 Female 140 37.0 4 1.06 0.361 The cryptic NUP98–NSD1 fusion involves the nucleoporin gene 98 Male 230 60.8 4 1.06 (NUP98) on 11p15 and the non-homeobox Hemoglobin, g/dl 0.323 gene NSD1 in chromosomal band 5q35. NUP98 encodes a Median 9.5 8.7 98-kDa protein of the nuclear pore complex, and is known to Range 4.0–16.3 5.0–13.4 fuse to at least 21 different fusion gene partners in chromosomal WBC count, Â 109/l 0.224 rearrangements of various hematopoietic disorders.4 NSD1 Median 28.9 80.4 contains two distinct nuclear receptor interaction domains, as Range 0.9–274.0 4.5–300.0 well as a SET domain and multiple PHD fingers, both of which are Platelet count, Â 109/l 0.584 frequently found in transcriptional regulators and may be involved Median 93.7 144.4 in chromatin remodeling.5,6 NSD1 is thought to function as both a Range 3.0–950.0 28.0–337.0 transcriptional coactivator and a corepressor. The NUP98–NSD1 Bone marrow blasts, % 0.039 fusion gene has been shown to induce AML in vivo, which sustains Median 53.5 71.1 self-renewal of myeloid stem cells in vitro, and enforces expression Range 8.0–99.5 24.0–98.5 7 of the HOXA7, HOXA9, HOXA10 and MEIS1 proto-oncogenes. Correlating mutations Recently, Hollink et al.8 published a study on NUP98–NSD1 in 293 FLT3-ITD (n ¼ 73/378) 67 17.7 6 1.6 0.001 pediatric and 808 adult cytogenetically normal AML (CN-AML) FLT3-TKD (n ¼ 16/344) 16 4.2 0 0.0 0.680 CEBPA (n ¼ 45/377) 45 11.9 0 0.0 0.358 cases. The NUP98–NSD1 fusion gene has been described in this MLL-PTD (n ¼ 56/375) 56 14.8 0 0.0 0.271 single study with a frequency of 16.1% in pediatric and 2.3% in RUNX1 (n ¼ 90/348) 90 23.8 0 0.0 0.089 adult AML patients with distinct characteristics (for example, ASXL1 (n ¼ 104/371) 104 27.5 0 0.0 0.070 WT1 (n ¼ 27/366) 23 6.1 4 1.1 0.001 mutual exclusiveness with NPM1) and dismal prognosis. The aim of our study was to further investigate the frequency and clinical Abbreviations: NUP98, nucleoporin gene 98; WBC, counts. relevance of the NUP98–NSD1 fusion transcript in 378 adult cases Significant values are written in bold.

Accepted article preview online 14 August 2012; advance online publication, 4 September 2012

& 2013 Macmillan Publishers Limited Leukemia (2013) 233 – 258