Many types of leukaemia have been associated with chromosomal translocations that form oncogenic fusion proteins. In the October issue of Nature Genetics, Anne Hagemeijer and colleagues describe a unique fusion between NUP214 and ABL1 that occurs only on episomes in human T-cell acute lymphoblastic leukaemia (T-ALL) cells. This genetic aberration could not be detected by conventional cytogenetics and is the first episomal fusion gene associated with human cancer.

In studying the role of the tyrosine kinase ABL1 in T-ALL, the authors observed that ABL1 was amplified in about 6% of patients with this cancer. Interestingly, these gene amplifications were all extra-chromosomal. Further analysis of the ABL1-carrying episome revealed that it contained several other genes, including NUP214, and that all these genes were derived from the chromosome region 9q34. The authors discovered that this chromosomal region becomes amplified and circularized, leading to formation of an episome that carried an in-frame fusion between the carboxyl terminus of ABL1 and the amino terminus of NUP214. Examination of leukaemic cells from 175 patients with T-ALL revealed that 10 (6%) expressed NUP214–ABL1.

How might this fusion protein contribute to cell transformation? First, it contains the coiled-coil oligomerization motif of NUP214, which mediates the protein–protein interactions of this nuclear-pore protein. Second, it contains the SH3, SH2 and kinase domains of ABL1, allowing it to function as a constitutively active tyrosine kinase. Hagemeijer and colleagues found that the fusion protein was also expressed by three T-ALL cell lines, and that one of the known substrates of ABL1, CRKL, was constitutively phosphorylated in these cells. Importantly, treatment of these cells with imatinib, a selective inhibitor of ABL1 kinase activity, decreased CRKL phosphorylation in all three cell lines and inhibited the proliferation of one cell line. Phosphorylation of CRKL was also inhibited by imatinib in primary bone-marrow cells isolated from one of the patients found to carry the fusion gene.

Additional leukaemia-associated mutations were also detected in bone-marrow cells of these patients. These mutations were included in the authors' multistep model of T-ALL pathogenesis, beginning with deletion of tumour suppressors ( CDKN2A and CDKN2B ), overexpression of a transcription factor (HOX11) and, finally, expression of the constitutively active tyrosine kinase NUP214–ABL1. As expression of the NUP214–ABL1 fusion protein was associated with an especially aggressive course of disease, imatinib will be an important new therapeutic approach for these patients. The authors also emphasize that genome-analysis tools such as high-resolution array comparative genome hybridization are important for detecting cancer-associated genetic amplifications and deletions that cannot be detected by conventional cytogenetics.