Synthesis and radiopharmacological evaluation of a novel 18F-labeled cyclooxygenase-2 inhibitor based on dihydropyrrolo[3,2,1-hi]indole core structure

Objectives
Cyclooxygenase-2 (COX-2), a key player in inflammation, is an attractive target for functional characterization of solid tumors by PET because its overexpression has been associated with chemo-/radioresistance and poor prognosis in cancer. We recently developed a novel series of selective COX-2 inhibitors based on a tricyclic core structure with IC50 values in the nanomolar range1 and herein report on the 18F-labeling and evaluation of a promising candidate.
Methods
5-(4-[18F]Fluorophenyl)-4-[4-(methylsulfonyl)phenyl]-1,2-dihydropyrrolo[3,2,1-hi]indole ([18F]2) was synthesized according to our recently reported 18F-fluorination and McMurry cyclization approach2 with modifications reported herein. [18F]2 was evaluated in vitro in cell lines with different COX-1/COX-2 expression patterns. In vivo dynamic small animal PET imaging and biodistribution studies were performed in NMRI nu/nu mice bearing a COX-2-positive A2058 tumor-xenograft.
Results
18F-Fluorination under standard conditions2 was hampered by basic hydrolysis leading primarily to side product [18F]1b. Optimization experiments focused on the use of different bases with varying concentrations (K2CO3, KHCO3, KH2PO4) or no base using the ‘minimalist approach’3. As one result, the use of decreased amounts of K2CO3 (5 instead of 20 μmol) effectively suppressed hydrolysis and gave [18F]1a in high yield (Figure 1). An automated synthesis comprising mild 18F-fluorination, McMurry cyclization, and purification using a TracerLabFX-N module provided [18F]2 in 16% isolated RCY (d.c.) with a molar activity of 45-106 GBq/μmol at EOS. A LogDpH7.4 of 4.66 and a CHI IAM value of 48 indicated high lipophilicity and non-specific binding. Cell uptake was independent of COX-2 expression. Biodistribution and PET studies revealed highest uptake of [18F]2 in liver and adipose tissue but only low accumulation in A2058 tumors (tumor/muscle < 1) at 60 min post injection. Celecoxib pre-injection (20 mg/kg) did not significantly change tumor uptake although a trend towards decreased radiotracer uptake was observed by PET in a subset of mice.
Conclusions
Despite of a high COX-2 selectivity and metabolic stability, [18F]2 did not emerge as suitable radiotracer for imaging COX-2 in vitro and in vivo, likely due to its high lipophilicity and fast hepatobiliary excretion.4 Future efforts for the development of COX-2-targeted radiotracers should focus on adaption of lipophilicity and/or use of targeted delivery systems.
References
1Laube et al. J. Org. Chem. 2015, 80, 5611-5624. 2Kniess et al. Biorg. Med. Chem. 2012, 20, 3410-3421. 3Richarz et al. Org. Biomol. Chem. 2014, 12, 8094-8099. 4Gassner et al. ChemistrySelect, 2016, 1, 5812–5820.
Figure 1.: Radiosynthesis of [18F]2 by 18F-fluorination and McMurry cyclization. S156: Poster 22nd International Symposium on Radiopharmaceutical Sciences
J