Abstract
Inv(3)(q21q26)/t(3;3)(q21;q26) is recognized as a distinctive entity of acute myeloid leukemia (AML) with recurrent genetic abnormalities of prognostic significance. It occurs in 1-2.5% of AML and is also observed in myelodysplastic syndromes and in the blastic phase of chronic myeloid leukemia. The molecular consequence of the inv(3)/t(3;3) rearrangements is the juxtaposition of the ribophorin I (RPN1) gene (located in band 3q21) with the ecotropic viral integration site 1 (EVI1) gene (located in band 3q26.2). Following conventional cytogenetics to determine the karyotype, fluorescent in situ hybridization (FISH) with a panel of bacterial artificial chromosome clones was used to map the breakpoints involved in 15 inv(3)/t(3;3). Inv(3) or t(3;3) was the sole karyotypic anomaly in 6 patients, while additional abnormalities were identified in the remaining 9 patients, including 4 with monosomy of chromosome7 (−7) or a deletion of its long arm (7q-). Breakpoints in band 3q21 were distributed in a 235 kb region centromeric to and including the RPN1 locus, while those in band 3q26.2 were scattered in a 900 kb region located on each side of and including the EVI1 locus. In contrast to most of the inversions and translocations associated with AML that lead to fusion genes, inv(3)/t(3;3) does not generate a chimeric gene, but rather induces gene overexpression. The wide dispersion of the breakpoints in bands 3q21 and 3q26 and the heterogeneity of the genomic consequences could explain why the mechanisms leading to leukemogenesis are still poorly understood. Therefore, it is important to further characterize these chromosomal abnormalities by FISH.
- Inv(3)(q21q26)
- t(3;3)(q21;q26)
- breakpoint distribution
- chromosomal abnormalities
- fluorescent in situ hybridization
The revised 2008 WHO classification of tumors of hematopoietic and lymphoid tissues recognized acute myeloid leukemia (AML) with inv(3)(q21q26)/ t(3;3)(q21;q26) as a distinctive entity of AML with recurrent genetic abnormalities of prognostic significance (1). Patients with inv(3)/t(3;3) frequently demonstrate normal or elevated platelet count, atypical megakaryocytes and multilineage dysplasia in bone marrow as well as minimal to no response to chemotherapy and poor clinical outcome (2-4). Inv(3) and t(3;3) occur in 1-2.5% of all FAB types of AML, except M3 (5, 6). They are also observed in myelodysplastic syndromes (MDS) (7) and in blastic phase of chronic myeloid leukemia (CML) (8).
The molecular consequence of the inv(3)/t(3;3) rearrangements is the juxtaposition of the RPN1 gene (ribophosphorin 1 gene located in band 3q21) with the EVI1 gene (ecotropic viral integration site-1 gene located in band 3q26.2) (9). EVI1 is a nuclear transcription factor that plays an essential role in the proliferation and maintenance of hematopoietic stem cells (10-12) and can inhibit myeloid differentiation (13). Two alternative forms exists, one generated from EVI1, the other MECOM (MDS1 and EVI1 complex locus) through intergenic splicing with MDS1 (myelodysplasia syndrome 1), a gene located 140 kb upstream of EVI1 (14). It has been suggested that the promoter of the house-keeping RPN1 could be responsible for the activation of the EVI1 gene in 3q26 (15). Indeed, the overexpression of EVI1 can be achieved not only through rearrangements of band 3q26, but also without the presence of 3q26 abnormalities, therefore indicating that other mechanisms can lead to EVI1 activation (16-22). Moreover, a substantial number of patients with 3q26 rearrangements do not overexpress EVI1 (23-25).
Furthermore, overexpression of GATA2 (GATA binding protein 2), located in band 3q21, was found in leukemia cell lines, as well as in patient samples with inv(3)/t(3;3), compatible with the hypothesis that deregulation of its transcription could contribute to leukemogenesis (26, 27). Indeed, GATA2, a member of a family of zinc finger transcription factors, is required in the early proliferative phase of hematopoietic development and is essential for proliferation of definitive hematopoietic stem cells (HSC) (28, 29).
Breakpoints in both regions are scattered over several hundred kilobases (kb). A first breakpoint cluster region (BCR) of about 30 kb located 15 kb centromeric of the RPN1 gene was identified (called telomeric BCR or BCR-T) (15, 30) but a number of breakpoints were found to map centromeric to that BCR region in another BCR located up to 60 kb from the first described BCR (called centromeric BCR or BCR-C) (5, 31-33). Breakpoints in band 3q26 are distributed on each side of the EVI1 gene, with t(3;3) and inv(3) breakpoints more likely to be located 5' and 3' of the EVI1 gene, respectively (9, 15, 16).
Over the past years, several groups have used P1-derived artificial chromosome (PAC) and bacterial artificial chromosome (BAC) clones to locate the breakpoints involved in inv(3)/t(3;3) in bands 3q21 and 3q26 (5, 33-39). A sole study tried to locate the breakpoints more precisely in 7 patients (27). In the present study, we used two BAC libraries to map the 3q21 and 3q26 breakpoints in 11 patients with inv(3) and 4 patients with t(3;3). We also compared our results to those obtained with the commercially available Cytocell Aquarius EVI1 Breakapart three-color probe (AmpliTech, Compiègne, France).
Patients and Methods
Patients. Patients diagnosed with malignant hemopathies at the Brest University Hospital (Hôpital Morvan) or at the Centre Hospitalier Le Foll in St Brieuc, a hospital located 150 km from Brest, are referred to the Cytogenetics Laboratory of the Brest University Hospital for conventional and molecular cytogenetic analyses. All the patient files from 1998 to 2010 were checked for the presence of inv(3)(q21q26)/t(3;3)(q21;q26) and included in the present study.
Conventional cytogenetics. Conventional cytogenetic analysis was performed on bone marrow cells at the time of diagnosis. Briefly, 24-hour unstimulated bone marrow culture was synchronized with fluorodesoxyuridine (10−7 M) for 17 h followed by thymidine (10−5 M) for 6 h before colcemide exposure and standard harvesting. R-Banding chromosomal analyses were performed according to standard procedures and the karyotypes described according to the International System for Cytogenetic Nomenclature (ISCN 2009) (40).
Fluorescent in situ hybridization (FISH) analyses with BAC clones. FISH studies were performed on the same fixed material as the conventional cytogenetic analyses. Cell preparations were stored in fixative at −20°C until use. To delineate the location of the breakpoints, FISH analyses were carried out using BAC clones mapping to bands q21 and q26 of chromosome 3.
We identified the BAC clones of interest through the Human Genome Browser Database of the Genome Bioinformatics Group at the University of California at Santa Cruz (http://genome.ucsc.edu/). They were then ordered from the Children's Hospital Oakland Research Institute in California (http://bacpac.chori.org/). Bacterial cultures were prepared from a single colony picked from a selective plate in the presence of chloramphenicol. Plasmids were obtained from bacterial cultures grown in the presence of chloramphenicol (10 mg/l). After lysing bacteria using 1% SDS/0.2 N NaOH, DNA was purified from RNA, proteins and other cellular contaminants. Probes were then labeled by nick translation in Spectrum Orange (Nick Translation Kit, Abbott, Rungis, France) or in fluorescein isothiocyanate (FITC; Prime-it Fluor Fluorescence Labeling Kit, Stratagene, Amsterdam, the Netherlands). All BAC clones were applied to normal lymphocyte metaphases to confirm their chromosomal location.
After hybridization, the slides were counterstained with 4-6-diamino-2-phenyl-indole-dihydrochloride. The preparations were examined using a Zeiss Axio Plan Microscope (Zeiss, Le Pecq, France). Image acquisition was performed using a CCD camera and analyzed using the In Situ Imaging System program (MetaSystems, Altlussheim, Germany).
Two libraries of 9 and 32 overlapping BACs covering bands 3q21 and 3q26 respectively were applied to the patients (Table I).
FISH analyses with commercially available probe. A FISH study using the Cytocell Aquarius EVI1 Breakapart probe (AmpliTech, Compiègne, France) was carried out on the metaphase preparations from all patients, as recommended by the manufacturer. The EVI1 Breakapart probe contains three probes: a probe labeled in Aqua of 562 kb in size centromeric to the EVI1 gene, a probe labeled in Spectrum Green of 181 kb covering EVI1 and its flanking regions and a probe labeled in Spectrum Orange of 124 kb telomeric of the EVI1 gene (telomeric of MYNN and covering LRRC34).
Results
Patients. From 1998 to 2010, 15 patients were found to have an inv(3)(q21q26) or a t(3;3)(q21;q26), (Table II). There were 10 males and 5 females. The mean age at the time of diagnosis was 57.8 years (standard deviation: 15.9 years). Nine patients had a diagnosis of AML and one of acute biphenotypic leukemia. One patient had AML evolving on chronic myelomonocytic leukemia and another was in the blastic phase of CML. Three patients had MDS, two of them developing AML within 2 years following the diagnosis of MDS.
R-Banding conventional cytogenetics showed 11 patients to have an inv(3)(q21q26) and the remaining 4 a t(3;3)(q21;q26). Inv(3) or t(3;3) was the sole karyotypic anomaly in 6 patients. Additional abnormalities were identified in 9 patients. Various anomalies led to a partial deletion of the short arm of chromosome 17 in 3 patients, while 4 other patients had monosomy 7 (−7) or a deletion of its long arm (7q-) (Table II).
Breakpoint mapping with BAC clones. Sequential FISH analyses with BAC clones were applied on metaphases of all 15 patients. Breakpoints in band 3q21 were distributed in a 235 kb region centromeric to and including the RPN1 locus (Figure 1). The breakpoint to one patient (P8) was localized in the RPN1 gene, whereas the breakpoints for 10 other patients mapped to a 72 kb region situated about 15 kb from the RPN1 locus.
Breakpoints in band 3q26.2 were distributed in a 900 kb region located on each side of and including the EVI1 locus (Figure 2). The breakpoints of all 11 inv(3) were localized in or centromeric to the EVI1 gene, in a region extending up to 180 kb from the EVI1 locus. The 3q26 breakpoints of the four t(3;3) were scattered in a wider region, one breakpoint (P14) being some 250 kb and another (P12) some 650 kb telomeric of the EVI1 locus.
Comparison of the breakpoint mapping using the EVI1 Breakapart probe and the BAC clones. Two sets of signal combinations were found among the 11 patients carrying an inv(3)(q21q26). Three patients (P2, P4 and P5) showed a split green signal in bands 3q21 and 3q26, with the blue signal being translocated to band 3q21 and the red signal remaining in band 3q26. This combination is compatible with a break having occurred in or close to the EVI1 gene. The remaining inv(3) patients showed the blue signal translocated to band 3q21, whereas both green and red signals were located in band 3q26. This combination is compatible with breakpoints localized farther centromeric of the EVI1 locus (Figure 2).
Two sets of signal combinations were found among the four patients carrying a t(3;3)(q21;q26). Two patients (P12 and P14) had a small derivative chromosome 3 (der(3) with a red signal and a large der(3) with a blue-green-blue-green-red signal combination compatible with a break having occurred telomeric to the EVI1 gene. The two remaining patients (P13 and P15) had a small der(3) with green and red signals and a large der(3) with a blue-green-blue-green-red signal combination, compatible with a break having occurred in or close to the EVI1 gene (Figure 2).
Discussion
Patients with inv(3)(q21q26) or t(3;3)(q21;q26) are now classified in a distinctive entity of AML with recurrent genetic abnormalities of prognostic significance (1). Although the inv(3)/t(3;3) is not restricted to AML, patients with MDS or CML in the blastic phase associated with these abnormalities have usually a progressive disease and a poor prognosis (7, 8).
Additional chromosomal abnormalities besides inv(3)/t(3;3) are frequently found, more particularly −7/7q-. A non-exhaustive search in the literature identified 278 patients, including ours, with inv(3) (191 patients, 68.7%) or t(3;3) (87 patients, 31.3%) (2, 3, 6, 37, 39, 41-44). Among these 278 patients, 190 (68.3%) had at least one additional abnormality, including 145 (76.3%) with −7/7q-. Retroviral integration experiments have shown that overexpression of the EVI1 gene was not sufficient to cause leukemia, indicating that development of malignancy likely depended on the occurrence of additional mutations (45). However, neither the exact link between inv(3)/t(3;3) and −7/7q-, nor the sequence of events, is fully understood (46). Recently, using array comparative genomic hybridization (array CGH), De Weer et al. identified two deleted regions in 7q35-36 in patients with inv(3)/t(3;3) overexpressing the EVI1 gene (47).
In contrast to most of the inversions and translocations associated with AML that lead to fusion genes, the inv(3)/t(3;3) does not generate a chimeric gene, but rather induces gene overexpression (10, 15, 26). Indeed, EVI1 overexpression is unlikely to be a consequence of the formation of fusion transcripts because of the position and transcription orientation of the genes involved in the rearrangement (27). The wide dispersion of the breakpoints in bands 3q21 and 3q26 and the heterogeneity of the genomic consequences necessitate further characterization of these chromosomal abnormalities by FISH.
Wieser et al. developed an interphase dual-color FISH assay using PAC clones to detect rearrangements of band 3q21 (33). They assembled PAC contigs covering both BCRs (centromeric and telomeric), which allowed them to map the majority of 3q21 breakpoints involved in inv(3)/t(3;3) (5). Wieser et al. also developed a similar assay using several overlapping PAC clones localized on the centromeric and telomeric sides of the EVI1 locus to detect rearrangements in band 3q26 (34). Other workers constructed dual-color probes located on either sides of the EVI1 locus using PAC or BAC clones (35-38). These constructions differed in size, ranging from a few hundreds kilobases to more than 1 megabase.
Shearer et al. developed a dual-color, double fusion FISH assay to detect RPN1–EVI1 gene fusion associated with inv(3)/t(3;3) using BAC clones. They labeled 3 BAC clones located in band 3q21 centromeric of GATA2, between GATA2 and RPN1, and telomeric of RPN1 in Spectrum Green and 11 BAC clones in band 3q26 covering the EVI1/MDS1 (MECOM) gene and beyond (39). They identified the fusion in 94% of the 47 samples analyzed.
We used a different approach because our goal was to map more precisely the breakpoints involved in inv(3)/t(3;3), and not only to verify the RPN1–EVI1 fusion by molecular cytogenetics. Our results showed a wide heterogeneity of breakpoints, as previously observed by Lahortiga et al. in 7 patients (27).
Breakpoints in band 3q21 were distributed in a 235 kb region, with breakpoints of 10 patients clustering in the so-called BCR-T and 2 in the BCR-C, with no preferential localization of the breakpoints involved in inv(3) or t(3;3). These results are compatible with those previously reported (15, 30-32). The distribution of the 3q26 breakpoints is much wider (about 900 kb), with a propensity for breakpoints involved in t(3;3) to be located telomeric of the EVI1 gene, as already shown (9, 15, 16).
No discrepancy between our results obtained with BAC clones and those using the EVI1 Breakapart probe was found. Therefore, this commercial probe could be used to screen not only patients with inv(3)/t(3;3), but presumably also patients with other chromosomal abnormalities involving band 3q26.
- Received July 19, 2011.
- Revision received September 1, 2011.
- Accepted September 2, 2011.
- Copyright© 2011 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved