Abstract
727
Objectives: Diffuse Intrinsic Pontine Glioma (DIPG) is an inoperable, treatment resistant pediatric brainstem tumor with a mean survival of less than one year. Recently, the DNA repair enzyme Poly(ADP-ribose)Polymerase 1 (PARP1) has been identified to be frequently overexpressed in DIPG and could provide a novel approach for image-guided characterization and treatment of the disease1,2. Here, we aim to show feasibility of DIPG imaging in genetically engineered mouse models (GEMM) using PARP1 targeted imaging agents, labeled with either a fluorescent or isotopic moiety to enable imaging from the cellular to the whole body level. Specifically, we aim to show accurate delineation of tumor margins and monitoring of tumor progression in vivo.
Methods: We assessed PARP1 expression in a set of human DIPG samples to confirm relevance of the target (n=5). To image PARP1, we used a fluorescent (“PARPi-FL”) and a 18F-labelled (“[18F]PARPi”) version of the PARP1 Inhibitor Olaparib, which display comparable affinity and specificity to nuclear PARP13-5. To generate DIPG like tumors, retroviral vector expressing DF1 cells (RCAS-PDFGB, RCAS-Cre), were injected into the brains of ntv-a/p53fl/fl mice. We examined PARP1 expression of the GEMM using immunohistochemistry (IHC) and compared it to the human samples. PARPi-FL accumulation in tumor bearing brains was observed on the cellular level using confocal microscopy after intravenous injection and compared to PARP1 expression. The ability of [18F]PARPi to delineate tumors was evaluated using PET/CT and autoradiography and compared to MRI, pathology and PARP1 expression (n=5). Specificity was determined using a blocking approach (pre-injection of Olaparib) (n=4/group). Tumor growth was characterized by weekly [18F]PARPi PET/CT imaging (n=5). Furthermore, we compared the imaging accuracy and contrast of [18F]PARPi imaging to the PET brain tracers [11C]Choline and [18F]FLT (n=5).
Results: The examined DIPG biopsy tissues showed PARP1 expression in >90% of cell nuclei, while normal brain tissue showed low PARP1 expression. The DIPG GEMM we used showed a similar PARP1 overexpression in the diffusely growing tumors, confirmed by H&E and Ki67 staining. PARPi-FL showed penetration and specific accumulation in tumor cell nuclei after i.v. injection, confirming the ability to delineate tumors. Autoradiographic images of [18F]PARPi showed clear tumor to background contrast in tumors as small as 1.5 mm. In vivo whole body PET/CT imaging revealed accumulation of [18F]PARPi in MRI confirmed tumors and histology (H&E, PARP1) revealed accuracy of tumor delineation. Pre-injection of Olaparib reduced the uptake to the % ID/g to the level of healthy mice, confirming specificity (p<0.05). We were able to monitor tumor growth by conducting weekly PET/CT imaging from 3-6 weeks post inoculation. Uptake of [18F]PARPi was in agreement with the DNA synthesis/metabolism PET tracers [11C]Choline and [18F]FLT, but at the same time can provide information on PARP1 expression to guide treatment decisions.
Conclusion: Our data show successful PARP1 imaging in a murine DIPG model on the whole body ([18F]PARPi) and cellular (PARPi-FL) level. This method may serve as novel method to longitudinally monitor disease progression and therapeutic response in preclinical studies and could ultimately help an informed patient selection for PARP1 Inhibitor treatment or the delivery of therapeutic radionuclides via PARP1 targeted molecules. Research Support: We thank the Molecular Cytology Core at MSKCC (P30 CA008748), the NIH (K25 EB016673), the Brain Tumor Center of MSK, the Imaging and Radiation Sciences Program, the German Research Foundation (DFG) and the Emily Tow Foundation for their generous funding.