BCR-ABL1 is frequently associated with CML and B-ALL, but is rarely found in T-ALL. We recently identified two variant ABL1 fusions in T-ALL: NUP214-ABL1, associated with episomal amplification of ABL1 in 6% of T-ALL cases, and EML1-ABL1, associated with the cryptic translocation t(9;14)(q34;q32) in 1 patient. Similar to BCR-ABL1, NUP214-ABL1 and EML1-ABL1 are constitutively activated tyrosine kinases that transform the Ba/F3 cell line to interleukin-3 (IL3) independent growth. In the case of NUP214-ABL1, however, the Ba/F3 cells need a significantly longer period to obtain the same level of proliferation compared to BCR-ABL1 and EML1-ABL1 transformed cells. In mouse bone marrow transplantation experiments, BCR-ABL1 induced CML with a latency of 3 weeks, while with EML1-ABL1 the mice developed disease after several months, and for NUP214-ABL1 no leukemia developed within 12 months after transplantation. These data suggest that EML1-ABL1 and NUP214-ABL1 are weaker oncogenes compared to BCR-ABL1.

To gain further insights in these differences, we generated a number of deletion constructs of BCR-ABL1, EML1-ABL1 and NUP214-ABL1 and assayed the respective proteins for autophosphorylation and for their ability to transform Ba/F3 cells. For BCR-ABL1, we observed that the coiled-coil domain is not strictly required for kinase activity and transformation of Ba/F3 cells, as reported previously. In contrast, the coiled-coil domain of EML1 is sufficient and required to generate a constitutively activated EML1-ABL1 fusion protein. In the case of NUP214-ABL1, the coiled-coil domains are required, but not sufficient to generate an activated NUP214-ABL1 fusion in Ba/F3 cells, and also deletion of the N-terminal and C-terminal regions of NUP214 results in a loss of activity of NUP214-ABL1. Additional experiments confirmed that EML1-ABL1 is activated through homodimerization, while the exact mechanism of activation of NUP214-ABL1 remains unclear. In contrast to BCR-ABL1 and EML1-ABL1, NUP214-ABL1 seems to have a lower kinase activity and lacks detectable phosphorylation of the activation loop of the kinase domain. NUP214-ABL1 interacts with the nuclear pore proteins NUP62, NUP88 and RanBP2, is partially localized at the nuclear envelope, and phosphorylates RanBP2.

In conclusion, we describe significant differences between BCR-ABL1, EML1-ABL1 and NUP214-ABL1. For BCR-ABL1, the coiled-coil domain is sufficient for kinase activation, however BCR-ABL1 can also be activated by coiled-coil independent mechanisms. These mechanisms explain the high kinase activity and strong transforming capacity of BCR-ABL1. For EML1-ABL1, the coiled-coil domain is the only domain that can activate the kinase, probably explaining its weaker transforming capacity in mouse bone marrow transplantation models when compared to BCR-ABL1. NUP214-ABL1 is a very weak oncogene, the coiled-coil domains of NUP214 need to cooperate with other domains to activate NUP214-ABL1 adequately in Ba/F3 cells. These results may explain why NUP214-ABL1 is always amplified in T-ALL patients, and why NUP214-ABL1 is associated with T-ALL and not with CML.

Disclosure: No relevant conflicts of interest to declare.

Author notes

*

Corresponding author

Sign in via your Institution