We showed before that cells transformed by BCR/ABL and other fusion tyrosine kinases (FTKs) such as TEL/ABL, TEL/JAK2 and TEL/PDGFR, inducing chronic myeloproliferative disorders (MPDs), and CD34+ chronic myeloid leukemia (CML) stem/ progenitor cells from chronic phase (CML-CP) and blast crisis (CML-BC) contain an excess of DNA double-strand breaks (DSBs) induced by reactive oxygen species (ROS) and genotoxic stress [Blood, 2005; Cell Cycle, 2006; DNA Repair, 2006; Cancer Res., 2008]. Recent studies also revealed that CD34+CD38− CML-CP and CML-BC stem cellenriched populations seem to display more DSBs than normal counterparts as measured by gamma-H2AX foci formation on DNA. Elevated levels of DSBs were also observed in leukemia cells expressing imatinib-resistant BCR/ABL kinase mutants. DSBs may cause apoptosis if not repaired or chromosomal aberrations if repaired unfaithfully. Numerous ROS- and radiation- induced DSBs are not lethal for BCR/ABL-positive leukemia cells; instead, they induce chromosomal instability implicating enhanced, but unfaithful repair [Leukemia, 2008]. The previous report [Mol. Cell, 2001] and ongoing studies demonstrated that BCR/ABL kinase (non-mutated and imatinib-resistant mutants) modulates expression of the mammalian RecA homologs RAD51, RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3, which are responsible for homologous recombination repair (HRR) of DSBs. RAD51 plays a key role in HRR in cells transformed by BCR/ ABL and other FTKs [Mol. Cell, 2001; Mol. Cell. Biol., 2002]. BCR/ABL stimulates the expression of, interacts with and phosphorylates RAD51 on Y315, which is located in a critical fragment of RAD51 essential for its filament formation on DNA. Accordingly, our recent results indicated that BCR/ABL-mediated RAD51[Y315] phosphorylation appears to be important for nuclear RAD51 foci formation in response to DNA damage. In addition to RAD51, BCR/ABL interacts directly with and phosphorylates RAD51B and XRCC2, but not other RecA homologs. Altogether, it appears that BCR/ABL can deregulate the expression and phosphorylation of some RecA homologs, which may have a significant impact on the efficiency and fidelity of DSB repair resulting in protection from apoptosis and chromosomal instability. Therefore, disassembly of BCR/ABL from RecA homologs should reduce the capability of CML cells to repair numerous ROS-induced DSBs and eventually trigger apoptosis. Based on this hypothesis we investigated the mechanisms of association between BCR/ABL and RecA homologs. Interactions between BCR/ ABL and RAD51 or RAD51B depend on the proline- rich (PP) regions of RAD51 and RAD51B, and the SH3 domain and SH2-catalytic domain (SH2-CD) linker of BCR/ABL, which form a pocket binding the PP regions. Disruption of the PP regions of RAD51 by P-L amino acid substitutions (PP-LL mutants) abrogated direct interaction with the BCR/ABL SH3-SH2-CD pocket. On the other hand, single amino acid substitutions in the BCR/ABL SH3-SH2-CD pocket, which eliminated its capability of binding the PP regions, prevented complex formation with RAD51 and RAD51B. In addition, RAD51 and RAD51B may interact with members of the BCR/ABL proteome such as Grb2 and Shc (RAD51 and RAD51B), and c-CrkL (only RAD51B), but not Gab2 and c-Cbl. 32Dcl3 murine hematopoietic cells expressing BCR/ABL SH3-SH2-CD pocket mutant, where single amino acid substitutions disrupted its direct interaction with RAD51, displayed a slower proliferation rate and responded poorly to genotoxic stress despite intact kinase activity in comparison to cells transformed with non-mutated BCR/ABL. Interestingly, expression of RAD51 PP-LL mutant eliminated BCR/ABL-transformed leukemia cells, without any toxic effect on normal counterparts. These results suggest that the interaction between BCR/ABL and RAD51 may be targeted for selective elimination of leukemia cells and/or suppression of genomic instability. To test this hypothesis we are employing the peptide aptamer strategy targeting RAD51 PP regions in CD34+ cells obtained from imatinib-sensitive and imatinib-resistant CML patients and healthy volunteers in vivo and in vitro. In summary, we hypothesize that mechanisms regulating the association of BCR/ABL with RAD51 and other mammalian RecA homologs may be explored for the planning of more effective anti-tumor modalities.

Disclosures: No relevant conflicts of interest to declare.