Heteroaryl derivatives of suvorexant as OX1R selective PET ligand candidates: Cu-mediated 18F-fluorination of boroxines, in vitro and initial in vivo evaluation

Background: The orexin receptor (OXR) plays a role in drug addiction and appears as a tumor marker in colon carcinoma. Subtype-selective OXR PET ligands have not yet been reported. The present work deals with the development of 18F-labeled OXR ligands dervived from selective OX1R antagonist JH112. Methods: Applying computational analysis, medicinal chemistry, integrated OXR binding studies, Cu-mediated 18F-fluorination and initial small animal PET studies, we evaluated a series of OXR ligands (1a-1f), varying the heteroarene scaffold of FH112 by 5-fluorobenzoxazole, 5-/6-fluorobenzthiazole and 6-fluoroquinoxaline. Receptor binding studies were preformed on HEK293T cells transiently expressing the OX1R and OX2R. Cu-mediated 18F-fluorination was performed on BPin and boroxine precursors. In vitro assays for logD7.4, stability in plasma, plasma-protein binding and potential P-gp-mediated transport in the Caco-2 monolayer model were performed. The brain uptake of 18F-labeled ligands was studied by dynamic PET imaging in rats. Results: Computational analysis predicted that fluorine substitution (1e) and introduction of the fluorobenzothiazole scaffold (1f) would be suitable for maintaining high OX1R affinity. After multi-step synthesis of 1a-1f, in vitro OXR binding studies confirmed molecular dynamics calculations and revealed single-digit nanomolar OX1R affinities for 1a-f, ranging from 0.69 nM-2.5 nM. The benzothiazole 1f showed high OX1R affinity (Ki = 0.69 nM), along with 77-fold subtype selectivity over OX2R. Cu-mediated 18F-fluorination of the nonselective OXR ligands benzothiazole [18F]1c and quinoxaline [18F]1d yielded RCY of 44% and 56%, respectively, after 10 min, compared to low RCY for benzoxazole [18F]1a of 12% after 20 min. The nonselective OXR ligand [18F]1c and the selective OX1R ligand [18F]1f gave total activity yields of 14% and 22%, respectively, using boroxine precursors for Cu-mediated 18F-fluorination with 5 min reaction time and a total synthesis time of 50-60 min. [18F]1c and [18F]1f were stable in plasma and serum in vitro, with logD7.4 of 2.28 ([18F]1c) and 2.37 ([18F]1f), and high plasma-protein binding of 66% and 77%, respectively, reflecting only marginal differences in brain uptake of [18F]1c (0.17 %ID/g) and [18F]1f (0.15 %ID/g). 1c showed higher passive permeability than 1f in vitro, which was consistent with the faster brain clearance of [18F]1c compared to [18F]1f. However, preinjection of the OXR antagonist suvorexant did not significantly block [18F]1c or [18F]1f uptake in the rat brain. Pretreatment with cyclosporin A to study the role of P-gp in limiting brain accumulation moderately increased brain uptake of [18F]1c and [18F]1f. Accordingly, in vitro experiments demonstrated that the P-gp inhibitor zosuquidar only moderately inhibited polarized, basal to apical transport of 1c (p<0.05) and had no influence on transport of 1f (n.s.), indicating that P-gp does not play a relevant role in brain accumulation of [18F]1c and [18F]1f in vivo. Conclusions: The in vitro and in vivo results of [18F]1c and [18F]1f provide a solid basis for further development of suitable OXR PET ligands for brain imaging. This could be achieved through structural changes that would result in lower blood protein binding and 18F labeling methods that would allow for improved molar activities.