Pharmacokinetics, Safety, And Tolerability Of Olaparib And Temozolomide For Recurrent Glioblastoma: Results Of The Phase I Oparatic Trial

Background. The poly(ADP-ribose) polymerase (PARP) inhibitor olaparib potentiated radiation and temozolomide (TMZ) chemotherapy in preclinical glioblastoma models but brain penetration was poor. Clinically, PARP inhibitors exacerbate the hematological side effects of TMZ. The OPARATIC trial was conducted to measure penetration of recurrent glioblastoma by olaparib and assess the safety and tolerability of its combination with TMZ.Methods. Preclinical pharmacokinetic studies evaluated olaparib tissue distribution in rats and tumor-bearing mice. Adult patients with recurrent glioblastoma received various doses and schedules of olaparib and low-dose TMZ in a 3 + 3 design. Suitable patients received olaparib prior to neurosurgical resection; olaparib concentrations in plasma, tumor core and tumor margin specimens were measured by mass spectrometry. A dose expansion cohort tested tolerability and efficacy of the recommended phase II dose (RP2D). Radiosensitizing effects of olaparib were measured by clonogenic survival in glioblastoma cell lines.Results. Olaparib was a substrate for multidrug resistance protein 1 and showed no brain penetration in rats but was detected in orthotopic glioblastoma xenografts. Clinically, olaparib was detected in 71/71 tumor core specimens (27 patients; median, 496 nM) and 21/21 tumor margin specimens (9 patients; median, 512.3 nM). Olaparib exacerbated TMZ-related hematological toxicity, necessitating intermittent dosing. RP2D was olaparib 150 mg (3 days/week) with TMZ 75 mg/m2 daily for 42 days. Fourteen (36%) of 39 evaluable patients were progression free at 6 months. Olaparib radiosensitized 6 glioblastoma cell lines at clinically relevant concentrations of 100 and 500 nM.Conclusion. Olaparib reliably penetrates recurrent glioblastoma at radiosensitizing concentrations, supporting further clinical development and highlighting the need for better preclinical models.