Targeting of Quiescent and Proliferating CML Stem Cells By DNA Repair Inhibitors

Tyrosine kinase inhibitors (TKIs) such as imatinib, dasatinib and nilotinib revolutionized the treatment of chronic myeloid leukemia in chronic phase (CML-CP). However, it is unlikely that TKIs will "cure" a majority of CML patients due to the presence of TKI-refractory quiescent leukemia stem cells (LSCs) and/or TKI-resistant proliferating LSCs. Moreover, a sub-population of patients does not respond favorably to TKI therapy and/or accumulates TKI-resistant BCR-ABL1 mutants and additional chromosomal aberrations (high risk/poor responders = HR/PR). This process may lead to the disease relapse and/or malignant progression to fatal chronic myeloid leukemia in accelerated phase or blast phase (CML-AP/BP). Therefore new treatment modalities are necessary to improve therapeutic outcome of CML HR/PRs.

We found that CML LSCs and LPCs, including quiescent LSCs, from HR/PRs accumulate the highest levels of reactive oxygen species (ROS)-induced DNA double strand breaks (DSBs) in comparison to "good responders" and normal counterparts (Bolton-Gillespie et al., Blood, 2013). DSBs are the most lethal DNA lesions, but CML cells can tolerate them because two major repair mechanisms, homologous recombination (HR) and non-homologous end-joining (NHEJ), are hyper-activated (Slupianek et al., Mol. Cell, 2001; Oncogene, 2005; DNA Repair, 2006; Cancer Res., 2011; Blood, 2011; Nowicki et al., Blood, 2005). It appears that CML cells, and especially these from HR/PRs are "addicted" to DSB repair pathways to survive pro-apoptotic challenge from lethal DSBs.

There are critical differences between DSB repair in normal and CML cells due to downregulation of BRCA1 in HR pathway and DNA-Pkcs in NHEJ pathway (Slupianek et al., Cancer Res., 2011; Podszywalow-Bartnicka et al., Cell Cycle, 2014). Therefore, proliferating LSCs/LPCs employ RAD52-dependent alternative HR, in contrast to BRCA1-mediated HRR in normal counterparts. Quiescent LSCs use PARP1-mediated alternative NHEJ instead of DNA-PKcs -dependent NHEJ, which is predominant in normal quiescent HSCs. We hypothesized that simultaneous targeting of PARP1 and RAD52 will trigger "dual synthetic lethality" eradicating quiescent LSCs and proliferating LSCs/LPCs.

To test this hypothesis we generated Parp1-/-Rad52-/- double knockout mice, which did not display any detectable problems in hematopoietic system and other organs. Leukemogenesis in SCLtTA/p210BCR-ABL1/Parp1-/-Rad52-/- mice was much prolonged in comparison to SCLtTA/p210BCR-ABL1/Parp1-/-, SCLtTA/p210BCR-ABL1/Rad52-/- and SCLtTA/p210BCR-ABL1 animals; 33% of SCLtTA/p210BCR-ABL1/Parp1-/-Rad52-/- mice did not develop leukemia. Moreover, BCR-ABL1 transfected Parp1-/-Rad52-/- bone marrow cells formed the lowest number of colonies when compared to BCR-ABL1 transfected Parp1-/-, Rad52-/-, and wild-type cells.

Next, we tested anti-leukemia effect of a combination of PARP1 inhibitor (FDA approved Lynparza) and RAD52 small molecule inhibitor (RAD52smi) identified by high-throughput screen. Simultaneous administration of RAD52 and PARP1 inhibitors completely eradicated clonogenic activity of TKI-treated CML-CP HR/PRs and CML-AP LSCs/LPCs and almost exhausted quiescent LSCs without affecting normal cells. These inhibitors were also very effective against Ph+ ALL cells.

In conclusion "dual synthetic lethality" triggered by simultaneous targeting of PARP1 and RAD52 represents a novel strategy which may eradicate quiescent LSCs and proliferating LSCs/LPCs not only from CML but also from BCR-ABL1 -positive ALL.