RAD51 recruitment but not replication fork stability associates with PARP inhibitor response in ovarian cancer patient-derived xenograft models.

Abstract

Poly(ADP-ribose) polymerase (PARP) inhibitors (PARPis) are currently used to treat BRCA1/2 mutant cancers. Although PARPi sensitivity has been attributed to homologous recombination (HR) defects, other roles of HR factors have also been linked to response to PARPi, including replication fork protection. In this study, we investigated PARPi sensitivity in ovarian cancer patient-derived xenograft (PDX) models in relation to HR proficiency and replication fork protection. Analysis of BRCA1/2 status showed that in our cohort of 31 ovarian cancer PDX models 22.6% harbored a BRCA1/2 alteration (7/31), and 48.3% (15/31) were genomically unstable as measured by copy number alteration analysis. In vivo, PARPi olaparib response was measured in 15 selected PDX models. Functional assessment of HR using ex vivo irradiation-induced RAD51 foci formation identified all olaparib-sensitive PDX models, including four models without BRCA1/2 alterations. In contrast, replication fork protection or replication speed in ex vivo tumor tissue did not correlate with olaparib response. Targeted panel sequencing in olaparib-sensitive models lacking BRCA1/2 alterations revealed a MUS81 variant as a possible mechanism underlying PARPi sensitivity. Combined, we show that ex vivo RAD51 analysis effectively predicts in vivo olaparib response and revealed a subset of PARPi-sensitive, HR-deficient ovarian cancer PDX models, lacking a BRCA1/2 alteration.