Radiosynthesis, in vitro and in vivo evaluation of a 7-(2-[¹⁸F]fluoroethoxy)-6-methoxyquinazoline derivative for imaging PDE10A with PET

Objectives: To improve some of the cognitive symptoms of schizophrenia, a 6,7-dimethoxy-4-pyrrolidinylquinazoline has been designed as a potent and selective brain penetrable inhibitor of phosphodiesterase 10A (PDE10A, K_i,PDE10A = 4 nM [1]). Based on this structure we developed a 7-[¹⁸F]fluoroethoxy-derivative [¹⁸F]I as potential PET radiotracer for imaging PDE10A in brain.
Methods: Non-radioactive reference compounds and precursors were prepared by multi-step syntheses and screened for their PDE10A inhibition as well as selectivity in enzyme activity studies. Radiolabelling of the derivative of highest PDE10A inhibition was initially performed via conversion of 1,3-bistosyloxyethane into [¹⁸F]fluoroethyltosylate using n.c.a. K[¹⁸F]F-K2.2.2-carbonate complex, and subsequently direct etherification of a deprotonated 7-hydroxy-derivative II to [¹⁸F]I (Fig. 1, left-hand). Afterwards, one-step radiosynthesis was developed by direct substitution of a 7-tosyloxy-analogue III with n.c.a. [¹⁸F]fluoride (Fig. 1, right-hand). [¹⁸F]I was purified by SPE and semi-preparative radio-HPLC. Samples were monitored by radio-TLC and -HPLC. Lipophilicity (logD_7.0-7.4) was determined by shake-flask as well as HPLC methods. Further pharmacological characterisation of [¹⁸F]I included in vitro determination of PDE10A affinity (K_D,PDE10A, PDE10A transfected SF21 cells), in vivo biodistribution and brain uptake studies, metabolism and ex vivo brain autoradiography in female CD-1 mice, with validation of target specificity by homologous competition (I, 5 mg/kg at 15 min before radiotracer) and pre-treatment with high PDE10A specific MP-10 (5 mg/kg at 15 min before radiotracer).

Results: Two-step radiosyntheses of [¹⁸F]I resulted in radiochemical yields (RCY) of 18-29% (3.5-4.5 h, based on [¹⁸F]F- aqueous solution) and radiochemical purities (RP) of 92-99%. Improvement was obtained by direct radiofluorination: RCY of 17-40% (3-4 h), RP ≥ 99% and specific activities of 110-1110 GBq/mol. By homologous competition assays a KD,PDE10A of 14 nM was estimated. The logD_7.0-7.4 was determined with ~ 2.6. According to this a sufficiently high initial brain uptake has been observed (2.3%ID/g at 5 min p.i. in striatum). However, radioligand binding in vivo (1.14%ID/g 60 min p.i.) was not inhibited by competition with I (1.3%ID/g) as well as MP-10 (1.4%ID/g). In plasma and brain, respectively, 70% and 96% of the radioactivity detected at 30 and 60 min p.i. corresponded to native radioligand. No evidence for defluorination of the radioligand was obtained.
Conclusions: Convenient radiochemical results, a moderate lipophilicity and a high PDE10A affinity indicate [¹⁸F]I to be a suitable radiotracer. However, further structural optimization is needed to improve the in vivo properties and to make this radioligand appropriate for neuroimaging of PDE10A with PET.
Research Support: Europäischer Fond für regionale Entwicklung (EFRE).
References: [1] Chappie, T.A. et al. (2007), J. Med. Chem., 50, 182-185.