Letters to the Editor 1778 patients who presented with mutant DNA methyltransferase 3 ACKNOWLEDGEMENTS alpha (DNMT3A) at diagnosis. This led them to assume that This work was supported by Leuka¨miehilfe Steiermark and the START-Funding- mutations in DNMT3A may cause genetic instability and thereby Program of the Medical University of Graz, Graz, Austria (both to AZ). induce FLT3-ITD, ultimately leading to relapse of AML. We recently described mutations in DNMT3A in 13 of 37 (35%) patients diagnosed with secondary AML (sAML)2 arising from either AUTHOR CONTRIBUTIONS myelodysplastic syndromes (MDS) or chronic myeloproliferative AZ was the principal investigator of this study, analyzed the data and wrote the neoplasias (MPN). In agreement with Wakita et al.,theDNMT3A paper. GH and AW performed sequence analyses and analyzed the data. FQ mutations detected at sAML diagnosis were already present at the was responsible for statistical analyses. HS contributed AML samples and was time of the precedent MDS/MPN in 10 of 10 (100%) cases, involved in data analysis, and in writing of the paper. All authors read and suggesting that mutant DNMT3A is a founder event in the contributed to the final version of the paper. development of sAML. FLT3-ITD is a known progression marker from MDS to AML with low frequencies in the MDS and high A Zebisch1, G Hoefler2, F Quehenberger3,AWo¨lfler1 and H Sill1 3,4 frequencies in the secondary leukemia samples, respectively. We, 1Division of Hematology, Medical University of Graz, Graz, Austria; therefore, asked whether the existence of DNMT3A mutations at 2Institute of Pathology, Medical University of Graz, Graz, Austria and MDS/MPN diagnosis might be accountable for the accumulation of 3Institute of Medical Informatics, Statistics and Documentation, FLT3-ITD during the clonal evolution to sAML and reanalyzed all 37 Medical University of Graz, Graz, Austria 2 sAML specimens described in Fried et al. Screening for mutations in E-mail: [email protected] nucleophosmin 1 (NPM1)andtheFLT3 tyrosine kinase domain (FLT3- TKD), as well as for the presence of FLT3-ITD was performed as previously described.5,6 Informed consent was obtained from all REFERENCES individuals and the study was approved by the institutional review 1 Wakita S, Yamaguchi H, Omori I, Terada K, Ueda T, Manabe E et al. Mutations of board of the Medical University of Graz, Graz, Austria. Statistical the epigenetics-modifying gene (DNMT3a, TET2, IDH1/2) at diagnosis may induce correlation of DNMT3A with FLT3-ITDwascalculatedbyFisher’sexact FLT3-ITD at relapse in de novo acute myeloid leukemia. Leukemia 2012, (e-pub test using R 2.15.2 (http://www.r-project.org). ahead of print 8 November 2012). doi:10.1038/leu.2012.317. We identified FLT3-ITD in 8 of 37 (22%) and point mutations 2 Fried I, Bodner C, Pichler MM, Lind K, Beham-Schmid C, Quehenberger F et al. within the FLT3-TKD in 3 of 37 (8%) patients with sAML. Only 3 of Frequency, onset and clinical impact of somatic DNMT3A mutations in therapy- 37 (8%) patients exhibited mutations in NPM1, which is less than related and secondary acute myeloid leukemia. Haematologica 2012; 97: 246–250. 7 3 Dicker F, Haferlach C, Sundermann J, Wendland N, Weiss T, Kern W et al. Mutation the frequencies reported in de novo AML. Importantly, FLT3-ITD analysis for RUNX1, MLL-PTD, FLT3-ITD, NPM1 and NRAS in 269 patients with MDS and DNMT3A mutations occurred independently of each other or secondary AML. Leukemia 2010; 24: 1528–1532. (P ¼ 0.69; Table 1), indicating that the presence of mutant DNMT3A 4 Zebisch A, Czernilofsky AP, Keri G, Smigelskaite J, Sill H, Troppmair J. Signaling at diagnosis of MDS/MPN does not result in a higher frequency of through RAS-RAF-MEK-ERK: from basics to bedside. Curr Med Chem 2007; 14: FLT3-ITD after transformation into sAML. These data raise a couple 601–623. of issues. First, the fact that different results were reported by 5 Zebisch A, Staber PB, Delavar A, Bodner C, Hiden K, Fischereder K et al. Two Wakita’s and our group pinpoint different mechanisms of de novo transforming C-RAF germ-line mutations identified in patients with therapy- and secondary leukemogenesis. This assumption is further related acute myeloid leukemia. Cancer Res 2006; 66: 3401–3408. corroborated by the low frequency of NPM1 mutations observed 6 Fried I, Wolfler A, Quehenberger F, Hoefler G, Sill H, Zebisch A. Mutations 8,9 inDNMT3A and loss of RKIP are independent events in acute monocytic leukemia. in our and previous cohorts of sAML patients. Second, as both Haematologica 2012; 97: 1936–1937. the groups analyzed a limited number of samples, further 7 Kayser S, Dohner K, Krauter J, Kohne CH, Horst HA, Held G et al. The impact of investigations are necessary to proof that mutations in DNMT3A therapy-related acute myeloid leukemia (AML) on outcome in 2853 adult patients in fact cause genetic instability and, in this way, trigger the with newly diagnosed AML. Blood 2011; 117: 2137–2145. acquisition of secondary genetic aberrations. Therefore, studies 8 Milosevic JD, Puda A, Malcovati L, Berg T, Hofbauer M, Stukalov A et al. Clinical implying next generation sequencing technologies are most significance of genetic aberrations in secondary acute myeloid leukemia. Am J appropriate as they allow an unbiased, genome-wide approach Hematol 2012; 87: 1010–1016. for the assessment of both, mutation frequencies, as well as 9 Schnittger S, Bacher U, Haferlach C, Alpermann T, Dicker F, Sundermann J et al. mutational patterns in a wide variety of genes. Characterization of NPM1-mutated AML with a history of myelodysplastic syn- dromes or myeloproliferative neoplasms. Leukemia 2011; 25: 615–621. 10 Pichler MM, Bodner C, Fischer C, Deutsch AJ, Hiden K, Beham-Schmid C et al. Eva- luation of mutations in the isocitrate dehydrogenase genes in therapy-related and CONFLICT OF INTEREST secondary acute myeloid leukaemia identifies a patient with clonal evolution to IDH2 The authors declare no conflict of interest. R172K homozygosity due to uniparental disomy. Br J Haematol 152:669–672.

Natural history of acute lymphoblastic leukemia in neurofibromatosis type 1 monozygotic twins

Leukemia (2013) 27, 1778–1781; doi:10.1038/leu.2013.55 have revealed not a linear sequence of mutation acquisition but rather a clonal development model based on complex branching similar to Darwin’s evolution trees and natural selection. Indeed, a frequently occurring prenatal first hit is followed by the acquisition of Although the natural history of acute lymphoblastic leukemia (ALL) critical and secondary copy number alterations (CNAs) in different is clinically silent, identical twin pairs suffering from concordant subclones, which gain a selective advantage giving rise to leukemia.2 leukemia provide a unique and tractable model to better understand Here, we report a pair of female monozygotic monochorionic the clonal evolution of this disease.1 Recently, genetic studies on ALL twins (referred to as T1 and T2, respectively) affected with

Accepted article preview online 21 February 2013; advance online publication, 12 March 2013

Leukemia (2013) 1745 – 1791 & 2013 Macmillan Publishers Limited Letters to the Editor 1779 Neurofibromatosis type 1 (NF1). Both patients were diagnosed benign neurofibromas3 and predispositions to malignancies,4 with concordant ALL, aged 6 and 6.5 years. Genome-wide array mainly myeloid leukemia.5,6 Monoallelic loss of NF1 followed by analysis of these identical twins led us to hypothesize that they its bi-allelic inactivation through a second hit has a role in neo- might represent a bona fide model of leukemia clonal evolution plastic progression of somatic cells.7,8 Nevertheless, the sequence of associated to NF1. events occurring during clonal evolution from NF1 mutations to the NF1 is an autosomal dominant disorder caused by mutations of development of leukemia still remains poorly understood. the NF1 gene, located at chromosome band 17q11.2, encoding for T1 and T2 were diagnosed with NF1 in the first year of life. They neurofibromin, a GTPase-activating protein that negatively reg- showed NF1 clinical features and sequencing data confirmed the ulates the activation of Ras family members. Clinical features of NF1 presence of an identical inactivating germline NF1 mutation include alterations of skin pigmentation (cafe`-au-lait macules), (2763insA) in both the children (data not shown).

Figure 1. (A) Copy number abnormalities detected in the diagnostic samples of T1 and T2 twins. (a): Representative copy number (CN) plots for T1 and T2 diagnostic and T2 relapse samples are shown. Homozygous profile of allele peaks (AP) reveals a copy number neutral LOH of the whole 17q arm. (b–e): Weighted log2 ratio values (WLR) for each probe showed: 10q deletions different between T1 and T2 (b) and hemizygous focal deletions involving BACH2, BX247990 and TCF12 genes (c–e). (B) Copy number abnormalities detected at relapse of T2 twin. Array plots show representative additional abnormalities detected in the relapse clone of T2: Ikaros (a), ETV6 (b) and C20orf94 (c) focal deletions. WLR ¼ weighted log2 ratio.

& 2013 Macmillan Publishers Limited Leukemia (2013) 1745 – 1791 Letters to the Editor 1780 The clinical, immunophenotypic and cytogenetic data of T1 and chromosome 17, where the NF1 gene is located (Figure 1A). Both T2 at diagnosis and relapse are summarized in Supplementary diagnostic samples of the twins (but not the T2 relapse sample) Table 1. At diagnosis, T1 bone marrow (BM) aspirate analysis showed chromosome 7 monosomy; we did not ascertain the confirmed diagnosis with common B-cell precursor ALL with parental origin of this chromosome. Nine additional chromosomal karyotype 45,XX,-7,del(9)(p12),del(10)(q23)[8]/46,XX[12]; peripheral abnormalities were found to be specific for each twin, five for T1 blood analysis performed on T2 at that time was normal. Five and four for T2 (Figure 1A and Supplementary Figure 2). months after T1 onset, T2 was diagnosed with common ALL with The T2 relapse clone showed the same LOH 17q and 100% BM blasts infiltration, karyotype 45,XX, À 7,del(10)(q22)[14]. del(15)(q21.3) found in its diagnostic sample, and as expected, Immunoglobulin/T-cell receptor (Ig/TcR) monoclonal rearrange- several adjunctive CNAs as heterozygous deletions affecting single ments found at diagnosis (T1 and T2) and relapse (only T2) were genes (IKZF1, ETV6, C20orf94) (Supplementary Figure 1B), trisomy used to quantify the minimal residual disease (MRD) at three 21 and a complex rearrangement of chromosome 20 (Supple- follow-up time points (day þ 15, day þ 33 and day þ 78) mentary Figure 2). Deletions associated with TCR clonal rearrange- according to the AIEOP-BFM ALL2000 protocol.9 ments at diagnosis and relapse, visible and distinct for both twins, Twins were stratified according to the criteria described by are listed in Supplementary Table 5. Conter et al.10: T1 was classified as MRD high risk (HR), while T2 The role of NF1 mutations during clonal evolution of ALL is still was MRD intermediate (IR) risk. T1 was subjected to BM trans- not completely understood. We report, for the first time, the plantation 8 months after diagnosis and remained in clinical occurrence of LOH as a prenatal lesion predisposing to leukemia remission at 90 months, while T2 relapsed after 13 months in a pair of twins suffering from NF1. Furthermore, we have from diagnosis as common ALL and karyotype 47,XX,del(9) characterized the succession of events leading to leukemia from (p21), þ 21[11]/46,XX[5]. the prenatal phase in one of these NF1 twins. At diagnosis, T1 and T2 showed completely different Ig/TCR Indeed, the combination of genome-wide copy number analysis rearrangements as shown in Supplementary Table 2. Highly and Ig/Tcr rearrangements monitoring has led us to hypothesize a sensitive (X1.0 Â 10 À 4) cross analysis of diagnostic markers in model of clonal evolution of leukemia in NF1 twins (Figure 2). both the twins did not show any common rearrangements The uniqueness of 17q LOH as a common feature between twins (Supplementary Table 3). Three clonal rearrangements were also and its persistency from diagnosis to relapse in T2 suggests that 17q found in the T2 relapse sample (Supplementary Table 2) not LOH is most probably a single cell or a clonal event occurring related to the T2 diagnostic ones. prenatally in one twin. This conclusion is also supported by previous By rearrangement backtracking, we detected both T2 relapse data showing shared and identical gene fusions in other clonal rearrangements in the diagnostic sample of T2 but not T1. monozygotic twin pairs with concordant ALL;1 it is likely that They accounted for a clone corresponding to about 1% of cells blood cell chimeras might have facilitated the spread of clonal (1.8 Â 10 À 2 and 8 Â 10 À 3, respectively) (Supplementary Figure 1, leukemia or progeny from one twin to the other due to their Supplementary Table 3). monochorionic status in the uterus.1 The identification of the same The genome-wide CNA analysis (see Supplementary table 4) NF1 mutation (2763insA) in T1 and T2 confirmed it to be a primary indicated that T1 and T2 at diagnosis and T2 at relapse shared event linked to the disease that preceded LOH. Moreover, IgH and only one common aberration, namely the copy number neutral TCR rearrangements were clonally distinct. We cannot, therefore, loss of heterozygosity (LOH) of the whole long arm of rule out that both twins were at a greater risk of leukemia because

T2DX LOH 17q Vg2 Jg2.3 NF1+ del TCF12 VH1JH6 amp(2)(pter→p25.3) -7 VH3JH4 del(10)(q23.1) del(14)(q32.13)

NF1+ T2REL

NF1+ COMMON Vg9 Jg2.3 Vg9 Jg2.3 PROGENITOR NF1+ VH6JH4 VH6JH4 NF1+/- NF1+ NF1+ ~1% LOH 17q LOH 17q LOH 17q del TCF12 del TCF12 del TCF12 del(9)(p13.1) del IKZF1 T1DX del ETV6 LOH 17q LOH 17q del C20orf94 NF1+ LOH 20q LOH 17q +21 -7 Others (table S4) del BACH2 9p deletions Vg3 Jg1.3 del(10)(q23.33) VH4JH4 DH2JH4

Time (years) BIRTH 6 6,5 9,0 Figure 2. Clonal evolution in NF1 ALL twins. The time flow diagram indicates the hypothetical succession of genetic events during clonal evolution of ALL in NF1 twins.

Leukemia (2013) 1745 – 1791 & 2013 Macmillan Publishers Limited Letters to the Editor 1781 of their constitutive NF1 status; in this scenario, selective pressure AUTHOR CONTRIBUTIONS due to the loss of the normal NF1 allele resulted in both of them MG and AL performed SNP arrays, MG and CM performed cytogenetic analyses, losing the entire long arm of chromosome 17, but independently SS performed Ig/TcR clonality analyses and monitoring; GC designed the and along with uniparental disomy of the 17q carrying the research and supervised the contribution of all the authors; MG wrote the constitutive NF1 mutation. If that were to be the case, the paper, with the supervision of GC and AB; CM and CD follow the patients and 17q- could have arisen either pre- or postnatally. We note that provided clinical data. NF1-associated ALL (rather than CMML or AML) is extremely 6,11 rare. It is therefore unlikely that these twins developed ALL M Galbiati1,3, A Lettieri1,3, C Micalizzi2, S Songia1, independently. C Morerio2, A Biondi1, C Dufour2 and G Cazzaniga1 After birth, each clone had an independent evolution, as 1Centro Ricerca Tettamanti, Clinica Pediatrica, Universita` di shown by specific Ig/TCR rearrangements and the accumulation of Milano-Bicocca, Ospedale San Gerardo, Monza, Italy and different lesions. These findings further suggest that 17q 2UO Ematologia Clinica Sperimentale, LOH occurred very early during differentiation, before somatic IRCCS G. Gaslini, Genova, Italy recombination. The monosomy 7 (frequently described in E-mail: [email protected] myeloid disorders and rarely associated with lymphoid malig- 3These authors contributed equally to this work. nancies)11,12 should have occurred as an independent and postnatal acquisition, as sustained by the loss of this aberration at T2 relapse. REFERENCES Few additional aberrations were then required for the progres- 1 Greaves MF, Maia AT, Wiemels JL, Ford AM. Leukemia in twins: lessons in natural sion to full leukemia in both the twins. Interestingly, we found new history. Blood 2003; 102: 2321–2333. lesions herewith reported for the first time as deleted in leukemia, 2 Greaves M. Cancer stem cells: back to Darwin? Semin Cancer Biol 2010; 20: in particular, a 292-kb deletion on chromosome 6 involving the 65–70. BACH2 gene and a focal deletion of 130 kb on chromosome 15 3 Jett K, Friedman JM. Clinical and genetic aspects of neurofibromatosis 1. Genet involving the TCF12 gene. BACH2 is a human B-cell specific Med 2010; 12:1–11. transcriptional repressor; it is involved in B-cell differentiation and 4 Laycock-van Spyk S, Thomas N, Cooper DN, Upadhyaya M. Neurofibromatosis has been described as a tumor suppressor gene;13 TCF12 (also type 1-associated tumours: their somatic mutational spectrum and pathogenesis. named HEB) is a member of the basic helix–loop–helix E-protein Hum Genomics 2011; 5: 623–690. 14 5 Steinemann D, Arning L, Praulich I, Stuhrmann M, Hasle H, Stary J et al. Mitotic family and is involved in B- and T-cell commitment. The altered recombination and compound-heterozygous mutations are predominant expression of these genes could contribute to the differentiation NF1-inactivating mechanisms in children with juvenile myelomonocytic leukemia arrest and the uncontrolled proliferation of leukemic blasts. and neurofibromatosis type 1. Haematologica 2010; 95: 320–323. However, an in-depth study of these lesions might reveal new 6 Stiller CA, Chessells JM, Fitchett M. Neurofibromatosis and childhood leukaemia/ leukemia pathogenetic mechanisms. lymphoma: a population-based UKCCSG study. Br J Cancer 1994; 70: 969–972. Backtracking of Ig/TCR and CNA analysis of the T2 relapse 7 Balgobind BV, Van Vlierberghe P, van den Ouweland AM, Beverloo HB, sample strongly suggests that the relapse might have originated Terlouw-Kromosoeto JN, van Wering ER et al. Leukemia-associated NF1 from an intermediate stage preceding the first leukemia inactivation in patients with pediatric T-ALL and AML lacking evidence for presentation in T2 (Figure 2). This stage was formed by a clonal neurofibromatosis. Blood 2008; 111: 4322–4328. 8 Haferlach C, Grossmann V, Kohlmann A, Schindela S, Kern W, Schnittger S et al. cell population representing about 1% of leukemia at first Deletion of the tumor-suppressor gene NF1 occurs in 5% of myeloid malignancies presentation derived from a post-natal pre-leukemic clone with and is accompanied by a mutation in the remaining allele in half of the cases. a TCF12 deletion. This cell population underwent further clonal Leukemia 2012; 26: 834–839. evolution through the acquisition of additional chromosomal 9 van der Velden VH, Cazzaniga G, Schrauder A, Hancock J, Bader P, Panzer- aberrations involving known lesions, such as ETV6 and Ikaros Grumayer ER et al. Analysis of minimal residual disease by Ig/TCR gene deletions, the last one recently described as negative prognostic rearrangements: guidelines for interpretation of real-time quantitative PCR data. factor in pediatric BCP-ALL patients.15 In addition, it is likely that Leukemia 2007; 21: 604–611. other genes located in regions of CNA might have contributed to 10 Conter V, Bartram CR, Valsecchi MG, Schrauder A, Panzer-Grumayer R, Moricke A the progression to full leukemia. et al. Molecular response to treatment redefines all prognostic factors in children and adolescents with B-cell precursor acute lymphoblastic leukemia: In conclusion, the in-depth analysis of clonal and tumorigenic results in 3184 patients of the AIEOP-BFM ALL 2000 study. Blood 2010; 115: events in a twin pair has let us establish a chronological series of 3206–3214. transforming events that allows a pre-natal pre-leukemic clone in 11 Scrideli CA, Baruffi MR, Rogatto SR, Valera ET, Defavery R, Tone LG. B lineage a patient with a predisposing disease to become fully leukemic acute lymphoblastic leukemia transformation in a child with juvenile myelo- and, subsequently, to acquire additional lesions that could favor monocytic leukemia, type 1 neurofibromatosis and monosomy of chromo- the relapse of the disease. some 7. Possible implications in the leukemogenesis. Leuk Res 2003; 27: 371–374. 12 Cutting R, Taj M, Vora A. Childhood precursor B-cell acute lymphoblastic CONFLICT OF INTEREST leukaemia and monosomy 7 with phenotypic shift at relapse; evidence for the stem cell origins of monosomy 7. Br J Haematol 2008; 140: 584–586. The authors declare no conflict of interest. 13 Kobayashi S, Taki T, Chinen Y, Tsutsumi Y, Ohshiro M, Kobayashi T et al. Identi- fication of IGHCdelta-BACH2 fusion transcripts resulting from cryptic chromoso- mal rearrangements of 14q32 with 6q15 in aggressive B-cell lymphoma/leukemia. ACKNOWLEDGEMENTS Genes Chromosomes Cancer 2011; 50: 207–216. This paper has been partially supported by Fondazione Tettamanti, Fondazione 14 Quong MW, Romanow WJ, Murre C. E protein function in lymphocyte develop- Cariplo, AIRC (Associazione Italiana Ricerca sul Cancro) and MIUR (Ministero ment. Annu Rev Immunol 2002; 20: 301–322. dell’Istruzione, dell’Universita´ e della Ricerca), ERG spa, Cambiaso and Risso, 15 Mullighan CG, Su X, Zhang J, Radtke I, Phillips LA, Miller CB et al. Deletion of IKZF1 Rimorchiatori Riuniti, SAAR Depositi oleari portuali. and prognosis in acute lymphoblastic leukemia. N Engl J Med 2009; 360: 470–480.

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

& 2013 Macmillan Publishers Limited Leukemia (2013) 1745 – 1791