ORIGINAL ARTICLE the MLL Recombinome of Acute Leukemias

The MLL recombinome of acute leukemias

C Meyer1,23, B Schneider1,23, S Jakob1, S Strehl2, A Attarbaschi2, S Schnittger3, C Schoch3, MWJC Jansen4, JJM van Dongen4, ML den Boer5, R Pieters5, M-G Ennas6, E Angelucci7, U Koehl8, J Greil9, F Griesinger10, U zur Stadt11, C Eckert12, T Szczepan´ski13, FK Niggli14, BW Scha¨fer14, H Kempski15, HJM Brady15, J Zuna16, J Trka16, LL Nigro17, A Biondi18, E Delabesse19, E Macintyre19, M Stanulla20, M Schrappe21, OA Haas2, T Burmeister22, T Dingermann1, T Klingebiel8 and R Marschalek1

1Institute of Pharmaceutical Biology/ZAFES/Diagnostic Center of Acute Leukemia, University of Frankfurt, Frankfurt/Main, Germany; 2CCRI, Children’s Cancer Research Institute, Vienna, Austria; 3Laboratory for Leukemia Diagnostics, Department of Internal Medicine III, University Hospital Grosshadern, Ludwig-Maximilians-University, Munich, Germany; 4Department of Immunology, Erasmus MC, Rotterdam, The Netherlands; 5Department of Paediatric Oncology/Haematology, Erasmus MC, Sophia Children’s Hospital, Rotterdam, The Netherlands; 6Department of Cytomorphology, University of Cagliari, Cagliari, Italy; 7Haematology and Oncology Hospital ‘A Businco’, Cagliari, Italy; 8Pediatric Hematology and Oncology, University Frankfurt, Frankfurt, Germany; 9Department of Pediatric Hematology and Oncology, University Children’s Hospital, Tuebingen, Germany; 10Department of Hematology and Oncology, Goettingen, Germany; 11Department of Pediatric Hematology and Oncology, Eppendorf, Hamburg, Germany; 12Department of Pediatric Oncology and Hematology, Charite´ – Virchow Campus, Berlin, Germany; 13Department of Pediatric Hematology and Oncology, Silesian Academy of Medicine, Zabrze, Poland; 14University Children’s Hospital, Department of Oncology, Zuerich, Switzerland; 15Molecular Haematology and Cancer Biology Unit, Institute of Child Health, University College London, London, UK; 16Department of Paediatric Haematology/Oncology, Charles University Prague, Prague, Czech Republic; 17Center of Pediatric Hematology Oncology, University of Catania, Catania, Italy; 18M.Tettamanti Research Center, University of Milano-Bicocca, Monza, Italy; 19Biological Hematology, AP-HP and INSERM EMIU210, Paris, France; 20Paediatric Haematology/Oncology, Medical School of Hannover, Hannover, Germany; 21Paediatric Department, University of Schleswig-Holstein, Campus Kiel, Germany and 22Charite´ – Benjamin Franklin Campus, Med. Klinik III, Berlin, Germany

Chromosomal rearrangements of the human MLL gene are a leukemias.1,2 For instance, the presence of MLL rearrangements hallmark for aggressive (high-risk) pediatric, adult and therapy- is an independent dismal prognostic factor in infant anti- associated acute leukemias. These patients need to be lymphoblastic factor (ALL), whereas childhood ALL patients identified in order to subject these patients to appropriate . therapy regimen. A recently developed long-distance inverse with MLL AF4 fusions are usually treated according to high-risk PCR method was applied to genomic DNA isolated from protocols. Thus, the identification of MLL gene fusions is individual acute leukemia patients in order to identify chromo- necessary for rapid clinical decisions resulting in specific somal rearrangements of the human MLL gene. We present data therapy regimens. Current procedures to identify MLL rearran- of the molecular characterization of 414 samples obtained from gements include cytogenetic analysis, fluorescence in situ 272 pediatric and 142 adult leukemia patients. The precise localization of genomic breakpoints within the MLL gene and hybridization (FISH) experiments (e.g. MLL split-signal FISH), the involved translocation partner genes (TPGs) was deter- specific reverse transcriptase-PCR (RT-PCR) and genomic PCR mined and several new TPGs were identified. The combined methods. This repertoire of technologies was recently extended data of our study and published data revealed a total of 87 by a long-distance inverse PCR (LDI-PCR) method that uses different MLL rearrangements of which 51 TPGs are now small amounts of genomic DNA to determine any type of MLL characterized at the molecular level. Interestingly, the four gene rearrangement at the molecular level.3 This includes most frequently found TPGs (AF4, AF9, ENL and AF10) encode nuclear proteins that are part of a protein network involved in chromosomal translocations, gene internal duplications, chro- histone H3K79 methylation. Thus, translocations of the MLL mosome 11q deletions or inversions, and MLL gene insertions gene, by itself coding for a histone H3K4 methyltransferase, are into other chromosomes, or vice versa, the insertion of presumably not randomly chosen, rather functionally selected. chromatin material into the MLL gene. Leukemia (2006) 20, 777–784. doi:10.1038/sj.leu.2404150; To gain insight into the frequency of MLL rearrangements, published online 2 March 2006 unscreened and prescreened pediatric and adult leukemia Keywords: MLL; chromosomal translocations; partner genes; acute leukemia patients were analyzed. Prescreening tests (cytogenetic analysis, FISH, Southern blot, RT-PCR or NG2-positivity) were performed at different European centers, where acute leukemia patients are enrolled in different study groups (Interfant-99, AMLCG, Introduction GMALL). With the exception of a few patients, all prescreened MLL rearrangements were successfully analyzed and patient- Chromosomal rearrangements involving the human MLL gene specific MLL fusion sequences were obtained. Data obtained are recurrently associated with the disease phenotype of acute from the literature and results obtained in this study are summarized in a color map where all 51 translocation partner Correspondence: Dr R Marschalek, Institute of Pharmaceutical genes (TPGs) and their specific breakpoint regions have been Biology/ZAFES, University of Frankfurt, Biocenter, N230, 303 Marie- assigned. The applied color code will enable all investigators to Curie Str. 9, D-60439 Frankfurt/Main, Germany. identify compatible intron–intron fusions between the MLL gene E-mail: [email protected] 23These authors contributed equally to this work. and all yet characterized TPGs. Moreover, we provide a Received 24 November 2005; revised 3 January 2006; accepted 11 complete list of 87 MLL rearrangements of which 36 still await January 2006; published online 2 March 2006 molecular characterization. The MLL recombinome C Meyer et al 778 Materials and methods decapping enzyme), BCL9L and ARHGEF17. With the exception of BCL9L and ARHGEF17, all of them have been reported in a Patient material technical paper by the authors.3 The BCL9L gene encodes a Genomic DNA was isolated from bone marrrow and/or structural constituent of ribosomes (with similarity to the peripheral blood samples of all patients and sent to our center. Drosophila legless gene) and is located on chromosome Patient samples were obtained from the Interfant-99 study group 11q23.5 The fusion between MLL and BCL9L deletes a (Rotterdam, The Netherlands), the AMLCG study group (Mu- chromosomal area of about 300 kb and fused MLL intron 8 nich, Germany) and the GMALL study group (Berlin, Germany). tail-to-tail with the 30-non-translated region (30-NTR) of BCL9L. Informed consent was obtained from all patients or patients’ Therefore, no functional fusion protein can be produced. The parents/legal guardians and control individuals. ARHGEF17 gene is located on chromosome 11q13 and encodes a protein of 1510 amino acids (164 kDa) that belongs to the family of Rho guanine nucleotide exchange factors involved in Long-distance inverse PCR experiments signaling pathways.6 The in-frame fusion between MLL intron All DNA samples were treated and analyzed as described.3 12 and ARHGEF17 intron 1 occurred on both fusion alleles and Briefly, 1 mg genomic patient DNA was digested with restriction involved two homologous chromosomes 11. enzymes and religated to form DNA circles before LDI-PCR analyses. Restriction polymorphic PCR amplimers were isolated from the gel and subjected to DNA sequence analyses to obtain Different possibilities to cause genetic MLL aberrations the patient-specific fusion sequences. In general, human MLL rearrangements are initiated by DNA cleavage or a DNA damage situation, which induces DNA repair via the non-homologous end joining (NHEJ) DNA repair Results pathway.7,8 Although the MLL gene is predominantly involved in reciprocal chromosomal translocations (Figure 1a), other Identification of MLL rearrangements genetic rearrangements were observed in this study or were Acute leukemia patients carrying MLL rearrangements are not described in the literature. Gene-internal partial tandem exceeding a total of 800 cases per year in Europe (about 300 duplications (PTD) of specific MLL gene portions (duplication pediatric and about 500 adult leukemia patients). To analyze the of introns 2–9, 2–11, 4–9, 4–11 or 3–8; Figure 1b) are frequently recombinome of the human MLL gene, 414 acute leukemia observed in AML patients.9–11 MLL PTD are being discussed to samples from different European centers were analyzed over a mediate dimerization of the MLL N-terminus, a process that period of 24 months (272 pediatric and 142 adult patients). One seems to be sufficient to mediate leukemogenic transforma- Hundred and seventy-six out of 414 leukemia samples were not tion.12 The third possibility is deletions on the long arm of prescreened for MLL aberrations or were classified as ‘MLL- chromosome 11, which affect only portions of the MLL gene, negative’ cases. A total of seven MLL rearrangements are but always lead to functional MLL fusion genes (TPG . MLL or identified in the group of not prescreened pediatric leukemia MLL . TPG, depending on whether the deletion involves genes patients, which corresponds to the expected frequency of located centromeric or telomeric relative to the MLL gene; 5–10%. By contrast, 238 patients with acute leukemia were Figure 1c). In the latter case, only the derivative(11) chromo- prescreened for MLL rearrangements (see Table 1a). These ‘MLL- some is created (e.g. ARHGEF12). A fourth mechanism is the positive’ leukemias correspond to 107 pediatric and 131 adult inversion of a chromosome 11 segment, represented by the acute leukemia patients. From these patients, 206 were MLL . PICALM fusion (Figure 1c). A more complicated picture successfully analyzed, whereas for 32 acute leukemia patients, emerges when chromosome 11 material (including portions of no MLL rearrangement was detected (false-positive prescreen- the MLL gene) is inserted into other chromosomes, or vice versa, ing: 3/32; prescreened only by NG2-positivity: 9/32; poor DNA the insertion of chromosome material (including portions of a quality: 3/32; failure to produce a positive result: 17/32). The TPG) into the breakpoint cluster region of the MLL gene (e.g. latter failures (17/32) might be explained either by false-positive MLLT10; Figure 1d). A combination of both events leads to prescreenings or by the intrinsic limitation of the applied reciprocal insertions. In rare cases, the described insertion method. which allows the identification of MLL rearrangements mechanism involves the MLL gene and two different TPGs, a only when they occurred within the breakpoint cluster region of mechanism that leads to complex three-partner translocations. the MLL gene. However, chromosomal breakpoints have The insertion of chromatin fragments is necessary because some already been mapped outside the breakpoint region.4 TPGs are not transcribed in the telomeric direction, as the MLL The most frequent rearrangements were t(4;11)(q21;q23) gene. Therefore, a ‘simple’ reciprocal translocation would lead involving the MLLT2 (AF4) gene, t(9;11)(p22;q23) involving to a head-to-head or tail-to-tail fusion of the MLL gene with the MLLT3 (AF9) gene, t(11;19)(q23;p13.3) involving the MLLT1 these TPGs. All recombination events – or any combination (ENL) gene, t(10;11)(p12;q23) involving the MLLT10 (AF10) thereof – are sufficient to explain all yet known MLL rearrange- gene and t(6;11)(q27;q23) involving the MLLT4 (AF6) gene, ments. The different genetic mechanisms convert the MLL respectively (see Table 1b). These translocations account for protein into an oncogenic derivative, necessary for the onset of a about 85% of all investigated leukemia samples. All identified preleukemic or leukemic clone, and subsequently, to initiate TPGs in pediatric and adult leukemia patients are summarized acute leukemia. in Table 1b.

MLL rearrangements and the ‘intron code’ Novel translocation partner genes On the basis of results and data obtained from the literature, a A total of seven new TPGs were discovered: ACACA total of 51 TPGs are now characterized at the molecular level. (acetyl-CoA carboxylase alpha), SELB (elongation factor for These 51 TPGs are summarized in Figure 2, where all MLL selenocysteine), SMAP1 (glycoprotein involved in erythropoietic fusion genes are shown schematically with their speciﬁc exon/ development), TIRAP (TLR4 adaptor protein), DCPS (mRNA intron structure. In addition, all introns were artiﬁcially

Leukemia The MLL recombinome C Meyer et al 779 Table 1 Combined data of analyzed patient material (a) Diagnostic study centers in alphabetical order and (b) Distribution of MLL partner genes in pediatric and adult leukemia patients

(a) MLL-positive MLL-negative

Center Total P A Pre Np Total P A Total Np Fp N/C DNA Fail

Berlin 101 1 100 101 F 87 1 86 14 F 37F 4 Cagliari 11 F 11 F 11 FFF11 11 FF F F Catania 22 22 F 21 1 18 18 F 41FF F 3 Frankfurt 11 11 F 47 44F 77FF F F Goettingen 4 2 2 4 F 22F 2 FFF 2 F Hamburg 1 1 F 1 F 11FFFFFFF London 12 12 F 84 88F 43FF F 1 Munich 9 F 99 F 4 F 45FFF F 5 Paris 26 10 16 26 F 26 10 16 FFFFFF Prague 11 11 F 11 F 10 10 F 1 FF 1 FF Rotterdam 24 23 1 20 4 18 17 1 6 4 FF 11 Tuebingen 150 150 F 1 149 7 7 F 143 143 FF F F Vienna 27 24 3 27 F 24 21 3 3 FFF F 3 Zabrze 3 3 F 3 F 22F 1 FF 1 FF Zuerich 2 2 F 2 F 22FFFFFFF

Total 414 272 142 238 176 213 103 110 201 169 3 9 3 17

Pediatric 272 107 165 169 158 2 1 8 Adult 142 131 11 32 11 3 7 2 9

(b) Pediatric Adult

Gene Total Total AML ALL Pediatric ext AL* Total AML ALL

MLLT2 (AF4) 102 23 F 23 33* 79 F 79 MLLT3 (AF9) 33 24 12 12 20* 9 7 2 MLLT1 (ENL) 26 16 F 16 16* 10 3 7 MLLT10 (AF10) 11 10 6 4 1 1 F MLLT4 (AF6) 8 7 5 2 1 1 F ELL 6 4 4 F 22F EPS15 3 3 1 2 FFF MLLT6 (AF17) 3 1 1 F 22F ABI1 1 1 1 FFFF ACACA 1 1 1 FFFF AF15Q14 1 FFF 11F AF1Q 1 1 F 1 FFF ARHGEF17 1 1 1 FFFF BCL9L 1 1 F 1 FFF CXXC6 1 FFF 11F LAF4 1 1 F 1 FFF PICALM 1 1 F 1 F -FF PNUTL 1 FFF 11F RPS3 1 1 1 FFFF SELB 1 1 F 1 FFF SEPT6 1 1 1 FFFF SMAP1 1 FFF 11F TIRAP/DCPS 1 FFF 11F 2p21 1 1 F 1 FFF 10p21 1 1 1 FFFF 22q21 1 1 F 1 FFF MLL PTD 1 FFF 11F MLL inversion 1 1 1 FFFF Internal deletion 1 1 F 1 FFF

Total 213 103 36 67 69* 110 22 88 P: pediatric leukemia patients; A: adult leukemia patients; Pre: pre-screened patient samples; Np: not pre-screened; Fp: veriﬁed false-positive; N/C: pre-screened either by NG-2 positivity or by cytogenetic experiments; DNA: poor DNA quality; fail: failure to detect an MLL rearrangement. AML: acute myeloid leukemia; ALL: acute lymphoblastic leukemia.*External pediatric patients: data obtained from the diagnostic unit at the University of Erlangen-Nuernberg, analyzing the I-BFM pediatric patients carrying pre-screened t(4;11), t(9;11) and t(11;19) chromosomal translocations; the ﬁve most frequent MLL translocations are shown on top. Gene names are given. MLL PTD: partial tandem duplication of the MLL gene.

Leukemia The MLL recombinome C Meyer et al 780 color-coded: colorless introns are named ‘type 0’ introns, whereas introns disrupt the three nucleotides of a given codon either by red and blue introns are ‘type 1’ and ‘type 2’ introns. A ‘type 0’ 1-intron-2 or 2-intron-1, respectively (see also Marschalek intron disrupts a given open reading frame exactly between two et al.13). Chromosomal rearrangements within the MLL gene triplet codons (triplet–intron–triplet), whereas ‘type 1’ and ‘type 2’ occur mostly in ‘type 0’ introns (between MLL exons 9 and 14)

Figure 1 For caption refer page 781

Leukemia The MLL recombinome C Meyer et al 781 and a corresponding ‘type 0’ intron of a recombination partner however, 36 genetic loci were identified only by cytogenetic gene. Figure 2 allows also to identify ‘compatible’ gene fusions, experiments and still await their molecular characterization. as it predicts MLL fusions at the mRNA level in cases where According to our data, the most frequent TPGs in MLL- introns of different colors were recombined (e.g. MLL fusions mediated acute leukemias are MLLT2 (AF4), MLLT3 (AF9), with RPS3, MLLT6 or MSF). In these rare cases, alternative splice MLLT1 (ENL), MLLT10 (AF10) and MLLT4 (AF6), which are events are fusing compatible exons – flanked by identical intron responsible for about 85% of all investigated rearrangements color – to guarantee functional fusions. To this end, Figure 2 is (see Table 1b, pediatric and adult patients). A second diagnostic helpful to design appropriate RT-PCR experiments. center at the University of Erlangen-Nuremberg has investigated only pediatric leukemia patients prescreened for t(4;11), t(9;11) and t(11;19) translocations. They have successfully analyzed 33 t(4;11), 20 t(9;11) and 16 t(11;19) leukemia patients in parallel Arbitrary MLL fusions and ‘spliced fusions’ to our study (see Table 1b, marked by asterisk) by using different In very rare cases, the MLL gene is fused to TPGs without genomic PCR approaches.14,15 Upon taking these data together, producing functional products. In one patient who carried a the five different chromosomal translocations mentioned above reciprocal translocation between the MLL and the SELB gene, no account for about 90% of all MLL-positive leukemias. Routine detectable fusion mRNA species from the two reciprocal fusion diagnostic methods based on RT-PCR might be restricted to products MLL . SELB and SELB . MLL were detected.3 A second these few TPGs in order to identify the vast majority of MLL gene unusual observation was identified in a patient with an rearrangements. MLL . TIRAP fusion resulting from an interstitial deletion. The An important result of this study is the establishment of involved TIRAP intron 7 was located within the 30-untranslated patient-specific DNA sequences that can now be used for region of the TIRAP gene (TIRAP–11q24; Figure 2). Indeed, no monitoring of minimal residual disease (MRD) by quantitative MLL-TIRAP fusion mRNA could be identified. However, a gene PCR techniques. In general, chromosomal fusion alleles located about 5 kb telomeric of TIRAP, the DCPS gene, was used represent the most reliable markers for MRD studies and have for splicing to generate an in-frame MLL fusion.3 This ‘spliced several advantages over surrogate markers such as IgH or T-cell fusion’ exists only at the RNA level, and thus, establishes receptor rearrangements (as they may represent tumor sub- another mechanism by which a functional MLL fusion can be populations). Owing to the fact that a given MLL fusion allele is generated. A similar situation was found for the MLL . BCL9L genetically stable and a mono-allelic marker for each tumor fusion mentioned above. As the MLL gene was fused tail-to-tail cell, a more reliable quantification and tracing of residual tumor with the 30-NTR of BCL9L, no functional fusion mRNA could be cells becomes possible. For each of these 213 acute leukemia encoded by the genomic fusion of both genes. Experimental patients, at least one MLL fusion allele was identified and attempts to identify spliced fusions with telomer located genes characterized by sequencing. A first prospective study was of BCL9L (FOXR1, etc.) are still ongoing. already initiated and verified the reliability of these genomic markers for MRD monitoring.16 Therefore, the use of these MRD markers will contribute to stratification, improved treatment and Discussion outcome of leukemia patients. The analysis of the MLL recombinome allows to classify MLL Several conclusions can be drawn from this study. Genomic fusion partner genes into functional categories. Interestingly, the DNA turned out to be a good source to identify different MLL most frequent TPGs in MLL translocations encode nuclear rearrangements. The applied method allowed the identification proteins (AF4, AF9, ENL and AF10) that were recently identified of chromosomal translocations, MLL gene-internal duplications, to belong to the same nuclear protein network (see Figure 3). chromosome 11 inversions, chromosomal 11 deletions and the This may indicate that TPGs are not randomly chosen, rather insertion of chromosome 11 material into other chromosomes, functionally selected. Briefly, AF4 and AF9 colocalize at distinct or vice versa, the insertion of chromatin material of other nuclear foci,17 and the disruption of the AF4/AF9 protein chromosomes into the MLL gene. Moreover, it allowed to interaction induces apoptosis in t(4;11) cells.18 By contrast, the identify unknown TPGs and/or complex rearrangements that AF4 protein seems to be required for growth and differentiation cannot be analyzed by the resolution of cytogenetic or FISH of lymphocytic progenitors.19 ENL was shown to interact analyses. The successful analysis of more than 200 MLL fusion directly with the AF4 and AF10 proteins,20 and in addition, alleles led to the discovery of new TPGs and a large variety of AF10 binds to hDOT1L.21 The hDOT1L protein is a non-SET- different mechanisms that fuse the MLL gene to different TPGs or domain protein that is able to mediate the methylation of lysine MLL-spliced fusions in order to create oncogenic protein 79 of histone H3 proteins. H3K79me seems to be a prerequisite variants. Within the past years, 87 genetic aberrations involving for elongation of RNA polymerase II. Both the MLL and hDOT1L the human MLL gene located on chromosome 11 band q23 have protein guarantee a site-specific histone methylation pattern that been described (see Table 2 in Supplementary Information). play a major role in transcriptional initiation, elongation and, Fifty-one TPGs are now characterized on the molecular level; subsequently, maintenance of transcription.22–24 The limiting

Figure 1 General possibilities to generate MLL rearrangements. The MLL gene is shown as a red arrowhead, whereas potential translocation partner genes (TPGs) are shown as green, blue or gray arrowheads; der(11): derivative (11) chromosome; der(Z): reciprocal derivative (Z) chromosome (Z: any chromosome). Small colored triangles (red, green or blue) at the side of chromosomes: recombination sites. Numbers in the right lower corners indicate the amount of these genetic events out of the 51 TPGs summarized in this paper. (a) Chromosomal translocations. Translocation partner genes of the human MLL gene are transcriptionally orientated into the directions of telomers; two possibilities are shown: TPGs are located either on the p (left) or on the q arm (right). (b) Gene-internal partial tandem duplications of the MLL gene. Duplicated gene areas include MLL introns 2–9, 2–11, 4–9, 4–11 or 3–8. (c) MLL inversions and deletions. Inversions are created by turning around a chromosome 11 fragment. Deletions are caused by fusion of MLL with a telomer-located gene (MLL . TPG fusions). (d) Insertions of a TPG-containing chromosomal fragment into the MLL gene, or vice versa, the insertion of an MLL-containing chromosomal fragment into another chromosome. If chromatin material of both chromosomes involved is reciprocally inserted, two fusion genes are created. In very rare cases, a third TPG (gray arrowhead) is involved to cause complex three-partner gene fusions (shown only on the right–hand side). For Figure refer page 780

Leukemia The MLL recombinome C Meyer et al 782

Figure 2 For caption refer page 783

Leukemia The MLL recombinome C Meyer et al 783 Figure 2 Molecular-characterized translocation partner genes of the human MLL gene. All currently known TPGs are shown according to their chromosomal localization (listed from chromosome 1 to X). The gene names are given on the left. Chromosomal positions are given on the right as well as the genetic abnormality that is leading to a functional MLL fusion. Top: MLL gene structure with the two published exon nomenlatures. Below: exon/intron structures of all known TPGs. Introns are shown in white, red or blue, depending on the speciﬁc intron type (‘type 0’ intron ¼ white; ‘type 1’ intron ¼ red; ‘type 2’ intron ¼ blue). The MLL breakpoint cluster region of the MLL gene contains only ‘type 0’ introns. Compatible fusions between MLL and TPGs occur predominantly in gene areas containing a ‘type 0’ intron. Chromosomal positions and the type of genetic alterations are shown on the right (TL ¼ translocation; Ins ¼ insertion; Del ¼ deletion; Inv ¼ inversion; Spl ¼ spliced fusion). Red ﬂashes indicates particular introns (or exons) of TPGs found to be involved in MLL translocations. Light gray boxes: exons of non-translated regions. For Figure refer page 782

Figure 3 The AF4/AF9/ENL/AF10/DOT1L protein network links epigenetics to leukemogenesis. (a) MLL, EZH and SUVH39 are modifying histone H3 proteins by their specific methyltransferase activity. Subsequent DNA methylation processes via DNMT3 and DNMT1 (de novo and maintenance) are silencing the chromatin. HP1: heterochromatin protein 1; Pc: polycomb. (b) The most frequent MLL translocation partner genes encode nuclear proteins that form a protein network: AF4 interacts with AF9 and ENL. In addition, ENL binds to AF10 and the non-Set-domain protein hDOT1L. The latter protein is able to methylate lysine 79 of histone H3 proteins (H3K79). In conjunction with histone H3 lysine 4 methylation (H3K4), this leads to active chromatin and allows elongation of RNA polymerase II. factor of this complex seems to be the AF4 protein, as it has a Reinald Repp, Thorsten Langer, Jo¨rn-D Beck, Markus Metzler und very low abundance in mammalian cells owing to its interaction Thomas Leis for providing and sharing unpublished information of with the two E3 ubiquitin ligases SIAH1 and SIAH2.25 On taking their ongoing study (Grant 2002.032.1 from the Wilhelm Sander together, disruption of either component of this protein network Foundation) and Michael Karas for critical comments. This study might compromise important cellular functions. is supported by Grant GEN-AU Child, GZ 200.071/3-VI/2a/2002 Although most MLL translocations are associated with poor to OAH and Grant 2001.061.1 from the Wilhelm-Sander- outcome in infant and childhood acute leukemia, it can be Foundation to RM, TD and TK. assumed that a systematic analysis of the MLL recombinome will allow to draw conclusions on certain aspects of hemato- References malignant transformation processes. The analyses of different MLL fusion partner genes may help to categorize those for (1) 1 Pui CH, Gaynon PS, Boyett JM, Chessells JM, Baruchel A, Kamps their subcellular localization, (2) their cellular function, (3) W et al. Outcome of treatment in childhood acute lymphoblastic specific protein domain structures (e.g. dimerization domains) leukaemia with rearrangements of the 11q23 chromosomal region. and, finally, (4) their ability to interact with other proteins. These Lancet 2002; 359: 1909–1915. classifications have to be supplemented by functional studies 2 Pui CH, Chessells JM, Camitta B, Baruchel A, Biondi A, Boyett JM et al. Clinical heterogeneity in childhood acute lymphoblastic attempting to demonstrate the oncogenicity of different MLL leukemia with 11q23 rearrangements. Leukemia 2003; 17: fusions in the future by retroviral transduction of hematopoietic 700–706. stem/precursor cells. At the end, this will help to classify the 3 Meyer C, Schneider B, Reichel M, Angermueller S, Strehl S, large variety of different MLL translocations into different risk Schnittger S et al. Diagnostic tool for the identification of MLL groups, and thus will lead to a better stratification and treatment rearrangements including unknown partner genes. Proc Natl Acad of leukemia patients. Sci USA 2005; 102: 449–454. 4 Reichel M, Gillert E, Angermuller S, Hensel JP, Heidel F, Lode M et al. Biased distribution of chromosomal breakpoints involving the MLL gene in infants versus children and adults with t(4;11) ALL. Acknowledgements Oncogene 2001; 20: 2900–2907. 5 Katoh M, Katoh M. Identification and characterization of human This work was made possible by and conducted within the BCL9L gene and mouse Bcl9l gene in silico. Int J Mol Med 2003; framework of the International BFM Study Group. We thank 12: 643–649.

Leukemia The MLL recombinome C Meyer et al 784 6 Teuffel O, Betts DR, Thali M, Eberle D, Meyer C, Schneider B et al. throughput method for detection and sequencing genomic fusion Clonal expansion of a new MLL rearrangement in the absence of sites in t(4;11). Br J Haematol 2004; 124: 47–54. leukemia. Blood 2005; 105: 4151–4152. 16 Burmeister T, Marschalek R, Schneider B, Meyer C, Go¨kbuget N, 7 Reichel M, Gillert E, Nilson I, Siegler G, Greil J, Fey GH et al. Fine Schwartz S et al. Monitoring minimal residual disease by structure of translocation breakpoints in leukemic blasts with quantification of genomic chromosomal breakpoint sequences in chromosomal translocation t(4;11): the DNA damage-repair model acute leukemias with MLL aberrations. Leukemia 2006, 19 of translocation. Oncogene 1998; 17: 3035–3044. January, [E-pub ahead of print]. 8 Richardson C, Jasin M. Frequent chromosomal translocations 17 Erfurth F, Hemenway CS, de Erkenez AC, Domer P. MLL fusion induced by DNA double-strand breaks. Nature 2000; 405: partners AF4 and AF9 interact at subnuclear foci. Leukemia 2004; 697–700. 18: 92–102. 9 Schichman SA, Caligiuri MA, Gu Y, Strout MP, Canaani E, 18 Srinivasan RS, Nesbit JB, Marrero L, Erfurth F, LaRussa VF, Bloomfield CD et al. ALL-1 partial duplication in acute leukemia. Hemenway CS. The synthetic peptide PFWT disrupts AF4–AF9 Proc Natl Acad Sci USA 1994; 91: 6236–6239. protein complexes and induces apoptosis in t(4;11) leukemia cells. 10 Caligiuri MA, Strout MP, Schichman SA, Mrozek K, Arthur DC, Leukemia 2004; 18: 1364–1372. Herzig GP et al. Partial tandem duplication of ALL1 as a recurrent 19 Isnard P, Core N, Naquet P, Djabali M. Altered lymphoid molecular defect in acute myeloid leukemia with trisomy 11. development in mice deficient for the mAF4 proto-oncogene. Cancer Res 1996; 56: 1418–1425. Blood 2000; 96: 705–710. 11 Schnittger S, Kinkelin U, Schoch C, Heinecke A, Haase D, 20 Zeisig DT, Bittner CB, Zeisig BB, Garcia-Cuellar MP, Hess JL, Haferlach T et al. Screening for MLL tandem duplication in 387 Slany RK. The eleven–nineteen-leukemia protein ENL connects unselected patients with AML identify a prognostically unfavorable nuclear MLL fusion partners with chromatin. Oncogene 2005; 24: subset of AML. Leukemia 2000; 14: 796–804. 5525–5532. 12 Martin ME, Milne TA, Bloyer S, Galoian K, Shen W, Gibbs D et al. 21 Okada Y, Feng Q, Lin Y, Jiang Q, Li Y, Coffield VM et al. hDOT1L Dimerization of MLL fusion proteins immortalizes hematopoietic links histone methylation to leukemogenesis. Cell 2005; 121: 167–178. cells. Cancer Cell 2003; 4: 197–207. 22 Jenuwein T, Allis CD. Translating the histone code. Science 2001; 13 Marschalek R, Nilson I, Lochner K, Greim R, Siegler G, Greil J 293: 1074–1080. et al. The structure of the human ALL-1/MLL/HRX gene. Leuk 23 Lachner M, Jenuwein T. The many faces of histone lysine Lymphoma 1997; 27: 417–428. methylation. Curr Opin Cell Biol 2002; 14: 286–298. 14 Langer T, Metzler M, Reinhardt D, Viehmann S, Borkhardt A, 24 Zhang Y, Reinberg D. Transcription regulation by histone Reichel M et al. Analysis of t(9;11) chromosomal breakpoint methylation: interplay between different covalent modifications sequences in childhood acute leukemia: almost identical MLL of the core histone tails. Genes Dev 2001; 15: 2343–2360. breakpoints in therapy-related AML after treatment without 25 Bursen A, Moritz S, Gaussmann A, Dingermann T, Marschalek R. etoposides. Genes Chrom Cancer 2003; 36: 393–401. Interaction of AF4 wildtype and AF4 Á MLL fusion protein with 15 Metzler M, Brehm U, Langer T, Viehmann S, Borkhardt A, Stanulla SIAH proteins: indication for t(4;11) pathobiology? Oncogene M et al. Asymmetric multiplex-polymerase chain reaction – a high 2004; 23: 6237–6249.

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

Leukemia