Predicting PARP inhibitor sensitivity and resistance

This article refers to:

Evaluation of candidate biomarkers to predict cancer cell sensitivity or resistance to PARP-1 inhibitor treatment

The concept of precision therapy in cancer implies an ability to customize treatment according to the molecular makeup of the patient and the tumor. While many pharmaceutical agents are described as targeted therapies, their precise role in cancer treatment is unclear. Often the new drug’s target is well-described, but the precise indication of when it should be used (i.e., a prospectively validated predictive marker) is often not defined. PARP inhibitors exemplify this paradigm well. The molecular target, poly ADP-ribose polymerase 1 (PARP1), and the interaction between drug and target are well-defined, and small-molecule inhibitors of PARP1 (PARP1i) have shown considerable promise for the treatment of BRCA1- and BRCA2-deficient breast, ovarian and prostate cancer.¹ Recent work suggests PARP1i may also have utility in the treatment of cancers with deficiencies in other DNA damage response proteins,² including ATM^3-5 and Mre11.⁶ Nevertheless, experience in triple-negative breast cancer, a limited surrogate for DNA repair deficiency, has tempered enthusiasm for their use, emphasizing that the identification of molecular predictors of PARP1i sensitivity and resistance to guide their use remains a priority. Here, Bartek and colleagues attempt to address this issue directly, describing several key observations.⁷

First, they extend prior findings to show that depletion of either Nbs1 or Mre11, components of the MRN complex required for activation of ATM and initiation of the homologous recombination DNA double-strand break repair pathway, also sensitize breast cancer cells to PARP1i. However, unlike in cells with depletion or inactivation of ATM, where mutation or depletion of p53 enhances sensitivity to the PARP1i olaparib,⁴ p53 status had little effect on PARP1i sensitivity in colon cancer cells depleted of Mre11 and/or Nbs1, either alone or in combination with DNA damaging agents, suggesting that Mre11 or Nbs1 deficiency has subtly different effects on PARP1i sensitivity compared with ATM deficiency. Second, the authors show that depletion of 53BP1 promotes PARP1i resistance, consistent with active NHEJ being required for PARP1i-induced cell death.^4,8

The successful clinical application of PARP inhibitors will require identification of predictive markers for tumors likely to be susceptible to PARP inhibition. Oplustilova and colleagues⁷ provide evidence that poly-ADP ribose, the product of PARP activity, could be a potential biomarker of cellular response to PARP inhibition. Finally, the authors address acquired resistance to PARP1i, an evolving problem clinically, by modulation of P-glycoprotein (P-gp). The authors show that the P-gp inhibitor Verapamil overcomes PARP1i efflux attributed to P-gp overexpression, thereby restoring sensitivity to PARP1i.⁷

Despite addressing these important questions, do these findings herald the arrival of a predictive marker for PARP1i? Unfortunately, the answer is probably not. Although these findings extend our understanding of the potential application of PARP1i to additional DNA repair-deficient cancers, there are several important caveats. The work described was performed on a range of different cell lines in vitro and has not been validated in in vivo models. In addition, the incomplete understanding of the influence of p53 on PARP1i in the context DNA repair deficiency limits its status as a predictor of sensitivity. Finally, overcoming resistance using P-gp inhibitors has had a disappointing course in modulating the efficacy of standard cytotoxic chemotherapy. Unless the third generation P-gp inhibitors are markedly more effective and less toxic, it seems unlikely that this is the best avenue to pursue in improving PARPi utility. As exemplified in HER2-positive breast cancer and in EGFR mutation-positive non-small cell lung cancer, it is a consistent finding that “molecular selection trumps clinical selection.”⁹ A validated predictive molecular marker can direct targeted therapy, ensuring that patients unlikely to respond are offered more useful intervention, and restrict expensive treatment to those most likely to benefit. While the work of Bartek et al.⁷ extends our understanding of the role of DNA repair deficiency in PARPi sensitivity, it seems that the subtle nuances of DNA repair pathway modulation make it challenging to develop such a marker for PARP inhibition. In the absence of a functional assay to test tumor DNA repair capacity, more work is required to precisely define which tumor molecular marker (or combination of markers) will best serve this purpose.