PET Imaging of PARP Expression Using 18F-Olaparib

Abstract

Poly(adenosine diphosphate ribose)polymerases (PARP) are an abundant family of enzymes with multiple cellular functions. PARP1, the best-studied of this family, is essential for the repair of single-strand DNA breaks through base excision repair (1). In the presence of defects in the homologous recombination DNA repair pathway (HRD), PARP1 also plays a critical role in maintaining the DNA repair machinery because the loss of PARP1 increases the number of DNA lesions normally repaired by HR (2). The breast cancer susceptibility genes (BRCA1/2) are central regulators of the HRD repair pathway (3,4). Tumors with loss-of-function BRCA mutations exhibit high levels of genomic instability, thus making them dependent on PARP1 for survival. The therapeutic potential in exploiting this dependency was first illustrated in 2 key studies demonstrating that PARP inhibition in BRCA mutant cancers leads to synthetic lethality (5,6). This was confirmed in initial clinical trials by marked clinical responses to PARP inhibition in patients with treatment-resistant BRCA mutant cancers (7,8), sparking significant enthusiasm for this therapeutic approach. Positive results in BRCA-mutant ovarian cancer led to the Food and Drug Administration approval of olaparib (9), for which testing was required to demonstrate the presence of a BRCA mutation. Since then, it has become clear that BRCA mutation status alone does not predict which patients will respond, as seen in an early study with olaparib, which showed objective responses in 24% of patients with ovarian cancer without a BRCA mutation (10). However, HRD markers may improve on patient selection. Niraparib and rucaparib were approved without the need for BRCA testing because phase 3 trials showed improved progression-free survival in patients either with a germline BRCA mutation or with tumors exhibiting HRD but without a germline BRCA mutation (11,12). Similar results were observed in prostate cancer patients; those with genetic mutations predicting the presence of any DNA repair defect had higher response rates to olaparib (13). Olaparib has since gained expansion of the initial approved indication for maintenance therapy without needing BRCA testing (14). Despite these advances in identifying patients who will respond, HRD markers still do not perfectly identify all responders to PARP inhibitors, nor do they predict the development of treatmentresistance mechanisms. The increase in progression-free survival in ovarian cancer patients without germline BRCA and positive HRD markers was consistently less than that seen in patients with germline BRCA in the niraparib and rucaparib trials, demonstrating the limitations of the HRD markers for predicting PARP inhibitor response. Additionally, treatment resistance can be present because of DNA replication fork stabilization, regardless of HR status, or may develop because of somatic reversion mutations that restore HR functionality during PARP inhibitor treatment (15–18). The only way to detect these alterations, whether before starting PARP inhibitor therapy or when assessing causes of treatment failure, is to obtain tumor tissue for testing. Presumably, these functional alterations affect PARP expression levels, but the study of this effect has been limited by the requirement of tissue for such assessments. Imaging PARP expression levels may be a novel approach for studying the relationship of PARP expression and PARP inhibitor efficacy in the context of HRD and functional alterations that may confer treatment resistance. The results of such studies could be used to identify rational therapy combinations with PARP inhibitors to maximize the therapeutic benefits while minimizing exposure to toxicities in patients who would not respond. Studies have shown that PARP inhibitor efficacy in vitro requires the presence of PARP1 because PARP inhibitors are ineffective at killing PARP1 knockout cells (19,20). Therefore, radiolabeled PARP inhibitors could be used to verify the presence and extent of tumoral PARP1 expression noninvasively in patients. Radiolabeled PARP inhibitors could further be used to characterize dynamic changes in tumoral PARP expression during treatment with PARP inhibitors or DNA-damaging agents that could then be related to the presence of HRD markers or other markers of treatment resistance. Human studies performed with 18F-fluorthanatrace demonstrate the potential for PARP imaging to detect differences in tumoral PARP expression in clinical studies. First-in-human studies demonstrated high 18F-fluorthanatrace binding in a single-patient example of biphenotypic hepatocellular carcinoma and cholangiocarcinoma (21). A subsequent trial in ovarian cancer patients demonstrated heterogeneously increased 18F-fluorthanatrace uptake that differed from 18F-FDG uptake (22). This study further demonstrated that 18F-fluorthanatrace uptake but not 18F-FDG uptake correlated with PARP levels assessed by autoradiography and immunohistochemical staining (22). The results from these initial studies indicate that radiolabeled PARP inhibitors promise to be highly useful in measuring changes in tumoral PARP expression levels during treatment. The relationship between PARP inhibitor binding in vivo and treatment efficacy could also be studied using radiolabeled PARP imaging agents, thus providing insights into dosing approaches for PARP inhibitor therapy. All PARP inhibitors in clinical use or under evaluation to date bind to the PARP binding site for Received Oct. 7, 2018; revision accepted Dec. 14, 2018. For correspondence or reprints contact: Delphine L. Chen, Washington University School of Medicine, Campus Box 8225, 510 S. Kingshighway Blvd., St. Louis, MO 63110. E-mail: [email protected] Published online Jan. 10, 2019. COPYRIGHT© 2019 by the Society of Nuclear Medicine and Molecular Imaging. DOI: 10.2967/jnumed.118.219733