Abstract

In 95% of chronic myeloid leukemia (CML) and in 25% of acute lymphatic leukemia (ALL) the t(9;22) translocation fuses the bcr gene on chromosome 22 to the abl gene on chromosome 9 and vice versa. On 22+ the different breakpoints leads to the formation of two different major fusion genes: the major breakpoint (M-bcr) related to CML and the minor (m-bcr) related to ALL. The chimaeric fusion gene on 22+ (Philadelphia-chromosome) encodes for the BCR/ABL protein, the p210(BCR/ABL) in CML and the p185(BCR/ABL) in Ph+ALL. The fusion gene on 9+ encodes for the reciprocal ABL/BCR proteins, the p40(ABL/BCR) in CML and the p96(ABL/BCR) in Ph+ALL. The respective ABL/BCR transcripts are detectable in 65% of CML and 100% of Ph+ ALL patients. The ABL/BCRs are BCR mutants and thus N-terminally truncated Rho-guanine-nucleotide exchange factors (Rho-GEF’s). It is known that the N-terminal truncation can confer transformation potential to Rho-GEFs, such as NET-1. In addition, both ABL/BCRs, like wt BCR, contain a C-terminal Rac-GTPase activating protein (GAP)-domain. CML-associated ABL/BCR (p96(ABL/BCR)) differs from the Ph+ ALL-associated p40(ABL/BCR) in that that it misses the ‘dbl homology domain’(DH domains) of potential oncogenic function. Hence it seems that Ph+ALL blasts, in contrast to CML-blasts, express, as a consequence of t(9;22) translocation, two oncogenic fusion proteins, the p185(BCR-ABL) as well as the p40(ABL/BCR) protein. Actually nothing is known about the contribution of the reciprocal t(9;22) translocation products to the CML- and the ALL-phenotype. Thus we studied the phenotype induced by the ABL/BCRs in hemopoietic progenitors.

Here we report that both ABL/BCRs i) lost the capacity of wt BCR to suppress the activation of RAC by its Rac-GAP domain, but did not influence the activation status of Rho or cdc42; ii) as a consequence of the deregulation of Rac the cytoskeleton modelling by BCR (Filopodia - cdc42-like phenotype) was altered in p40(ABL/BCR)- and p96(ABL/BCR)-expressing fibroblasts (stress fibers - Rho-like phenotype and “microspikes”, respectively); iii) the increase of migration of BCR-expressing 32D cells into a stroma cell-spheroid model was reverted in p40(ABL/BCR)- and p96(ABL/BCR)-expressing 32D cells; iv) adhesion to TNFalpha activated endothelial cell layer in the “flow chamber” was increased in BCR-positive 32D cells but not in p40(ABL/BCR)- and p96(ABL/BCR)-positive cells. Regarding their leukemogenic potential we showed that i) both ABL/BCRs, in contrast to wt BCR, activated RAS; ii) both ABL/BCRs were unable to transform fibroblasts and to render Ba/F3 cells factor-independent. iii) p96(ABL/BCR) increased the replating efficiency of Sca1+/lin- hemopoietic stem cells (HSC) by selecting a population of immature HSC exclusively expressing c-kit and Sca-1 more strongly than p40(ABL/BCR); iv.) both ABL/BCR blocked the myeloid differentiation of HSC v) the inoculation of p96(ABL/BCR)- or p40(ABL/BCR)-expressing HSC into lethally irradiated recipient mice led in the 40% and 60% of the cases, respectively, to a clinical picture of either acute leukemia or myeloproliferative syndrome within 2–9 months.

These data show for the first time that the t(9;22) leads to two leukemogenic fusion proteins - the BCR/ABL and the ABL/BCR - in CML as well as in Ph+ALL, which might represent an additional target for molecular therapy approaches.

Author notes

Corresponding author