Oncogene-Induced DNA Repair Defects Promote PARP1-Mediated “Dual Synthetic Lethality” To Eradicate Quiescent and Proliferating Leukemia Stem and Progenitor Cells

Leukemia stem cells (LSCs), including quiescent cells, are disease initiating and therapy-refractory cells. Therefore, even if a treatment clears a disease burden consisting mostly of leukemia progenitor cells (LPCs), it usually fails to eradicate LSCs and residual LPCs which developed therapy-resistance. Leukemia cells expressing BCR-ABL1 (CML, Ph+ALL), FLT3(ITD) (AML), and AML1-ETO (AML) accumulate high numbers of spontaneous DNA double strand breaks (DSBs). Leukemia cells can survive numerous DSB, the most lethal DNA lesions, due to enhanced DSB repair activity by homologous recombination (HR; active in S and G2 cell cycle phase) and/or non-homologous end-joining (NHEJ; active in G0/G1 and S cell cycle phase). Altogether, leukemia cells may be “addicted” to DSB repair pathways.

Normal cells usually employ BRCA1/2-RAD51 -mediated HRR and Ku70/86-DNA-PKcs -dependent NHEJ. BCR-ABL1 causes BRCA1 and DNA-PKcs deficiency, FLT3(ITD) inhibits Ku86, and AML1-ETO downregulates DNA-PKcs and RAD51. Accordingly, leukemia cells expressing these oncogenes are forced to employ alternative DSB repair pathways, such as PARP-LigIII –mediated NHEJ. Since NHEJ plays a predominant role in quiescent cells and also supports HR in proliferating cells, we postulated that targeting PARP should exert “dual synthetic lethality” to eradicate quiescent LSCs and proliferating LSCs/LPCs, with negligible effect on normal cells.

We showed that PARP inhibitor olaparib, which is in clinical trials for the treatment of solid tumors displaying BRCA1/2 mutations, abrogated NHEJ activity in BCR-ABL1 cells. Olaparib reduced the number of imatinib-naïve and imatinib-treated Lin^-CD34⁺CD38^-CTV^max quiescent CML-CP LSCs in hypoxia and normoxia mimicking bone marrow niche and arterial peripheral blood, respectively. In addition, olaparib enhanced the anti-proliferative effect of imatinib in Lin^-CD34⁺CML-CP and CML-AP LPCs. These effects are probably due to imatinib-mediated inhibition of BCR-ABL1 kinase-dependent anti-apoptotic activity and olaparib-induced increase of the number of lethal DSBs, resulting in accumulation of annexin V-positive apoptotic quiescent LSCs and hyper-activation of caspase-3 in imatinib+olaparib treated LPCs. Inhibition of PARP almost completely abrogates DSB repair in DNA-PKcs-deficient quiescent LSCs and diminishes resolution of ROS-induced stalled replication forks in BRCA1-deficient proliferating LSCs/LPCs. The combination of imatinib+olaparib did not affect normal quiescent hematopoietic stem cells, but exerted modest inhibitory effect on normal proliferating progenitors.

Since olaparib does not discriminate neither between PARP family members nor other enzymatic pathways involving NAD⁺, its long-term application may generate side-effects. Our genetic studies involving Parp1-/- mice and PARP1(E988K) catalytic-deficient mutant identified PARP1 as major player in NHEJ in BCR-ABL1 cells. Using high throughput screening we identified 5F2, a small molecule which abrogated histone 4-dependent PARP1 activation and exerted synthetic lethality in BRCA1-deficient, but not BRCA1-proficient carcinoma cells. 5F2, similarly to olaparib, reduced the number of imatinib-naïve and imatinib-treated Lin^-CD34⁺CD38^-CTV^max quiescent LSCs and inhibited colony formation by Lin^-CD34⁺LPCs. However 5F2, in contrast to olaparib, did not affect normal cells.

In addition, olaparib and/or 5F2 reduced the number of imatinib-naïve and imatinib-treated Ph+ALL cells harvested from patients at diagnosis. Moreover, PARP1 inhibitors exerted anti-leukemia effect against ponatinib-naïve and ponatinib-treated Ph+ALL cells carrying BCR-ABL1 T315I mutation, and against lestaurtinib/quizartinib-naïve and lestaurtinib/quizartinib-treated FLT3(ITD)-positive AML cells. PARP1 inhibitors also abrogated the growth of leukemia cells expressing AML1-ETO (AML), but not of these expressing PML-RAR (APL) or overexpressing HOXA9+MEIS1 (AML).

In conclusion, targeting PARP1 resulted in the induction of “dual synthetic lethality” and eradication of quiescent and proliferating CML cells displaying specific defects in DSB repair pathways. Similar effect is induced in other leukemias carrying specific, oncogene-induced DSB repair deficiencies. PARP1 inhibitors are currently tested in vivo using primary leukemia xenografts.