EVI1 activation in blast crisis CML due to juxtaposition to the rare 17q22 partner region as part of a 4-way variant translocation t(9;22)

Background Variant translocations t(9;22) occur in 5 to 10% of newly diagnosed CMLs and additional genetic changes are present in 60–80% of patients in blast crisis (BC). Here, we report on a CML patient in blast crisis presenting with a four-way variant t(9;22) rearrangement involving the EVI1 locus. Methods Dual-colour Fluorescence In Situ Hybridisation was performed to unravel the different cytogenetic aberrations. Expression levels of EVI1 and BCR/ABL1 were investigated using real-time quantitative RT-PCR. Results In this paper we identified a patient with a complex 4-way t(3;9;17;22) which, in addition to BCR/ABL1 gene fusion, also resulted in EVI1 rearrangement and overexpression. Conclusion This report illustrates how a variant t(9;22) translocation can specifically target a second oncogene most likely contributing to the more aggressive phenotype of the disease. Molecular analysis of such variants is thus warranted to understand the phenotypic consequences and to open the way for combined molecular therapies in order to tackle the secondary oncogenic effect which is unresponsive to imatinib treatment.


Background
Chronic myeloid leukemia (CML) has a typical indolent chronic phase that may last several years but will ultimately progress into acute myeloid leukemia (AML) or acute lymphoid leukemia (ALL). The typical associated translocation t(9;22)(q34;q11), leading to the BCR/ABL1 fusion gene and constitutive activation of the ABL1 tyrosine kinase on 9q34, is considered to be the initial transforming event [1]. Instead of the classical t(9;22) translocation, cryptic and variant translocations occur in about 5 to 10 % of all CML cases. CML in blast crisis is often accompanied by the presence of additional chromo-some aberrations [2]. Amongst those, activation of the EVI1 gene has been reported in a small percentage of patients [3]. Ectopic expression of the EVI1 gene is usually due to recurrent 3q26 translocations such as the t(3;21)(q26;q22) (AML1/EVI1) and the inv(3;3)(q21q26). Transcriptional activation of EVI1 can also occur in the absence of genomic rearrangement at this locus [4]. In general, EVI1 upregulation confers a poor prognosis in hematological malignancies [5].
In this study, we describe how the molecular characterization of the translocation breakpoints of a variant t(9;22) in a patient with CML in blast crisis lead to the discovery of involvement of the EVI1 locus. We also discuss the importance of the study of secondary genomic events present or occurring during blast crisis with respect to the development of strategies for treatment of CML patients in blast crisis.

Patient material
A diagnostic bone marrow sample was obtained from a patient who presented with features of chronic myeloid leukemia in blast crisis. The patient died nine months after diagnosis.
This study was approved by the Ethical Committee of the Ghent University Hospital (2003/273).

Cytogenetic analysis
Cytogenetic analysis of the diagnostic bone marrow sample was performed according to standard methods. A 24hour bone marrow culture was performed and chromosomes were G-banded with a trypsin-Giemsa stain (GTGbanding). Fifteen metaphases were analyzed and the karyotype was described according to the International System for Human Cytogenetic Nomenclature [6].

Fluorescence in situ hybridisation analyses
Dual-colour Fluorescence In Situ Hybridisation (FISH) was performed on the diagnostic bone marrow sample using a t(9;22) specific BCR/ABL1 dual color, dual fusion probe (Abbott Vysis, Germany). Commercial Whole Chromosome Paints (WCP) for chromosomes 3, 9, 17 and 22 (Metasystems, Germany) were used to unravel the translocations in the bone marrow sample. A commercial centromeric probe for chromosome 3 (Abbott Vysis, Germany) was used to facilitate identification of chromosome 3. All non-commercial clones were acquired from the Sanger Wellcome Trust Institute, Hinxton, Cambridge (United Kingdom). The EVI1 specific probes were labelled with biotin-16-dUTP (Roche Diagnostics, Belgium) and digoxigenin-12-dUTP (Roche Diagnostics, Belgium) and FISH analysis was performed as previously described [7,8]. Biotin labelled probes were detected with Fluorescein Iso-ThioCyanate (FITC)-conjugated anti-biotin antibodies (Invitrogen, Belgium) and digoxigenin labelled probes with TetramethylRhodamine IsoThioCyanate (TRITC)conjugated anti-digoxigenin antibodies (Roche Diagnostics, Belgium). A minimum of 100 nuclei and a minimum of 5 metaphases were scored using a fluorescence microscope (Axioplan 2, Zeiss, Belgium) and images were captured using a black and white CCD camera and images were processed with the ISIS software program (MetaSystems, Germany).

Real-time quantitative RT-PCR
Total RNA was extracted from the diagnostic bone marrow sample using the Trizol LS reagent (Invitrogen, Belgium) according to the manufacturer's recommendations. cDNA was prepared from 2 μg of total RNA with the iScript cDNA Synthesis Kit (Bio-Rad, Belgium) according to the manufacturer's instructions.
QRT-PCR for the BCR/ABL1 fusion transcript was done according to the Europe Against Cancer program (EAC) protocol [11].
FISH with the BCR/ABL1 dual-color dual-fusion probe revealed two fusion signals together with a single green and red signal in interphase nuclei. In metaphases, one fusion was located on the Philadelphia chromosome whereas the second fusion signal was present on the marker chromosome instead of on the der(9)t(9;22)(q34;q11.2) ( Fig. 2A). Real-time quantitative RT-PCR (qRT-PCR) confirmed the presence of the BCR/ABL1 fusion and revealed a major MBCR transcript (data not shown).
Further FISH characterisation of this t(9;22) variant translocation, revealed involvement of chromosomes 3 and 17. FISH analysis with two probe combinations for the EVI1 locus (EVI1 centromeric and overlapping; EVI1 telomeric and overlapping) revealed a breakpoint 5' and telomeric of EVI1 and translocation of the telomeric chromosome 3 segment to the der(9) (Fig. 2B). The translocation of the ABL1/BCR segment from the der(9) to the der(17) was accompanied by translocation of the distal 17q segment immediately distal to the EVI1 locus on the der(3). FISH with whole chromosome paints confirmed these observations (Fig. 2C-D).
Ectopic expression of the full length EVI1 transcript was detected with qRT-PCR using the 5' located EVI1 primer pair (Fig. 3A). qRT-PCR with the 3' located cEVI1 primer pair was able to detect additional ectopic expression 5' EVI1 variant transcripts (Fig. 3B). No expression of the MDS1/EVI1 or MDS1 transcripts could be detected with qRT-PCR (data not shown).

Discussion
We describe the molecular characterization of a complex 4-way t(3;9;17;22) translocation, which in addition to BCR-ABL1 gene fusion also resulted in EVI1 overexpres-Representative karyotype from bone marrow of the patient Figure 1 Representative karyotype from bone marrow of the patient. Arrows indicate a derivative chromosome 3 with an addition, a marker chromosome, two derivative chromosomes 9 (der(9)), del(16)(q23), the loss of a chromosome 17 and a Phchromosome. sion. Based on FISH and karyotype data we propose a twostep mechanism [12] for generating the complex 4-way translocation, i.e. first the formation of a classical t(9;22), followed by a subsequent three-way translocation involving chromosomes 3, 17 and the derivative chromosome 9.

FISH analysis on metaphases from bone marrow slides
The latter rearrangement was the result of translocation of the ABL1/BCR segment from the der(9) to chromosome 17, translocation of the distal 17q segment to the 3q26 locus on chromosome 3 and translocation of the distal part of chromosome 3 to the der(9). Further molecular investigation using FISH indicated that the 3q26 breakpoint located immediately distal of the 5' end of the EVI1 locus. As expected qRT-PCR revealed ectopic expression of EVI1. A third and final step encompassed the duplication of the der(9) in this patient (Fig. 4A-C).
EVI1 overexpression is a well established unfavourable prognostic marker in AML patients, and is often found in progressive or blast crisis CML [4,13,14]. It is therefore likely that, like for the additional alterations such as trisomy 8 and isochromosome 17q, the observed EVI1 overexpression contributes to the aggressive phenotype of the blastic CML phase. The presence of the additional cytogenetic abnormality del(16)(q23) in this patient could also negatively influence disease progression.
Apart from the known recurrent secondary changes occurring during progression or blast phase in CML, most additional rearrangement breakpoints have remained unexplored at the molecular level. It can be assumed that such breakpoints target specific genes leading to additional proliferative or survival advantage to the tumour cell. The present study illustrates the importance of characterization of such secondary rearrangements or breakpoints in order to understand the molecular basis of the acute phase in CML.
Given the progress in CML treatment due to the development of 1 st and 2 nd generation small molecules (TK-inhibitors) such as imatinib, nilotinib and dasatinib [15], targeting of such additional oncogenic events now represents a new challenge for future CML treatment. Even though currently the number of targeted therapies is still limited, it can be anticipated that in the future, combination therapies with imatinib and compounds targeted at the secondary changes can become a reality. Therefore, we propose a systematic testing of 3q26 aberrations in CML patients in blast crisis or CML patients with an evolutionary disease course. Since to date only a few target genes of the transcriptional repressor EVI1 are known [16], further research will be needed to identify more genes implicated in EVI1 pathogenesis thus opening the way for development of novel targeted therapeutics.

Conclusion
In conclusion, we report on a CML patient in blast crisis presenting with a three-step variant Ph-positive chromosome rearrangement, involving the EVI1 locus. This report shows that the variant translocation can specifically target a second oncogene which most likely contributes to the EVI1 expression measured by qRT-PCR more aggressive phenotype of the leukemia. Molecular analysis of t(9;22) variants is needed to understand its phenotypic consequences and to open the way to combined molecular therapies in order to counteract the secondary oncogenic effect which is unresponsive to imatinib treatment.