Introduction
The PET radioligand [C-11]FLB-457 was developed in order to study extrastriatal tissues where the D2/D3 receptor densities are 1–2 orders of magnitude lower than in the striatum. Most PET/[C-11]FLB-457 studies have been quantified using the simplified reference tissue model (SRTM), assuming that the D2/D3 receptor density in the cerebellum is negligible. The present study investigated the effects of specific binding in the reference region on estimates of radioligand affinity (KD), regional binding potential (BP) and receptor occupancy.
Methods
Ex vivo saturation studies were conducted in 17 Sprague-Dawley rats by either varying the dose of [C-11]FLB-457 injected or pre-dosing with stable ligand. The in vivo affinity was estimated from the single-site binding model that included a parameter for non-specific uptake (Bns), and from Scatchard analyses using either Bns or cerebellar uptake as an estimate of free ligand concentration. Ten male volunteers underwent a 90-min dynamic PET study in the ECAT/EXACT3D tomograph after a bolus injection of [C-11]FLB-457. The volume of distribution (VD) was estimated from ROI-derived time-activity curves (thalamus, hippocampus, frontal cortex, cerebellum) using a two-tissue compartment four-rate constant model with a metabolite-corrected plasma input function. BP was calculated as VD ratio minus 1, assuming that the cerebellum is a suitable reference region. The apparent BP and occupancy (relative reduction in BP) were estimated using equations derived to account for the presence of specific binding in the reference region and compared to values estimated without specific binding.
Results
In rat brain, radioligand binding decreased with increasing dose of stable ligand in all tissues, including the cerebellum. From the single-site binding model parameters, the cerebellar BP(=Bmax/Bns) is estimated to be ∼1.5. The in vivo KD is overestimated by a factor of ∼5, if the cerebellum is used as an estimate of non-specific uptake. Using D2/D3 receptor density (Bmax) measurements in postmortem human brains, the striatal BP is underestimated by ∼5% for the low affinity radioligand [C-11]raclopride (KD∼1 nM) compared to ∼50% for the high affinity radioligand [C-11]FLB-457 (KD∼0.02 nM). The cortical BP is underestimated by >50% for both radioligands, indicating that the D2/D3 receptor density in the cerebellum is too high relative to that in cortical regions for use as a reference region, regardless of the affinity of the radioligand. As the cerebellar BP is simulated to increase from 0.1 to 3, the occupancy is increasingly underestimated, but uniformly across regions of different receptor densities. However, if the occupancy is lower in the reference than in the target regions, low density regions appear to have a higher occupancy than high density regions.
Conclusion
Specific binding was estimated to account for more than half of the cerebellar uptake of [C-11]FLB-457, invalidating its use as an estimate of non-specific binding. The in vivo affinity of [C-11]FLB-457 may be even higher than previously estimated, thereby increasing the specific activity required in order to avoid the mass effect of co-injected stable FLB-457. Measurements of regional receptor occupancies using radioligands of different affinities may obscure or create differential occupancies between regions of different receptor densities.