Phd in MOLECULAR MEDICINE

Thesis title: A combination of PARP and CHK1 inhibitors efficiently antagonizes MYCN-driven tumors via an ATM-dependent pathway

MYCN amplification (MNA) is strongly correlated with poor prognosis and treatment failure in patients with neuroblastoma. Despite the intense multimodal therapies, about 50% of these patients succumb to their disease, making the search for effective and less toxic therapies an absolute priority. Activation of oncogenes such as MYCN is associated with high replication stress (RS) and DNA damage, making MYCN-driven tumors addicted to the activity of proteins involved in RS and DNA damage responses. Prompted by the relevant involvement of PARP proteins in controlling RS and DNA repair, we previously demonstrated that olaparib (a pharmacological inhibitor of PARP) enhanced MYCN-induced RS and caused cell death via mitotic catastrophe, in high-risk MYCN amplified (MNA) and overexpressing neuroblastoma cells. Inhibition of the CHK1-dependent S-phase checkpoint raised by olaparib anticipated and increased the occurrence of mitotic catastrophe. Together with the observation that both PARP and CHK1 expression is significantly increased in MNA compared to MYCN single copy (MNSC) neuroblastomas, our data suggested that MYCN-induced RS could make MNA neuroblastoma strongly addicted to PARP and CHK1 activity, providing a rationale to test their combined inhibition for therapeutic purposes. More recent work revealed that olaparib sharply increases the sensitivity of a large panel of neuroblastoma cells to the CHK1 inhibitor MK-8776. The efficacy of this combination is far higher in MNA cell lines than in MYCN single copy (MNSC) cells, in vitro. Indeed, the MK-8776+olaparib combination leads to accumulation of DNA damage and cell death in a synergistic and MYCN-dependent manner, allowing the use of suboptimal doses of MK-8776. Consistently, suboptimal doses of MK-8776 plus olaparib significantly reduce the growth of MNA neuroblastoma subcutaneous and orthotopic xenografts, and the growth of a SHH-MYCN medulloblastoma model, with no major toxicities. IMR-32, the most responsive MNA cell line, bears a putatively pathogenic ATM mutation. It has been reported that genetic loss of homologous recombination repair genes, including ATM, lead to synthetic lethality with PARPi and CHK1i, in the high-risk 11q-deleted neuroblastoma. Therefore, we asked whether ATM is involved in the responses to PARPi+CHK1i combination in IMR-32 cells and, more generally, in MNA neuroblastoma. CRISPR/Cas9-based genetic correction of mutant ATM shows that this mutation does not fully abolish ATM kinase function and is irrelevant for the biological responses to PARPi+CHK1i combination. Rather, ATM knock-out by CRISPR/Cas9 editing, or its pharmacological inhibition, reveals that ATM is required for the proapoptotic effect of PARPi+CHK1i combination, in MNA neuroblastoma cells. This is rather unexpected since literature data and our analysis performed in MNSC backgrounds indicate that loss of ATM function has opposite effects. In conclusions, our data highlight a new potential chemo-free strategy to treat MNA neuroblastoma. Furthermore, our studies on ATM kinase shed light on its involvement in shaping neuroblastoma cell responses to PARPi+CHK1i depending on their genetic background.