Constitutively activated mutants of the non-receptor tyrosine kinase ABL1 play a central role in the pathogenesis of clinically and morphologically distinct chronic and acute leukaemias (Chase & Cross, 2006). By far the most frequent and best-studied ABL1 fusion gene is BCR-ABL1, which results from an acquired reciprocal t(9;22)(q34;q11) (Shtivelman et al, 1985; Ben-Neriah et al, 1986). Other ABL1 fusion genes have been described but are uncommon. Imatinib is a specific inhibitor of several tyrosine kinases, including ABL1, and induces long-term complete haematological and cytogenetic remissions in most patients with chronic phase chronic myeloid leukaemia (CML) (Druker et al, 2001; Deininger et al, 2009). However, a substantial proportion of patients with advanced phase CML or BCR-ABL1 positive ALL are initially refractory to imatinib treatment or lose imatinib sensitivity over time and relapse (Apperley, 2007; Pui et al, 2008). NUP214-ABL1 and EML1-ABL1 have been shown to be imatinib sensitive in-vitro and two single case reports documented dasatinib sensitivity for NUP214-ABL1 and RCSD1-ABL1 in-vivo (Deenik et al, 2009; Mustjoki et al, 2009). This study investigated two B-ALL patients with acquired chromosomal rearrangements each involving the chromosomal band 9q34 and identified two novel ABL1 fusion genes potentially sensitive to imatinib treatment. The t(3;9)(p12;q34) was identified in a 16-year-old Caucasian female (Case 1) who was diagnosed with pre B cell ALL in October 2001, with an initial white blood cell (WBC) count of >50 000 × 109/l. Her past medical history was significant only for recurrent deep venous thrombosis (DVT) secondary to protein S deficiency. Bone marrow cytogenetics was 46,XX,t(3;9)(p12;q34). Fluorescent in situ hybridization (FISH) revealed loss of one copy of the p16 locus. Based on her age and initial WBC count she was considered high risk. Treatment was initiated according to the standard arm for high-risk patients of the COALL-97 protocol of the German Cooperative Study Group for Childhood Acute Lymphoblastic Leukaemia (COALL) (Escherich et al, 2010). Therapy was complicated by recurrent DVT, catheter-associated septicaemia and steroid-induced diabetes mellitus. She achieved a complete response and in February 2002 she was consolidated with a haplo-identical transplant from her father. Complications included grade III graft-versus-host disease of the skin and fungal pneumonia, both of which resolved on therapy. On last follow-up, in the spring of 2010, she continued in complete remission, with excellent functionality (attending medical school). The male B-ALL patient (Case 2) with t(5;9)(q23;q34) was diagnosed in April 2004, aged 29 years, when he presented with a WBC count of 161 × 109/l. His cells showed a classical B-cell immunophenotype and bone marrow cytogenetics showed a t(5;9)(q23;q34) and a 20q-. FISH showed evidence of the ABL1 gene on both derivative chromosomes. BCR was apparently not involved. He responded well to initial chemotherapy but relapsed early. He then received hyper-C-VAD (hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone) and again responded. He relapsed again at the end of 2004 and was started on imatinib; the effect of this was transient and despite other supportive measures, including hydroxycarbamide, he died in the spring of 2005. FISH on the ABL1 gene indicated that the t(3;9)(p12;q34) and the t(5;9)(q23;q34) targeted ABL1; however, no BCR-ABL1 fusion was seen by FISH or reverse transcription polymerase chain reaction (RT-PCR) in either case. Most ABL1 fusions reported to date result in the partner gene fusing to ABL1 exons 2 or 3. To identify the t(3;9) partner, we therefore performed 5′-RACE PCR from both these exons. These initial attempts failed despite the fact that normal 5′ABL1 sequence was readily obtained (data not shown). 5′ RACE primers were subsequently designed in ABL1 exon 4 (ENST00000318560). Sequencing of the products revealed several clones in which FOXP1 (ENST00000318789) exon 19 was fused in frame to ABL1 exon 4. The presence of the FOXP1-ABL1 fusion was confirmed initially by RT-PCR. The reciprocal ABL1-FOXP1 product was detectable by single step PCR. Cloning and sequencing revealed that the reciprocal product fused ABL1 exon 3 to an alternative RNA isoform (NM_032682) of FOXP1. The FOXP1-ABL1 fusion was specifically amplified from patient cells by single step PCR, but was not detectable in normal controls (Fig 1A, B). To further confirm the presence of the fusion the genomic breakpoint was amplified by long PCR, cloned, and sequenced. In Case 2, 5′ RACE PCR on the t(5;9)(q23;q34) from ABL1 exons 2 and 3 failed, after the previous cloning of the t(3;9) translocation, 5′-RACE PCR primers were designed in ABL1 exon 4 (ENST00000318560). Sequencing of the products revealed several clones in which SNX2 (ENST00000379516) exon 3 was fused in frame to ABL1 exon 4. The presence of the SNX2-ABL1 fusion was again confirmed by RT-PCR. The reciprocal ABL1-SNX2 product was not detectable by single step PCR. As shown in Fig 2A, B, the SNX2-ABL1 fusion was specifically amplified from patient cells by single-step PCR, but was not detectable in normal controls. To further confirm the presence of the fusion the genomic breakpoint was amplified by long PCR, cloned, and sequenced (data not shown).
This study characterized the genomic breakpoints in two B-ALL cases with chromosome 9q34 rearrangements and identified FOXP1 at 3p12 and SNX2 at 5q23 as novel ABL1 fusion partners. This brings the number of known ABL1 partner genes to nine. Fusion genes involving ABL1 are excellent drug targets, as exemplified by the activity of imatinib in BCR-ABL1 positive diseases, particularly CML. The clinical course of patients with ABL1 fusion genes other than BCR-ABL1 may be different and highly dependent on the underlying disease. Response rates and risk of relapse may be similar to CML in patients with chronic phase myeloproliferative neoplasms or similar to BCR-ABL1 positive ALL in patients with acute leukaemias, respectively. As the fusions described here were referred specifically for investigation of 9q rearrangements and were not part of a series that underwent systematic cytogenetic investigation, it is not possible to accurately determine the frequency of these abnormalities. However, we estimate that they probably account for <1% of ALL cases. In most previously described fusion genes the partner fuses with ABL1 exon 2. However, similar to the recently described fusion partners SFPQ and RCSD1 (Hidalgo-Curtis et al, 2008; Mustjoki et al, 2009), both partner genes described here are fused to ABL1 exon 4. Nearly the entire FOXP1 gene including the forkhead domain was fused in 5′ position to ABL1. The SNX2 gene is located at 5q23 and belongs to the sortin nexin (SNX) family, which functions within the endocytic network, including endocytosis, endosomal sorting and endosomal signalling (Cullen, 2008). The N-terminal domain of SNX2 without the PX domain is fused to ABL1 exon 4. Although rare, the potential sensitivity to imatinib implies that these fusions should be searched for in acute leukaemia. The screen for such uncommon, but clinically significant fusion transcripts can easily be included in multiplex PCR panels used in routine diagnostic work-ups. Targeted therapy has revolutionized the treatment of CML and there is considerable optimism that targeting ABL1 fusion genes other than BCR-ABL1 will bring at least short term benefits for those patients. On the other hand, targeted therapy is very expensive and it will be important to develop markers to identify those patients who are most likely to benefit from treatment.