Easily automated radiosynthesis of [18F]P10A-1910 and its clinical translation to quantify phosphodiesterase 10A in human brain

Our previous work showed that [18F]P10A-1910 was a potential radioligand for use in imaging phosphodiesterase 10A (PDE10A). Specifically, it had high brain penetration and specific binding that was demonstrated in both rodents and non-human primates. Here, we present the first automatic cGMP-level production of [18F]P10A-1910 and translational PET/MRI study in living human brains. Successful one-step radiolabeling of [18F]P10A-1910 on a GE TRACERlab FX2N synthesis module was realized via two different methods. First, formulated [18F]P10A-1910 was derived from heating spirocyclic iodonium ylide in a tetra-n-butyl ammonium methanesulfonate solution. At the end of synthesis, it was obtained in non-decay corrected radiochemical yields (n.d.c. RCYs) of 12.4 ± 1.3%, with molar activities (MAs) of 90.3 ± 12.6 μmol (n = 7) (Method I). The boronic pinacol ester combined with copper and oxygen also delivered the radioligand with 16.8 ± 1.0% n. d.c. RCYs and 77.3 ± 20.7 GBq/μmol (n = 7) MAs after formulation (Method II). The radiochemical purity, radionuclidic purity, solvent residue, sterility, endotoxin content and other parameters were all validated for human use. Consistent with the distribution of PDE10A in the brain, escalating uptake of [18F]P10A-1910 was observed in the order of cerebellum (reference region), substantial nigra, caudate and putamen. The non-displaceable binding potential (BPND) was estimated by simplified reference-tissue model (SRTM); linear regressions demonstrated that BPND was well correlated with the most widely used semiquantitative parameter SUV. The strongest correlation was observed with SUV(50–60 min) (R2 = 0.966, p < 0.01). Collectively, these results indicated that a static scan protocol could be easily performed for PET imaging of PDE10A. Most importantly, that [18F]P10A-1910 is a promising radioligand to clinically quantify PDE10A.