Fluorescence In Situ Hybridization Characterization of New Translocations Involving TEL (ETV6) in a Wide Spectrum of Hematologic Malignancies

RECENT MOLECULAR studies show that theETV6 gene (previously known as TEL), a member of the E26 transformation-specific (ETS)-family of transcription factors located at 12p13,1 is involved in different chromosomal translocations associated with human leukemias (Fig 1A). For the t(3;12)(q26;p13), t(5;12)(q33;p13), t(9;12)(q34;p13), t(12;21)(p13;q22), and t(12;22)(p13;q11) the translocation partners were identified.1-6 These result in the expression of a chimeric transcript consisting of ETV6sequences fused to MDS1/EVI1 (3q26), PDGFRB (5q33),ABL (9q34), AML1/CBFA2 (21q22), and MN1(22q11), respectively. In ETV6 translocations involvingPDGFRB, ABL, and AML1/CBFA2 the helix-loop-helix (HLH) dimerization domain of ETV6influences or stimulates the activity of the fusion partner. In leukemias with a t(12;22), the aberrant MN1-ETV6protein is believed to have transforming capacity and the DNA-binding domain is thought to be the functional contribution ofETV6.⁶ In myeloproliferative disorders with a t(3;12)(q26;p13), the chimeric transcript consists of the first two exons of ETV6, which code for no known functional domains, fused toMDS1/EVI1 sequences suggesting that the oncogenic potential of this translocation could result from the ETV6 promoter driving the transcription of MDS1/EVI1.² Molecular analysis of the t(6;12)(q23;p13) recently described by Bohlander et al,7 in a B-cell acute lymphoblastic leukemia (ALL) cell line identified a novel gene on chromosome 6 named STL. TheETV6 breakpoint was localized in intron 2, upstream to the exon encoding the HLH domain. However, no obvious new chimeric reading frames were found and the hypothesis that the t(6;12)(q23;p13) does not lead to a fusion protein with oncogenic potential but to the elimination of normal ETV6 function was presented.

It is clear that ETV6 is a versatile element at the center of a network of genes involved in hematologic malignancies through different molecular mechanisms that are only partially understood. Here we report six leukemia cases and one B-cell non-Hodgkin's lymphoma (NHL) case with new chromosomal rearrangements involving ETV6. The breakpoints of a t(5;12)(q31;p13), t(6;12;17)(p21;p13;q25), t(7;12)(p15;p13), t(7;12)(p12;p13), t(7;12)(q36;p13), t(12;13)(p13;q12), and t(12;?)(p13;?) were characterized by fluorescence in situ hybridization (FISH) using a panel of DNA probes including an ordered set of cosmids covering the entire ETV6gene.

Patient material was collected at the Center for Human Genetics in Leuven, Belgium, during the last 3 years. Some clinical and hematologic findings of the reported cases are summarized in Table 1.

One-day cultures of bone marrow cells were used for cytogenetic analysis in all cases. Ten to 31 R- and G-banded karyotypes were analyzed and classified according to the International System for Human Cytogenetic Nomenclature.8 

FISH was performed as previously described.9 Chromosome 12p abnormalities were studied with the cosmids 4H9A (D12S158, assigned to 12p13.3), 123C12 (CDKN1B), and 9 cosmids forETV6 ordered from 5′ to 3′ (179A6 - 15A4 - 67C6 - 50F4 - 132B11 - 242E1 - 163E7 - 54D5 - 148B6).10 The position of the different ETV6 exons in these cosmids is shown in Fig 1B. Additional FISH experiments were performed using chromosome 5, 6, 9, 12, 13, 17, 19, 20, 21, and 22 painting probes labeled with bio-16-dUTP (Cambio, Cambridge, UK) or Spectrum Orange-dUTP (Vysis, Stuttgart, Germany). Three yeast artificial chromosomes (YACs) assigned to 5q31q33 (773B7, 939F12, and 888A7), and 10 YACs assigned to the 7p15p12 region (764B12, 776A4, 207E1, 881B6, 669G6, 16BC5, 959B3, 772B11, 802D6, and 908B12) were selected from the sequence-tagged site (STS)-based map reported by Green et al,11 Keen et al, ¹² and Chumakov et al.13 The cosB (PDGFRB),14 YAC Y6,15 and cosmid ICRFCIO2D12118 hybridize to, respectively, 5q33, 14q32, and 21q22.3. Chromosomes 7, 12, and 13/21 were identified by cohybridization with Texas Red-5-dUTP labeled centromere probes (Dupont, Boston, MA) for chromosome 7 (p7t1), 12 (pBR12/D12Z3), and 13/21 (pUC 1.76) in combination with G-banding using DAPI counterstaining. Between 5 and 12 abnormal metaphases were studied for each experiment. The FISH data were collected on a Leitz DMRB fluorescence microscope (E. Leitz Inc, Wetzler, Germany) equipped with a cooled black and white CCD camera run by SmartCapture software (Vysis, Stuttgart, Germany).

Seven new chromosome 12p translocations affecting the ETV6 gene were identified in patients with different malignant hemopathies including chronic myeloid leukemia (CML), acute myeloid leukemia (AML), ALL, and B-NHL. Cytogenetic findings of all seven cases are presented in Table 2. In cases 1, 3, 4, 5, and 7 chromosomal aberrations were identified at diagnosis, whereas in case 2 and 6 the abnormal karyotypes appeared during the course of disease. The 12p aberrations were found to be the sole abnormality in karyotypes of three patients; in one case a t(12)(p13) was associated with a monosomy 13, whereas the karyotypes of the three remaining cases displayed complex multichromosomal changes. The potential involvement of ETV6 was evaluated by FISH with a panel of cosmid probes covering the complete gene.10 FISH results are summarized in Table 3.

A t(5;12)(q31;p13) was found in a patient with atypical CML associated with marked eosinophilia. FISH with ETV6 cosmids showed that 179A6 (exon 1 of ETV6) hybridized to the der(5), whereas cosmid 67C6 (intron 1), 50F4 (exon 2), and 148B6 (exon 8)hybridized to the der(12) (Fig 2A). The chromosome 5 breakpoint was analyzed with a PDGFRB (5q33) probe (cosB) and three YACs (773B7, 939F12, 888A7) assigned to 5q31q33. All these probes hybridized to the der(12) indicating that the breakpoint occurred proximal to 773B7, the most centromeric probe examined (Fig 3A).

A second 12p13 translocation resulting in an ETV6 rearrangement was found in a pediatric patient with AML and a previous history of myelodysplastic syndrome (MDS). Cytogenetically the translocation was described as a t(6;12)(p21;p13). FISH with the cosmid 148B6 showed a hybridization signal on a der(12), whereas the cosmid 179A6 hybridized to 17q25. This implied a three-way translocation t(6;12;17)(p21;p13;q25), which was confirmed by FISH using chromosome 6 and 17 painting probes. Further FISH analysis showed that cosmid 15A4 (intron 1) hybridized to both the der(12) and der(17) (Fig 2B), which locates the breakpoint on chromosome 12 in the first intron of ETV6(Fig 1C).

Two other translocations, both cytogenetically defined as a t(7;12)(p14;p13), were found in pediatric patients with AML-M0 or ALL-L2. FISH analysis showed that in the AML case the translocation breakpoint was flanked by 50F4 and 132B11 (Fig2C), and thus occurred near exon 2 of ETV6 (Fig 1C). In the second case, cosmid 179A6 (exon 1) spanned the breakpoint (Fig 2D), and three cosmids 3′ to 179A6 hybridized with the der(12). To determine the breakpoint on the translocation partner, FISH analysis with chromosome 7p YACs was performed (Fig 3B and C and Fig 4). All three YACs (959B3, 16BC5, and 669G6) from a contig covering the Retinitis Pigmentosa 9 locus at 7p15¹² hybridized to the der(7) in the AML case or the der(12) in the ALL case, indicating a different 7p breakpoint for them. FISH analysis was then performed with YACs centromeric (881B6, 207E1, 776A4, and 764B12) or telomeric (772B11, 802D6, and 908B12) toRP9.11,13 The breakpoint of the AML t(7;12) was found to be spanned by two overlapping probes, 802D6 (D7S516, D7S1808, and D7S2416/1790 kb) and 908B12 (D7S1808, D7S2416, and D7S2564/1300 kb). A third overlapping YAC, 772B11 (D7S516 and D7S1808/1500 kb), hybridized only to the der(7) chromosome. These results suggested the localization of the breakpoint near D7S2416 at 7p15. The breakpoint of the ALL t(7;12) was narrowed down to the area flanked by the YACs 776A4 and 764B12 (7p12).

In a karyotype of the other pediatric patient with AML, complex chromosomal abnormalities, including add(7)(q36) and del(12)(p12p13), were observed. FISH with 179A6 (exon 1) and 148B6 (exon 8) showed hybridization signals on add(7) and del(12) chromosomes, respectively, and implied a t(7;12)(q36;p12). The ETV6 breakpoint of this translocation was localized in a region covered by 50F4 that was split between both derivative chromosomes (Fig 2E). Because exon 2 is located at the 3′ end of 50F4, the breakpoint could be mapped to intron 1 of ETV6 (Fig 1C).

The sixth ETV6 translocation identified by FISH as a t(12;13)(p13;q12) was masked by complex chromosomal rearrangements not detected during cytogenetic analysis. FISH with ETV6 cosmids 179A6 (exon 1) and 148B6 (exon 8) yielded hybridization signals on a small, unidentifiable chromosome and the der(12), respectively, whereas cosmid 54D5 (exons 5, 6, 7, and 8) spanned the breakpoint. All cosmids 5′ to 54D5 hybridized with the same small marker chromosome. However, none of them, nor a cosmid for CDKN1B (123C12), showed hybridization signals on the second chromosome 12, which was cytogenetically defined as a del(12)(p12p13). FISH analysis with the 12p13.3 cosmid 4H9A (D12S158) yielded hybridization signals on this del(12)(p) chromosome, confirming that the deletion is interstitial. Hybridization of the ETV6 probes to the small marker suggested a cryptic t(12;21)(p13;q22), masked by the cytogenetically recognized t(9;12)(q13;p13), but FISH with a chromosome 21 paint and a 21qter cosmid (ICRFC102D12118) showed no rearrangement of chromosome 21. The use of painting probes for chromosome 9, 12, 13, and 20 allowed the identification of this small chromosome as a der(13) from a reciprocal t(12;13)(p13;q12) masked by an additional and probably secondary t(9;13;20)(p13;q14;p12) involving the der(12)t(12;13)(p13;q12) (Fig 2F).

The karyotype of a seventh patient with a B-NHL showed rearrangements of both 12p chromosomes, cytogenetically defined as a del(12)(p11) and dic(12;17)(p11;p11). FISH analysis with cosmid 179A6 (exon 1) yielded surprisingly only a single hybridization signal on the terminal end of the add(14)(q32) suggesting a t(12;14)(p11;q32). This reciprocal translocation was confirmed by FISH with a cosmid 4H9A (D12S158) showing a signal on the add(14)(q32) and with a 14q32 YAC (Y6), which hybridized to a chromosome defined cytogenetically as a del(12p). Only one Y6 signal was observed in all abnormal cells analyzed, indicating a cryptic del(14)(q32). Cosmid 148B6 (exon 8) did not hybridize with the add(14)(q32), showing an intragenic deletion of part of ETV6, which was then shown to extend as far asCDKN1B (123C12) (Table 3). However, cosmids 148B6 and 123C12 yielded hybridization signals with a chromosome band 5p suggesting a 5;12 translocation. As cosmid 179A6 did not hybridize with this der(5), the second ETV6 allele also appeared to be inactivated by an intragenic deletion. Moreover, the 12p13.3 cosmid 4H9A (D12S158) was found to be translocated to another chromosome identified as a der(17)t(12;17)(p13;q11) using the chromosome 17 painting probe. Chromosome 17 material was also detected on a dic(12;17)(p11;p11), on the add(1)(p34) and a partially painted chromosome 17 suggesting a t(1;17)(p34;q24), and on a long arm of one of five marker chromosomes. All these data indicated the involvement of three copies of chromosome 17 in different chromosomal aberrations. Translocation of the telomeric 12p13.3 sequences (4H9A) to 17q11, deletion of the 5′ end of ETV6, and localization of hybridization signals from 163E7 and 148B6 cosmids on a distal but not terminal end of a der(5) chromosome (Fig 2G, inset) suggested a masked translocation of unknown material to the retained 3′ region ofETV6 on a der(5). Despite using chromosome 5, 12, 19, 20, and 22 libraries, the origin of this material could not be identified by FISH because of its small size, complex aberrations, and insufficient material resources. Summarizing the FISH and G-banding results, the chromosome 12 abnormalities found in this patient can be described as follows: a der(5)t(5;12;?)(p11;p11p13;?), der(12)del(12)(p13p13)t(12;14)(p11;q32), dic(12;17)(p11;p11) del(12)(p13p13), der(14)t(12;14)(p11;q32), and der(17)t(12;17)(p13;q11).

The breakpoint of a previously reported MDS case with a t(10;12)(q24;p13)16 was also further refined using theETV6 cosmids. The ETV6 breakpoint occurred in intron 2 and was flanked by cosmids 50F4 and 132B11 (Fig 1C).

Seven new chromosomal aberrations involving the ETV6 gene were detected in patients with CML, AML, ALL, and B-NHL and were further characterized by FISH. Although cytogenetic analysis already indicated 12p abnormalities in all these cases, FISH studies for most of them showed new and unexpected chromosomal rearrangements.

The (5;12)(q31;p13) translocation found in atypical CML showed a breakpoint gene in the first intron of ETV6 and a chromosome 5 breakpoint centromeric to 773B7, mapped proximally to PDGFRB.These findings clearly indicate that this translocation is different from the t(5;12)(q33;p13) resulting in theETV6-PDGFRB fusion cloned by Golub et al.1 Insufficient material did not allow to identify the target gene at 5q31 by FISH; however, the breakpoint is localized in a region where many growth factors and hormone receptor genes have been mapped including members of the interleukin (IL) gene family (IL3, IL4, IL5, IL9, IL12B, IL13), interferon regulatory factor-1 gene (IRF1), early growth response-1 gene (EGR1), colony stimulating factor-2 gene (CSF2), CD14 antigen and cell division cycle 25C gene (CDC25C). It is interesting to note that the t(5;12)(q31;p13) presented here, as well as the translocation 5;12 resulting in the ETV6-PDGFRB fusion protein, occurred in patients with marked eosinophilia.

In the second case a three-way t(6;12;17)(p21;p13;q25) disrupting theETV6 gene in intron 1 was identified. The 5′ end ofETV6, including exon 1, was translocated to 17q25, whereas the remaining part of the gene with the HLH and ETS DNA-binding domains was retained on the der(12) and became juxtaposed to unknown 6p21 sequences. Among the candidate genes on 6p21 are PBX2(pre–B-cell leukemia transcription factor 2), TNFA (tumor necrosis factor-α), PIM1, CDKN1A/WAF1(cyclin-dependent kinase inhibitor 1A), CBFA1/AML3 (α subunit of a core binding factor), and CCND3 (cyclin D3).

Two apparently similar translocations, both cytogenetically defined as t(7;12)(p14;p13), appeared to be molecularly different after FISH analysis. Although both translocations have a breakpoint in the 5′ end of ETV6, they affected different regions on chromosome 7p (see Results). The breakpoint of the AML t(7;12) was near the D7S2416 locus at 7p15. The breakpoint is contained in the Centre d'études du Polymorphisme Humain YACs 802D6 and 908B2. The HOX-A gene cluster was mapped to 7p15 and can be considered candidate genes. However, polymerase chain reaction (PCR) analysis with expressed sequence tours developed forHOXA1 and HOXA4 did not detect the presence of these genes on either YAC excluding these candidates (results not shown). Of interest in a perspective of our finding is another published case of a t(7;12)(p15;p13) found in a 4-year-old boy with a minimallydifferentiated AML (French-American-British classification AML-M0).17 This could be a relevant association because the AML-M0 is poorly characterized at the cytogenetic level. The largest published series of AML-M0 comprises 21 adult patients with clonal abnormalities,18 and although 12p chromosomal aberrations were found in three patients, no t(7;12) was present among these cases. Moreover, involvement of this region in myeloid leukemia with a typical t(7;11)(p15;p15) or with a del(7p) has been already reported.19,20 

Another 7;12 translocation involving the long arm of chromosome 7 (q36) and ETV6 was found in a 1-year-old boy with AML. The ETV6breakpoint of this translocation was localized in intron 1 of ETV6 upstream to exons encoding the HLH domain. The partner gene at 7q36 remains unknown, among the candidate genes is a cyclin-dependent kinase 5 (CDK5) involved in a regulation of cell cycle.

FISH analysis of the sixth case diagnosed as ALL-L2 showed disruption of ETV6 by a cryptic t(12;13)(p13;q12) masked by additional multichromosomal changes not discernible by cytogenetic analysis. TheETV6 breakpoint of t(12;13) is covered by cosmid 54D5, containing exons 5 to 8 of ETV6, and the translocation is associated with loss of the second ETV6 allele as a consequence of del(12)(p12p13). The 13q12 region involved in the t(12;13) is known to carry the FGF9 (fibroblast growth factor 9), FLT1and FLT3 (fms-related tyrosine kinase 1 and 3), IPF1(insulin promoter factor 1), and BRCA2 (breast cancer 2) genes. Some other ALL cases with an analogous t(12;13) were previously reported.21-23 Although in some cases the breakpoints were interpreted differently, there is a possibility that all of them involve ETV6 and the same unknown sequences at 13q12.

The most complex karyotypic changes affecting both 12p chromosomes andETV6 were observed in a B-NHL case. FISH analysis showed that one chromosome 12 was rearranged by a t(12;14)(p11;q32) with a breakpoint typical for B-NHL lymphoma at 14q32 corresponding to the Ig heavy chain gene locus. The second chromosome 12p was involved in three different translocations, a t(5;12;?)(p11;p11p13;?), a t(12;17)(p13;q11), and a dic(12;17)(p11;p11). The 12p13 breakpoint in the t(5;12;?) that might have resulted in an ETV6 fusion transcript was localized in the 5′ end of the gene upstream to cosmid 163E7 and upstream to the HLH coding exons. The partner chromosome involved in this translocation could not be identified. FISH analysis resulted in the detection of two different cryptic microdeletions of 12p13 affecting both ETV6 alleles. One of them, independent from a t(12;14)(p11;q32), covered the 3′ end ofETV6 (163E7 and 148B6) and CDKN1B (123C12), whereas the second intragenic microdeletion was associated with a t(5;12;?)(p11;p11p13;?) and involved the 5′ end of ETV6(179A6). This latter finding indicated that the putative reciprocal chimeric transcript of a t(5;12;?)(p11;p11p13;?) containing the 5′ end ETV6 domain was absent in the malignant cells and did not play a significant role in the development and/or progression of this lymphoma. It is noteworthy that deletion of the secondETV6 allele is a commonly observed secondary phenomenon in pre–B-ALL cases with a t(12;21)(p13;q22).24,25 In two cases the deletion was shown to be intragenic inETV6,25,26 suggesting that the gene was probably the actual target of the deletion. It was already hypothesized that wild type ETV6 might reduce the oncogenic and growth-stimulating properties of the fusion proteins and that, therefore, its loss would provide an additional proliferative advantage to the malignant cells. The characterization of two new cases with biallelic ETV6 rearrangements [together with at least one more leukemia case characterized by a t(9;12;14)(q34;p13;q22)/ETV6-ABL and del(12)(p11;p13)²⁷; B-cell ALL cell line with a t(6;12)(q23;p13)/ETV6-STL and a microdeletion of the otherETV6 allele⁷; and a myeloid leukemia case with two translocations involving ETV6, namely, t(3;12)(q26,p13) and t(9;12)(p24;q15;p13)²] indicates that biallelic aberrations of ETV6 are not exclusively associated with a subtype of ALL with a t(12;21).

In summary, seven new chromosomal abnormalities affecting ETV6have been identified in patients with atypical CML associated with eosinophilia in MDS, in young patients with acute leukemias and, for the first time, in a case of B-NHL. In two leukemia cases the ETV6 aberrations appeared during the course of the disease suggesting that the disruption of this gene can not only initiate the development but also be involved in the progression of the malignant disorder. The latter can be supported by the karyotypic findings in a case of B-NHL where the ETV6 affected translocation was coexisting with a typical 14q32/IgH translocation regarded as a primary rearrangement in lymphomagenesis. The frequency of these new ETV6 translocations in hemopathies is difficult to evaluate because of multichromosomal aberrations impeding cytogenetic detection.

FISH analysis of seven ETV6 rearrangements showed that in six of them the breakpoint occurred in the 5′ end of ETV6upstream of exons coding for the HLH domain. Together with the previously reported t(10;12)(q24;p13),16 they might generate chimeric proteins with features of transcriptional activators analogous to the t(12;22)(p13;q11) fusion protein, but it is also conceivable that the oncogenic properties of some of these translocations result from the ETV6 promoter driving the partner gene as suggested for a t(3;12)(q26;p13) or from the disruption of as yet unidentified domain(s) encoded by the first two exons. In the remaining case with a t(12;13)(p13;q12), the breakpoint was localized in the 3′ end of ETV6, probably upstream to exons coding for its ETS DNA-binding domain as was previously reported for the t(5;12)(q33;p13), t(9;12)(q34;p13), and t(12;21)(p13;q22), which might lead to the expression of chimeric transcripts containing either the HLH domain or the ETS domain of ETV6. On the other hand, regarding that ETV6 is a candidate tumor suppressor gene, it is possible that some of the ETV6-related translocations, especially those associated with a complete or partial deletion of the second ETV6 allele, might inactivate the ETV6 gene as hypothesized for a t(6;12)(q23;p13).

The data presented here support the hypothesis of the multiple contributions of ETV6 in the pathogenesis of hematologic disorders and show the occurrence of ETV6 aberrations in hemopathies as diverse as CML, AML, ALL, and B-NHL. The cryptic deletions of the nontranslocated ETV6 allele found in two cases further emphasize the significance of the inactivation of the wild typeETV6 protein for the development and/or the progression of some hematologic malignant disorders.

The authors thank Dr C. Inglehearn (Institute of Ophthalmology, University of London, London, UK) and Dr F. Matsuda (Center for Molecular Biology and Genetics, Kyoto University, Sakyo-ku, Kyoto, Japan) for providing us with YACs specific for RP9 and 14q32, respectively. We are grateful to Magda Dehaen and the technicians of the leukemia laboratory of the Center for Human Genetics for their assistance.