Leukemia (2013) 27, 1745–1791 & 2013 Macmillan Publishers Limited All rights reserved 0887-6924/13 www.nature.com/leu

LETTERS TO THE EDITOR Recurrent involvement of ring-type zinc finger genes in complex molecular rearrangements in childhood acute myelogeneous leukemia with translocation t(10;11)(p12;q23)

Leukemia (2013) 27, 1745–1791; doi:10.1038/leu.2013.1 the MLL/MLLT10 fusion gene had been detected by conventional FISH and RT-PCR (Table S1). A specimen of the initial and remission sample was obtained in six children; in three of them a Complex rearrangements involving the MLL gene on chromosome relapse sample was available and additionally analyzed after 11q23 and MLLT10 on 10p have been reported in 15% of pediatric informed consent. By analyzing paired-end reads, we were able to patients with MLL rearranged acute myelogeneous leukemia describe these alterations in depth, revealing the precise pattern (AML). Owing to the opposite direction of MLL and MLLT10 of molecular rearrangement and finding other involved chromo- rearrangements (inversion and subsequent translocation or somes. For paired-end sequencing, DNA was isolated from insertion) with three or more breaks are required to result in a peripheral blood lymphocytes, and fragment libraries with a fusion gene.1,2 There are several additional reported median insert size of 450 bp were prepared. Samples of patients recombination partners of the MLL gene, in which a simple 1–3 were sequenced on a GAIIx platform; samples of patients 4–6 reciprocal translocation is insufficient due to incompatible on a HiSeq 2000 (Illumina Inc., San Diego, CA, USA). A orientation.3 However, even more complex rearrangements bioinformatical pipeline for read alignment, removal of duplicate consisting of additional involved chromosomes have been reads and variant detection was applied to find structural described in t(10;11).4 Chromosomal translocations t(10;11) alterations based on mapping coordinates, insert size and (p12;q14–23) always lead to fusion of the 50end of MLL and orientation of reads. By sequencing the remission sample we 30end of MLLT10. The MLLT10 fusion partner codes for two excluded inherited sequence variants. Copy number variations adjacent alpha helical domains and a leucine zipper motif were detected by coverage normalization, and computation of embedded within a region of 82 amino acids, proposed to be copy number ratios between diseased and remission sample. indispensable for induction of leukemia.5,6 The gold standard to Conventional PCR and capillary sequencing validated selected detect suspected translocations in AML patients are fluorescence translocations detected by paired-end read analysis. For further in situ hybridization (FISH) studies and subsequent reverse methodical details we refer to our supplemental file. transcriptase-polymerase chain reaction (RT-PCR). However, next Our pipeline was able to reveal the defining MLL/MLLT10 generation sequencing has emerged as a powerful tool to translocation with its precise genomic coordinates in each understand complex genetic traits. Here, we applied low- leukemic sample (except relapse in patient 4). Consistent with coverage whole-genome sequencing on six pediatric patients past publications we observed heterogeneity in breakpoints with t(10;11) AML (FAB-M4/M5), in which MLL rearrangement and within the breakpoint cluster regions of MLL and MLLT10,

Table 1. Upper rows: breakpoint (bp) heterogeneity and features in MLL/MLLT10 translocations

Patient 1 2 3 4 5 6 Bp in MLLT10 p12.31, intron p12.31, intron 5–6 p12.31, p12.31, intron p12.31, p12.31, intron 9–10 8–9 intron 5–6 intron 9–10 8–9 Bp in MLL q23.3, intron q23.3, intron 8–9 q23.3, q23.3, intron q23.3, intron q23.3, intron 8–9 10–11 intron 8–9 8–9 8–9 Second break q14.1, no gene q23.3, no gene q12.1, q23.3 RNF214, q14.3, no q13.4 RNF169, Chromosome 11 no gene intron 5–6 gene intron 2–3 Features Inversion and Inversion and translocation, CNV Insertion Insertion, Inversion Inversion and partial translocation, ( þ ) chromosome 1,13,21 ( À ) chromosome and insertion, CNV chromosome chromosome: 2 22 involved translocation ( þ )chromosome 8 16 involved

IR I I I RIR I Translocations 24 21 36 412 6 1 11 17 5 Deletions 3827 3108 1193 49 21 25 27 14 1 Inversions 15 13 14 16 0 0 11 12 3 Abbreviations: CNV, copy number variation; I, initial; R, relapse. In all patients, paired-end sequencing revealed the pattern of molecular rearrangement. ( þ ) indicates gains, ( À ) losses. Lower rows: number of detected structural variants. Patients 1–3 were sequenced on the GAIIx with a significant lower coverage, thus variants with at least two supporting reads have been counted; in patients 4–6 (HiSeq2000) variants had at least three supporting reads.

Accepted article preview online 9 January 2013; advance online publication, 15 February 2013 Letters to the Editor 1746

Figure 1. Patient 4 (upper panel) and patient 6 (lower panel) (a) FISH analysis: green color of the MLL probe identifies the proximal 50 part of the MLL gene, red color identifies the distal 30part of the gene. The MLL probe shows a MLL split signal on different chromosomes, on 10p and 11q, respectively, indicating an insertion in patient 4 and 6. (b) Karyotyping results. (c) Rearrangement profile: molecular pattern of rearrangement, revealed by paired-end sequencing. The illustration consists of a normal reference genome in the upper region with found paired-end reads (for example, m1 and m1a) aligned to the genome. As each read (for example, m1) is supposed to be orientated to its mate (for example, m1a), the type of rearrangement can be deduced; for example, upper panel: m4a and m4 orientated towards each other suggest a deletion, m1 and m1a on different chromosomes suggest a translocation; lower panel: m2 and m2a are located on the same strand and are orientated in the same direction, indicating an inversion. (d) Upper panel: schematic overview of the complex translocations between chromosome 10, 11 and 22, harboring the MLL/MLLT10 fusion gene. Furthermore, the fusion gene RNF214/MLL is depicted. The fusion sequence MLL/MLLT10 and t(22;11) was further validated by capillary sequencing, revealing the breakpoint at one-bp level. Lower panel: schematic overview of the translocation harboring the MLL/MLLT10 fusion gene and MLLT10/RNF169 fusion gene. The latter fusion sequence was further validated by capillary sequencing, revealing the breakpoint at 1-bp level. (e) CIRCOS plot of the initial sample: genomic landscape of interchromosomal translocations were scattered across the whole genome and allocated along the outer ring (chromosome ideograms). The inner ring represents copy number status in terms of gains and losses. However, only variants supported by 42 reads (in patient 1–3) and 43 reads (in patient 4–6) were transferred into this plot (f) CNV plot: deep blue-colored bars indicate the copy number determined by sequencing data in relation to the reference genome. In case of gains and losses bars are elevated or lowered.

Leukemia (2013) 1745 – 1791 & 2013 Macmillan Publishers Limited Letters to the Editor 1747 respectively. Furthermore, we detected numerous structural reads, we could exactly deduce the molecular anatomy of the variants in leukemic samples not found in the remission material rearrangement. In patients 3, 4 and 6, we found an inversion (Table 1). Briefly, we see classical t(10;11) rearrangements in followed by an insertion, in patient 1, 2 and 5 an inversion patients 3 and 5, which were also found by cytogenetics and FISH. followed by a translocation. Van Limbergen7 classified these In patient 3 (Figure S4), reads were found, suggestive of a missing reported molecular rearrangements as type I (patients 3, 4 and 6) region on chromosome 11 containing part of the MLL gene. and type II (patients 1, 2 and 5). In most patients we determined Further reads suggested that this MLL gene fragment was the breakpoints down to the nucleotide level by validating the subsequently inserted into the MLLT10 gene. In patient 5 breakpoint by PCR. We identified additional chromosomal fusions (Figure S5), reads indicated an inversion of a region of that were not being detected by routine cytogenetics; in patient 6 chromosome 11 that encodes partially for the MLL gene. we found a new subtype of rearrangement, which has not been Subsequently the inverted fragment and the q-terminal part was described before. Though FISH probes and cytogenetic banding translocated into the MLLT10 gene. Validation PCR and capillary techniques clearly identify the t(10;11) MLLT10/MLL translocation, sequencing verified the MLL/MLLT10 breakpoint in both patients. paired-end sequencing allowed to identify additional structural In patient 1 and 4, in contrast to FISH and cytogenetics, paired-end variants. It is certainly of great interest to know how many of the analysis revealed the involvement of additional chromosomal other previously reported t(10;11) harbor this type of rearrange- material inserted at the t(10;11) breakpoint. In patient 1 (Figure ment. In two patients we observed the reciprocal fusion of ring S1), we saw an additional breakpoint on 16q23.3. In patient 4 finger protein genes to the 30 MLL gene. RNF214 consists of a (Figure 1, upper panel), reads indicated the involvement of three conserved cysteine-rich domain that is able to bind zinc atoms. chromosomes (10, 11 and 22). We found read pairs on 11q23.3 Proteins with this domain are mostly involved in the ubiquiti- (m4, m4a) orientated towards each other but encompassing a 1- nation pathway of protein degradation. In recent publications, Mb-spanning region, suggestive of a ‘deletion’. Furthermore, at RNF169 was identified as a paralog to RNF168.8,9 RNF8/RNF168 both ends of this region we found reads, which had their origin in signaling is involved in the repair pathway of double stranded the intronic region 5–6 of the MLLT10 gene (p12.31), leading to the breaks; RNF169 is supposed to regulate this signaling pathway in MLL/MLLT10 fusion gene (m1, m1a), and at 22q12.3 (m3, m3a), magnitude.9 However, the precise function of RNF214 and RNF168, respectively. In conclusion, a 1-Mb-spanning region on 11q23.3 and their possible effects as reciprocal MLL fusion protein remains was inserted on the derivative chromosome 10 between the to be determined. Finally, copy number profiling was able to MLLT10 region (p12.31) and the inverted 22q12.3. Reads m2 and comprehensively disclose chromosomal gains and losses; for m2a suggest that the telomeric part of the MLLT10 gene had been example, in patient 6 the origin of a marker chromosome seen fused to 22q12.3 reciprocally. cytogenetically was clarified. In patient 6 (Figure 1, lower panel), we detected a new pattern In conclusion, we were able to decipher the complex MLL/ of rearrangement. We detected reads, which were on the same MLLT10 gene configuration found in our patients. Numerous strand and orientated in the same direction, instead of pointing reports have dealt with MLL/MLLT10 patients, also revealing towards each other. Therefore, m3 and m3a indicate an inversion complex karyotypes after different serial diagnostic methods. Our between 11q13.4 and 11q24.2 as the first step of rearrangement. method offers a time-efficient one-way method to discover the However, further paired-end reads suggest that only one part of underlying alteration at a high resolution on base pair level. This the inverted region on chromosome 11 was inserted into indicates that at least in patients with unclear FISH or cytogenetic chromosome 10. M2 and m2a suggested a ‘deleted region’ on findings, this method should be strongly considered, as sequen- the now inverted region between q13.4 and q23.3. M1 and m4a cing costs have also been decreasing gradually. Analytic pipelines on chromosome 11 and the corresponding mates m1a and m4 in for clinical diagnostic approaches will also need to be optimized the MLLT10 gene indicated that this particular missing region for detection of different types of variants and modified to the (q13.4:q23.3) on the previously inverted part of 11q was inserted clinical condition. into MLLT10 (intronic region 9–10). Thus, on the derivative As highlighted in our study, the involvement of ring-type Zinc chromosome 11 the region of q23.4 to q24.2 remains in the finger genes (examples also found in promyelocytic leukemia) is of inverted form. On chromosome 10, MLL exons 1–8 were fused particular interest as therapeutic strategies might evolve from in-frame to MLLT10 exons 9–24. future shared findings in other patients. Interestingly, we observed in two patients additional rearrange- ments between a ring finger protein gene and MLL, beside the CONFLICT OF INTEREST already described MLL–MLLT10 fusion. In patient 4, the rearrange- ment leads to an in-frame gene fusion RNF214/MLL (50-exon1- The authors declare no conflict of interest. 5RNF214-exon9-36MLL-30). In silico analysis predicts a 315 kDa protein (2883aa). In patient 6, we observe an in-frame gene fusion ACKNOWLEDGEMENTS 0 0 MLLT10/RNF169 (5 -exon1-8MLLT10-exon3-6RNF169-3 ; in silico We wish to thank the colleagues in the pediatric oncology centers and all parents prediction: 87 kDa protein, 781 aa). See Figure S6 for orientation who gave their consent to use the biological material from the minors. AB was and location of the corresponding genes. supported by grants from the BMBF. JB was funded by a grant of the Parents’ Copy number profiling revealed copy number variation in patient Leukemia Foundation Gieen. 2 (gains on chromosomes 1,14 and 21, losses on chromosome 2). In patient 6, cytogenetics detected an additional marker chromo- S Ghosh1, C Bartenhagen2, V Okpanyi1, M Gombert1, some and a loss of chromosome 17 (46,XY,der(10)t(10;11) V Binder1, A Teigler-Schlegel3, J Bradtke3,SRo¨ttgers3, (p12;q23)inv(11)(q13q23), der(11)t(10;11)(p12;q13),-17, þ mar(9)/45, M Dugas2 and A Borkhardt1 idem,-Y(2)/46,XY(3)). Copy number variation data identified a large 1Department of Pediatric Oncology, Hematology and Clinical gain on chromosome 8, but no loss of chromosome 17. Therefore, it Immunology, Medical Faculty, Center of Child and Adolescent Health, might be suggested that the marker consists of chromosome 17 Heinrich Heine University, Du¨sseldorf, Germany; and long arm material of chromosome 8 (Figure 1a and d). 2Institute of Medical Informatics, University of Mu¨nster, For a more detailed description of sequencing data and results, Mu¨nster, Germany and we refer to our supplementary data. In summary, in each sample 3Department of Pediatric Hematology and Oncology, we were able to reveal the complex mechanism leading to a Oncogenetic Laboratory, Justus Liebig University, translocation t(10;11) with at least three subsequent breakpoints Giessen, Germany and rearrangements. Owing to manual inspection of sequencing E-mail: [email protected]

& 2013 Macmillan Publishers Limited Leukemia (2013) 1745 – 1791 Letters to the Editor 1748 REFERENCES fusion of AF10 and MLL is resolved by fluorescent in situ hybridization analysis. 1 Chaplin T, Ayton P, Bernard OA, Saha V, Della Valle V, Hillion J et al. A novel class of Cancer Res 1995; 55: 4220–4224. zinc finger/leucine zipper genes identified from the molecular cloning of the 6 DiMartino JF, Ayton PM, Chen EH, Naftzger CC, Young BD, Cleary ML. The AF10 t(10;11) translocation in acute leukemia. Blood 1995; 85: 1435–1441. leucine zipper is required for leukemic transformation of myeloid progenitors by 2 Chaplin T, Bernard O, Beverloo HB, Saha V, Hagemeijer A, Berger R et al. MLL-AF10. Blood 2002; 99: 3780–3785. The t(10;11) translocation in acute myeloid leukemia (M5) consistently fuses the 7 Van Limbergen H, Poppe B, Janssens A, De Bock R, De Paepe A, Noens L et al. leucine zipper motif of AF10 onto the HRX gene. Blood 1995; 86: 2073–2076. Molecular cytogenetic analysis of 10;11 rearrangements in acute myeloid leukemia. 3 Meyer C, Kowarz E, Hofmann J, Renneville A, Zuna J, Trka J et al. New insights to Leukemia 2002; 16: 344–351. the MLL recombinome of acute leukemias. Leukemia 2009; 23: 1490–1499. 8 Chen J, Feng W, Jiang J, Deng Y, Huen MS. Ring finger protein RNF169 antagonises 4 Morerio C, Rapella A, Tassano E, Rosanda C, Panarello C. MLL-MLLT10 the ubiquitin-dependent signaling cascade at sites of dna damage. J Biol Chem fusion gene in pediatric acute megakaryoblastic leukemia. Leuk Res 2005; 29: 2012; 287: 27715–27722. 1223–1226. 9 Poulsen M, Lukas C, Lukas J, Bekker-Jensen S, Mailand N. Human RNF169 is a 5 Beverloo HB, Le Coniat M, Wijsman J, Lillington DM, Bernard O, de Klein A et al. negative regulator of the ubiquitin-dependent response to DNA double-strand Breakpoint heterogeneity in t(10;11) translocation in AML-M4/M5 resulting in breaks. J Cell Biol 2012; 197: 189–199.

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

Cell-surface expression of the TLR homolog CD180 in circulating cells from splenic and nodal marginal zone lymphomas

Leukemia (2013) 27, 1748–1750; doi:10.1038/leu.2013.3 family, with significant extracellular homology to the lipopolysac- charide receptor.6 Its expression is restricted to antigen presenting cells, such as B cells, monocytes, macrophages and dendritic cells, and anti-CD180 antibodies activate the majority of human and Marginal zone B-cell lymphomas (MZL) are classically characterized by mouse B cells in vitro.7,8 The literature describes opposite an indolent proliferation of CD5 À ,CD23À ,CD43À , CCND1 À and functions for CD180; however, it has been shown that CD180 CD10 À B cells that strongly express CD19, CD20, CD22, IgM and IgD, ligation induces significant B-cell proliferation and Ig production, even if CD5 þ cases have been characterized.1 The World Health in both humans and mice.7,8 Organization (WHO)2 classifies marginal B-cell lymphomas into splenic CD180 expression in B-CLL has been previously described to be marginal zone lymphoma (SMZL), nodal MZL and extranodal MZL of different from that in normal CD19 þ blood cells.9 In this paper, mucosa-associated lymphoid tissue (MALT).2 Among the lymphomas we evaluated CD180 expression using flow cytometry in two developed from the spleen, the splenic diffuse red pulp lymphoma independent series of B-LPD patients from two French with villous lymphocytes (SDRPLs) has been described as provisional hematology departments, including 60 cases of MZL that were entity close to but distinct from SMZL.2,3 centrally reviewed. For small to intermediate-sized circulating abnormal lympho- Ninety-six blood samples from patients referred for investiga- cytes, MZL diagnosis can be of great difficulty, particularly because tion or follow-up of B-LPD in the Oncology and Hematology there is no immunological-positive marker to distinguish MZL Department of the Strasbourg University Hospital were tested: 30 from atypical chronic lymphocytic leukemia (CLL), mantle cell patients with CLL, 16 patients with MCL, 15 patients with LPL and lymphoma (MCL) and lymphoplasmacytic lymphoma (LPL). More- 35 patients with MZL (25 splenic, 10 nodal without splenomegaly), over, if translocations involving MALT1 (t(11;18), t(14;18)), BCL10 in addition to 25 controls. Blood cells samples were collected after (t(1;14)) or FOXP1 (t(3;14)) are primarily associated with MALT, and signed informed consent was obtained (Comite´ de protection des del7q is frequent in SMZL, these cytogenetic markers are not personnes EST-IV, # AC-2008-438) in accordance with the always present. Therefore, there is no specific immunologic Declaration of Helsinki, and the Ethic Committee of the Faculty marker for splenic or nodal MZL. of Medicine of Strasbourg approved the study in 2011. Over 93% To improve the diagnosis of chronic B-cell lymphoproliferative of the patients were included in the study at the time of diagnosis, disease (B-LPD) with circulating cells, we have previously and all patients presented a low count of circulating residual developed an innovative mass spectrometry-based proteomic normal B cells. All cases of MCL displayed a t(11;14) and/or cyclin workflow on plasma membrane microparticles derived from D1 hyperexpression. All cases of CLL had a Matutes score of 4 or 5. B-LPD4 and proposed CD148 as a potential marker for MCL Only cases of MZL without plasma differentiation were selected in diagnosis panel.5 Using again the proteomic toolbox, a differential this study. Nine patients with SMZL displayed circulating villous proteomic analysis was applied here to compare the membrane lymphocytes. proteomes originating from MPs derived from the following three As previously described, normal B cells displayed strong different B-LPD: LPL, MCL and MZL. staining for CD180. CD27 þ and CD27 À cells were equally A list of 388 unique membrane proteins was obtained including stained and only a small percentage of CD180 À B cells were 59 CD annotated cell-surface proteins (Supplementary Table). observed in the controls (mean 5.3%, range 0.4–14%, data not Within this list of characterized cell-surface proteins, we focused shown). In agreement with previous studies9 malignant B cells of on CD180, also called Radio-Protective 105 (RP105) or lymphocyte CLL displayed weaker expression of CD180 (Po0.0001 Mann– antigen 64, which was considered overexpressed in the MZL Whitney U-test) (Figure 1b). sample by our semi-quantitative mass spectrometry analysis We observed that CD180 staining was also weak in MCL (peptide and spectral counting). CD180 is a leucine-rich repeat (Po0.001) and LPL (Po0.001) B cells (Figures 1a and b). CD180 type 1 membrane protein belonging to the Toll-like receptor (TLR) mean fluorescence intensities (MFIs) of CLL, MCL and LPL B cells

Accepted article preview online 10 January 2013; advance online publication, 1 February 2013

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