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 chromosome 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.

Detailed cytogenetic analysis of 15 G-banded metaphases from the diagnostic bone marrow revealed the following karyotype: 50, XY, add(3)(q26),+8, add(9)(q34),+add(9)(q34),+10,+12, del(16)(q23),-17, der(22)t(9;22)(q34;q11),+mar [cp15] (Fig. 1).

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).

Based on these FISH analyses, the karyotype was revised as 50, XY, der(3)(3pter→q26::17q22→qter),+8, der(9)(9pter→q34::3q26→qter),+der(9)(9pter→q34::3q26→qter),+10,+12, del(16)(q23), der(17)(17pter→q22::9q34q34::22q11→qter), der(22)(22pter→q11::9q34→qter) [cp15].

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).

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 overexpression. Based on FISH and karyotype data we propose a two-step 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. 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 41314. 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.

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 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.

ABL1: ABeLson murine leukemia viral oncogene homolog 1; ALL: Acute Lymfoid Leukemia; AML: Acute Myeloid Leukemia; BC: Blast Crisis; BCR: Breakpoint Cluster Region; CML: Chronic Myeloid Leukemia; EAC: Europe Against Cancer; EVI1: Ecotropic Viral Integration site 1; FISH: Fluorescence In Situ Hybridisation; FITC: Fluorescein IsoThioCyanate; MDS1: MyeloDysplasia Syndrome 1; qRT-PCR: Quantitative Reverse Transcription Polymerase Chain Reaction; TK: Tyrosine Kinase; TRITC: TetramethylRhodamine IsoThioCyanate; WCP: Whole Chromosome Paint.

The authors declare that they have no competing interests.

ADW carried out the genetic studies and drafted the manuscript. AC is the patients referring physician. BC helped with EVI1 expression analysis, JD performed the blood cell counts, BV analyzed the BCR-ABL1 expressions and JP performed the immunophenotyping. BP, NVR and FS helped to draft the manuscript.