Combined interphase fluorescence in situ hybridization elucidates the genetic heterogeneity of T-cell acute lymphoblastic leukemia in adults

Abstract

Background Molecular lesions in T-cell acute lymphoblastic leukemias affect regulators of cell cycle, proliferation, differentiation, survival and apoptosis in multi-step pathogenic pathways. Full genetic characterization is needed to identify events concurring in the development of these leukemias.Design and Methods We designed a combined interphase fluorescence in situ hybridization strategy to study 25 oncogenes/tumor suppressor genes in T-cell acute lymphoblastic leukemias and applied it in 23 adult patients for whom immunophenotyping, karyotyping, molecular studies, and gene expression profiling data were available. The results were confirmed and integrated with those of multiplex-polymerase chain reaction analysis and gene expression profiling in another 129 adults with T-cell acute lymphoblastic leukemias.Results The combined hybridization was abnormal in 21/23 patients (91%), and revealed multiple genomic changes in 13 (56%). It found abnormalities known to be associated with T-cell acute lymphoblastic leukemias, i.e. CDKN2A-B/9p21 and GRIK2/6q16 deletions, TCR and TLX3 rearrangements, SIL-TAL1, CALM-AF10, MLL-translocations, del(17)(q12)/NF1 and other cryptic genomic imbalances, i.e. 9q34, 11p, 12p, and 17q11 duplication, del(5)(q35), del(7)(q34), del(9)(q34), del(12)(p13), and del(14)(q11). It revealed new cytogenetic mechanisms for TCRB-driven oncogene activation and C-MYB duplication. In two cases with cryptic del(9)(q34), fluorescence in situ hybridization and reverse transcriptase polymerase chain reaction detected the TAF_INUP214 fusion and gene expression profiling identified a signature characterized by HOXA and NUP214 upregulation and TAF_I, FNBP1, C9orf78, and USP20 down-regulation. Multiplex-polymerase chain reaction analysis and gene expression profiling of 129 further cases found five additional cases of TAF_I-NUP214-positive T-cell acute lymphoblastic leukemia.Conclusions Our combined interphase fluorescence in situ hybridization strategy greatly improved the detection of genetic abnormalities in adult T-cell acute lymphoblastic leukemias. It identified new tumor suppressor genes/oncogenes involved in leukemogenesis and highlighted concurrent involvement of genes. The estimated incidence of TAF_I-NUP214, a new recurrent fusion in adult T-cell acute lymphoblastic leukemias, was 4.6% (7/152).

Introduction

Full genomic characterization is needed to trace pathogenic pathways in individuals with T-cell acute lymphoblastic leukemia (T-ALL), establish the impact of primary and secondary changes and determine how diverse genetic abnormalities interact/co-operate as T-ALL derives from combinations of diverse molecular lesions (multi-step hits). These lesions affect genes implicated in cell proliferation and/or survival (LCK and ABL1), self-renewal (NOTCH1), cell differentiation (HOX genes, MLL, LYL1, TAL1/2 and LMO1/2), and cell cycle control (CDKN2A/CDKN2B).1,2 The main causes of gene deregulation are: (i) oncogene activation with ensuing ectopic or over-expression, which is mainly due to juxtaposition with T-cell receptor loci (TCRB-HOXA, TCRA/D-HOX11, TCRA/D-LMO2, TCRA/D-LMO1);3,4 (ii) gain of function mutations (NOTCH1 and JAK1);5,6 (iii) tumor suppressor gene haploinsufficiency or inactivation, which is usually the result of deletion (CDKN2A-B) and/or loss of function mutation (PTEN);1,7 and (iv) chromosomal translocations producing fusion proteins which are associated with specific subgroups of T-ALL (CALM-AF10, NUP98-, MLL- and ABL1-fusions).1,2

New genomic techniques such as interphase fluorescence in situ hybridization (FISH), and array-comparative genomic hybridization have greatly improved the detection of genomic abnormalities in T-ALL.8 They bypass the inherent difficulties of poor normal cell proliferation in T-ALL and failure of conventional cytogenetics to identify cryptic molecular events such as the NUP214-ABL1 fusion, t(9;14)(q34;q32)/EML1-ABL1, C-MYB duplication or translocation, and cryptic del(17q)/NF1.9–13 In T-ALL gene expression profiling, assessing over- and under-expression of a myriad of genes, has elucidated distinct signatures that are associated with over-expression of LYL1, HOX11, TAL1, LMO1, LMO2, HOXA and HOX11L2 oncogenes.14,15

In the present study, we developed a newly designed strategy, i.e. combined interphase FISH, to increase our knowledge on the genetics of each individual case.

Design and Methods

Patients

This study recruited 152 adult patients (111 males, 41 females; median age 30; range, 17–64 years) with T-ALL who were enrolled in two consecutive Italian multi-center GIMEMA (Gruppo Italiano Malattie Ematologiche dell’Adulto) studies (LAL-0496 and LAL-0904) except for one 18-year old male patient who was enrolled in the AIEOP LLA 2000 protocol (Online Supplementary Methods). Combined interphase FISH was done on 23 cases for whom results of mutational analysis (NOTCH1 and FBW7) (Online Supplementary Methods) and multiplex-polymerase chain reaction (PCR) analysis (BCR-ABL1, PBX-E2A, SIL-TAL1, MLL-AF4, MLL-ENL, NUP98-RAP1GDS1) were available.16 The other 129 patients underwent a newly designed multiplex-PCR analysis or gene expression profiling.

Combined interphase fluorescence in situ hybridization

Specific DNA clones, ranging in size from 59 to 215 Kb, for genes/loci that have been implicated to date in T-ALL were selected (Table 1).3,17–19 Clones were labeled with spectrum orange and green (Vysis, IL, USA) for double-color break-apart tests or combined split FISH tests and applied on bone marrow and/or peripheral blood samples (Online Supplementary Methods).

The SIL-TAL1 fusion was investigated using a SIL-TAL1 FISH DNA probe (Dako Italia, Milan, Italy). In cases with 9q34/ABL1 deletion FISH was done with the LSI BCR-ABL1 ES dual color probe (Vysis, IL, USA) and four clones for the 9q34 band, centromere to telomere as follow: RP11-216B9, RP11-550J21, RP11-143H20, RP11-544A12. Abnormal hybridization patterns were: (i) a split signal (one fusion signal and separate green and red signals); (ii) duplication/trisomy (three fusion signals); (iii) deletion/monosomy (one fusion signal); and (iv) partial deletion (one fusion signal and one orange or one green signal). Cut-offs were the upper limit of 500 normal peripheral blood nuclei for each assay: split, duplication/trisomy patterns were considered positive when found in more than 5% of interphase cells; monosomy/deletion/partial deletion patterns were considered positive when found in more than 10% of cells.

Comparative genomic hybridization

Comparative genomic hybridization was performed in patient n. 6 to elucidate the chromosome 6q rearrangement and in patients n. 14 and 17 to investigate TCRB involvement in putative unbalanced translocations (Online Supplementary Methods).

Molecular studies

Reverse transcriptase PCR was performed as described elsewhere.20 The following primers were used to investigated TAF_I and NUP214: TAF_540F (exon 6) (5′-GAAGAGGCAGCATGAGGAAC-3′) + NUP_2916R (exon 20) (5′-TACTTTGGGCAAGGATTTGG-3′) for the first amplification round and TAF_747F (exon 7) (5′-TGACGAAGAAGGGGATGAGGAT-3′) + NUP_2601R (exon 18) (5′-ATCATTCACATCTTGGACAGCA-3′) for the nested PCR. Nested-PCR was used to monitor minimal residual disease in two patients for whom material was available. The TAF_540F/NUP_2916R and TAF_747F/NUP_2601R primers were also added to our multiplex-PCR for diagnostic screening of a cohort of 96 cases.16 Isoform-specific PCR was done using TAFa_283F (5′-GAAACCAAGACCACCTCCTG-3′) and TAFb_38F (5′-AGCTCAACTCCAACCACGAC-3′) primers (Online Supplementary Methods). A new reverse transcriptase PCR for SIL-TAL1 was set up (Online Supplementary Methods). Real-time quantitative PCR was performed to corroborate gene expression profile findings (Online Supplementary Methods).

Gene expression profiling and statistical analysis

We used the oligonucleotide arrays HGU133 Plus 2.0 gene chips, Affymetrix. Unsupervised clustering was performed as previously described21,22 and the distance between two genes was computed as one minus the correlation between standardized expression values across samples. Supervised analyses included analysis of variance (ANOVA) and t tests. ANOVA (p value <0.01) was used to compare the following subgroups: (i) five samples with normal or heterogeneous combined interphase FISH findings (patients n. 1, 5, 9, 15 and 16 of Table 2); (ii) two samples with the HOX11 rearrangement; (iii) four samples, defined as “HOXA”-positive because each had one of the following fusions inducing HOXA gene over-expression: CALM-AF10, MLL-ENL, MLL-translocation with an unknown partner and TCRB-HOXA;15,23–25 (iv) three SIL-TAL1-positive samples; and (v) four TAF_I-NUP214-positive samples. Patients n. 2, 12, 17, 18, 19 and 21 were excluded from supervised analysis, since each had a different aberration. t-tests (p value <0.01, fold change difference >1.5) were performed between the TAF_I-NUP214 and “HOXA” subgroups and TAF_I-NUP214-positive patients and the other T-ALL patients.

Results

Patients

The clinical and hematologic features of the 23 patients investigated in the preliminary study are summarized in Online Supplementary Table S1. Tables 2 and 3 show these patients’ cytogenetic and molecular findings. Eight (34.8%) had a normal karyotype, six (26.1%) had an abnormal karyotype and karyotype analysis was unsuccessful in the remaining nine cases (39.1%). Multiplex PCR detected SILTAL1 in one patients and MLL-ENL in another. NOTCH1 and FBW7 mutations were found in 11 and four cases, respectively (Table 2).

Table 1.Combined interphase-FISH assay genes and genomic clones.

Combined interphase fluorescence in situ hybridization

Combined interphase FISH detected diverse, multiple genomic aberrations, finding the aberrant clone in 10–98% of analyzed cells. The hybridization was abnormal in 21/23 patients, with 13 of these patients having two or more genetic changes (Tables 2 and 3). Abnormalities included CDKN2A-B deletions, del(6)(q16)/GRIK2, TCRB, TCRA/D TLX3 (Figure 1A), and MLL translocations, and CALM-AF10.26 Combined interphase FISH detected a cryptic del(1)(p32)/SIL-TAL1 in three cases and PCR confirmed the fusion in all, using only our new specific primers (Online Supplementary Results and Online Supplementary Figure S1). A del(17)(q12)/NF1 was detected in patients # 9 and 19.

Combined interphase FISH also unraveled new underlying cytogenetic mechanisms. In patient # 6, three separate hybridization signals were obtained from clone RP11-32B1 encompassing C-MYB (Figure 1B) and flanking clones RP11-557H15 and RP11-448D5. This finding, which is indicative of partial 6q trisomy and was validated by comparative genomic hybridization, may underlie C-MYB over-expression. In cases # 14 and 17 TCRB clones showed deletion of the 3′TCRB flanking probe (Figure 1C), indicating unbalanced translocations. In patient # 14, comparative genomic hybridization detected loss of 7q33-qter and gain of the 10q24-qter suggesting that a trisomic 10q was implicated in an unbalanced translocation with the TCRB gene. A double-color experiment combining RP11-1220K2, flanking 5′TCRB, and RP11-107I14, flanking 3′HOX11, demonstrated that TCRB and HOX11 were juxtaposed (Figure 1D). In patient # 17, comparative genomic hybridization revealed that the entire 9q arm was trisomic, but lack of biological material precluded further studies.

Combined interphase FISH detected nine cryptic imbalances: dup(17)(q11.2)/NF1, dup(9)(q34)/ABL1-NOTCH1, dup(11p), dup(12)(p13)/ETV6, del(5)(q35)/TLX3, del(12) (p13)/ETV6, del(7)(q34)/TCRB, del(14)(q11)/TCRA/D, and del(9)(q34)/ABL1. In two cases with ABL1/9q34 deletion, the LSI BCR-ABL ES dual-color probe proved that the deletion extended centromeric to ABL1 (Figure 1E). Four additional 9q34 probes defined its endpoints centromerically, between RP11-216B9 and RP11-550J21, within the TAF_I gene and telomerically, between RP11-143H20 and RP11-554A12 (Figure 1F) within the NUP214 gene. These findings suggest that the 5′TAF_I and the 3′NUP214 were juxtaposed.

TAF_I-NUP214 fusion

Reverse transcriptase PCR gave an 802 bp product in case # 3 and a 643 bp in case # 4; the amplification products detected by nested PCR were 280 bp and 121 bp, respectively. Cloning experiments and sequence analysis showed that nucleotide 813 (exon 7) of TAF_I was fused to nucleotide 2389 (exon 17) of NUP214 in patient # 3 and nucleotide 813 (exon 7) of TAF_I was fused to nucleotide 2548 (exon 18) of NUP214 in patient # 4 (Figure 2A). These two samples were used as positive controls when screening 96 additional T-ALL patients. Multiplex reverse transcriptase PCR found the TAF_I-NUP214 fusion transcript in three out of these 96 patients (Figure 2B). In each of the five patients with TAF_I-NUP214 fusion, isoform-specific reverse transcriptase PCR detected TAF_Iα-NUP214 and TAF_Iβ-NUP214 fusion transcripts (Figure 2C). Molecular cloning in patients # 3, 4, X, and Y showed both isoforms had the same TAF_I-NUP214 fusion point. The TAF_I-NUP214 fusion was found during minimal residual disease monitoring in both patients for whom material was available (Online Supplementary Table S2).

Table 2.Conventional cytogenetics, molecular and combined interphase-FISH findings in 23 patients with T-ALL.

Gene expression profiling showed over-expressed HOXA genes in patients with T-cell lymphoblastic leukemia with the TAF_I-NUP214 fusion

Gene expression profiles were evaluable in 24 samples (22/23 samples from patients undergoing combined inter-phase FISH and 2/3 samples from patients found to be TAF_I-NUP214-positive by multiplex-PCR). Unsupervised analysis selected 3557 probe sets, corresponding to 2795 genes, but did not reveal any clear subgroups. ANOVA selected 406 probe sets, corresponding to 372 genes, and recognized five specific clusters, i.e. samples with heterogeneous combined interphase FISH findings, HOX11, “HOXA”, TAL1 and TAF_I-NUP214 (Figure 3A). The TAF_I-NUP214 cluster was characterized by over-expression of HOXA3, HOXA5, HOXA7, HOXA9 and HOXA10 (P value <0.005) suggesting a similarity with the “HOXA” cluster, despite less up-regulation (Figure 3B). t-testing highlighted TAF_I down-modulation and higher NUP214 expression in TAF_I-NUP214-positive patients than in “HOXA” patients (Figure 3C). FNBP1, C9orf78 and USP20 at the 9q34 deleted region were significantly (P<0.01) down-regulated in the four TAF_I-NUP214+ samples (Figure 3C).

Quantitative PCR detected a significant (P<0.05) down-modulation of TAF_I transcripts and a significant over-expression (P<0.001) of the NUP214 3′ region in TAF_I-NUP214-positive patients compared with all other clusters (data not shown). Gene expression profiling analysis, validated in 33 additional cases of T-ALL (data not shown), detected two cases with a TAF_I-NUP214 signature confirmed by FISH (Online Supplementary Table S2) and by reverse transcriptase PCR (data not shown).

Discussion

Remarkable progress was made in the understanding of the genetics of T-ALL when conventional cytogenetics were combined with new technologies such as FISH, reverse transcriptase PCR, and microarray analysis. In reviewing reports we found that although at least 25 recurrent genes/loci had been identified as playing roles in T-ALL leukemogenesis, few data were available on concomitant molecular events in each individual patient. The results presented here show how we successfully set up a combined interphase FISH assay that was suitable for in-depth characterization of TALL in clinical and research laboratories. It detected cryptic aberrations and elucidated preferential and forbidden associations of multiple genetic events involved in the pathogenesis of each case. Combining at least six molecular assays in one slide, it provided major savings in time, costs and biological material.

This open, combined interphase FISH showed an abnormal hybridization pattern in 91% of T-ALL patients, 56% of whom had multiple rearrangements even though karyotyping and multiplex-PCR had shown genetic defects in only 26% and 8.7% of patients, respectively. It reliably detected typical, recurrent abnormalities such as 9p21/CDKN2A and 6q16/GRIK2 deletions, SIL-TAL1, CALM-AF10, and TCR-rearrangements and found, for the first time in two adult patients, the monoallelic del(17)(q12)/NF1, which had already been reported in three children with T-ALL.13 This finding confirms that NF1, a tumor suppressor gene, plays a role in the onset/development of diverse hematologic malignancies. Molecular lesions such as the NOTCH1 mutation, but not the characteristic del(9)(p21)/CDKN2A/B or del(6)(q16)/GRIK2 were also observed in these two patients.

In association with comparative genomic hybridization, combined interphase FISH revealed new cytogenetic mechanisms underlying typical T-ALL molecular lesions. In one case MYB duplication, due to a partial 6q22q25 trisomy, was associated with a TCRB-LMO2 rearrangement and del(6q)/GRIK2, confirming previous reports of MYB duplication being present together with other genetic lesions.12 A hitherto unknown deletion of the TCRB 3′ flanking region, which corresponded to an unbalanced der(7)t(7;10) (q34;q24) translocation, produced the TCRB-HOX11 rearrangement.

Table 3.Summary of the 57 aberrations detected by CI-FISH in 21/23 adult T-ALL patients.

Focusing on del(9)(q34)/ABL1, one of the four recurrent cryptic chromosome imbalances detected by combined interphase FISH (Tables 2 and 3), we found del(9)(q34) produced TAF_I-NUP214 as previously described in acute myeloid leukemia and in 5.4% of cases of pediatric TALL.20,27 TAF_I (official name SET), encodes for TAF_Iα and TAF_Iβ isoforms which are localized to the nucleus and have different N-terminal sequences deriving from an alternative first exon. They are chromatin remodeling proteins involved in DNA replication and transcription.28,29 NUP214, an FG nucleoporin interacting with hCRM1 in nucleus-cytoplasmic traffic, is localized at the cytoplasmic side of the nuclear pore complex. It also rearranges with ABL1 in T-ALL bearing episomal or intrachromosomal amplification.9 Remarkably, NUP214 like NUP98, another nucleoporin with a putative nucleus-cytoplasmic shuttling function, is involved in the pathogenesis of T-ALL and myeloid malignancies.

As observed in children, TAF_I-NUP214-positive T-ALL adults had a specific gene expression signature with HOXA gene cluster over-expression which was not as marked as in the so-called “HOXA” cluster T-ALL, i.e. patients with CALM-AF10, MLL-translocation and HOXA rearrangements (Figure 3). The gene expression signature was also characterized by NUP214 up-regulation and TAF_I down-regulation. Furthermore, down-regulation of FNBP1, C9orf78, and USP20, mapping within the cryptic del(9)(q34), suggested haploinsufficiency of these genes.

Figure 1.Representative FISH detection of A) TLX3 unbalanced translocation in patient n. 21: the TLX3 assay gives one fusion and one orange signal. (B) Trisomy of 6q including C-MYB in patient n. 6: clone RP1-32B1 spanning C-MYB gives three orange signals. The green signal corresponds to RP1-258B3 for GRIK2 which, in this case, underwent monoallelic deletion. (C and D) Unbalanced der(7)t(7;10)/TCRB-HOX11 in patient n. 14: the TCRB assay gives 1 fusion and 1 orange signal (C) and the double-color assay combining RP11-1220K2 (green) and RP11-107I14 (orange) results in one fusion, two orange and one green signal (D). Panels E and F) Cryptic del(9)(q34) in patient n. 3: metaphase FISH with the LSI BCR-ABL1 produces two green and one orange signals (E); a double-color experiment with RP11-544A12 orange and RP11-143H20 green shows one fusion signal on normal 9 and an orange signal on del(9)(q34) (F).

Figure 2.Characterizing TAF_I-NUP214 fusion transcripts. Panel A) TAF_I-NUP214 fusion transcript cloning and sequencing in patients 3 (upper schema) and 4 (lower schema). Arrows indicate the primers used in PCR amplification. Panel B) Multiplex PCR screening study used specific primers for E2A control gene amplification (680 bp). RA multiplex PCR: Upper gel includes oligonucleotides listed in right boxes. TAF_I-NUP214 fusion transcripts were confirmed in patients 3 (280 bp) and 4 (121 bp) and detected in another three patients, i.e. X (121 bp) and Y (121 bp) and W (121bp). RB multiplex PCR: Lower gel includes oligonucleotides listed in right boxes. Only the control gene was amplified. Panel C) TAF_I nucleotide 353 (exon Iα) is fused in-frame to TAF_I nucleotide 77 (exon 2) in the TAF_Iα/NUP214 isoform. TAF_I nucleotide 76 (exon Iβ) is fused in-frame to TAF_I nucleotide 77 (exon 2) in the TAF_Iβ-NUP214 isoform. Sequence numbers refer to GenBank accession NM_003011.1 for TAF_I and NM_005085.2 for NUP214. The TAF_I exon I_ sequence is reported as NM_001122821.1.

In order to establish the incidence of TAF_I-NUP214 in adult T-ALL we screened 129 additional patients, finding five positive cases: 3/96 screened by multiplex PCR and 2/33 screened by gene expression profiling. Thus, 7/152 adults with T-ALL carried TAF_I-NUP214, giving an estimated incidence of 4.6%, which is the same as in children.27 This subgroup of adult T-ALL has an immature phenotype and one or more additional genomic abnormalities: NOTCH1 mutations in four cases, del(12)(p13)/ETV6 in two cases, del(11)(p13)/LMO2 plus del(11)(q14)/CALM in two cases, del(5)(q35)/TLX3, del(6)(q16)/GRIK2, and del(9)(p21)/CDKN2A-B in one case each. Therefore, from a pathogenic point of view, in T-ALL, TAF_I-NUP214 cooperates with various different concomitant molecular events, all of which concur to determine a poor response to induction therapy. In fact, four of the six patients for whom follow-up information was available died within 12–24 months of diagnosis due to refractory disease or relapse; one patient, who has been followed up for 3 months at the time of writing, had TAF_I-NUP214 detectable by reverse transcriptase PCR after induction therapy, indicating that molecular remission had not been achieved and that the patient’s prognosis was presumably poor. Indeed the only one of these six patients with sustained complete remission for 29 months had undergone allogeneic bone marrow transplantation (Online Supplementary Table S2). Interestingly, six of the seven cases studied for TAF_I isoforms, carried both the TAF_Iα-NUP214 and TAF_Iβ-NUP214 fusion transcripts. Both fusion transcripts retained the TAF_I amino terminal region which is essential for protein dimerization and, consequently, chromatin remodeling activity. Further studies will help to clarify the roles and interrelationships of the oncogenic TAF_I-NUP214 isoforms.

Figure 3.Gene expression profiling A) GEP of 24 adult T-ALL samples. Samples bearing TAF_I-NUP214 are characterized by a specific signature. Each row represents a probe set, each column represents a single sample. The color scale indicates the relative levels of expression: dark blue the lowest levels of expressions, red the highest levels of expression. Red indicates high expression and blue low expression. B) Expression of HOXA cluster genes resulted from ANOVA (P<0.005). Panel C) Expression levels of genes differentially expressed between TAF_I-NUP214 and T-ALL samples.

In conclusion, combined interphase FISH is a powerful, flexible method as the assay can be extended beyond the panel of 25 genes that were tested in the present study. It provides in-depth molecular characterization in at least 90% of adults with T-ALL, and might be proposed in clinical laboratories as a surrogate for more advanced, expensive technologies. In the research setting, its integration with mutational analysis, PCR, and gene expression profiling in prospective studies of large consecutive T-ALL series will accelerate our understanding of the biology of T-ALL leukemogenic pathways and the design of a genome-based classification.

Acknowledgments

the authors would like to thank GIMEMA (Gruppo Italiano Malattie Ematologiche dell’Adulto), Dr. G A Boyd for assistance in the preparation of the manuscript and Roche Molecular Systems.

Footnotes

• PG and RLS contributed equally to this manuscript.
• Funding: this work was supported by a grant from IAP (Interuniversity Attraction Poles, University of Leuven, Belgium), AIRC (Associazione Italiana Ricerca sul Cancro), PRIN-MIUR (Programmi di Ricerca Cofinanziati-Ministero per l’Istruzione, l’Università e la Ricerca Scientifica, Italy), Fondazione Cassa di Risparmio di Perugia, Associazione “Sergio Luciani”, Fabriano, Italy; and the European Community. BC received a grant by FIRC (Fondazione Italiana Ricerca sul Cancro).
• The online version of this article has a supplementary appendix.
• Authorship and Disclosures CM designed the study, supervised all the results and wrote the paper. PG designed and supervised the molecular analysis (LE, SG, LB) and drafted the paper. RLS designed and supervised the FISH analysis (VP and BC) and drafted the paper. EV and CaM performed the mutational analysis. GLB and CaM performed the CGH analysis. SC and MM performed the GEP analysis. AV, AB, MM, MFM, AG, RF, and BC provided clinical, immunophenotypic and cytogenetic information on patients.
• The authors reported no potential conflicts of interest.

• Received April 21, 2009.
• Revision received June 9, 2009.
• Accepted July 8, 2009.