Abstract
81
Objectives Use of TSPO PET ligands is hampered by the lack of a reference tissue to define non-displaceable binding (VND). Pseudo-reference regions used with [11C]PK11195 (1) rely on an untested assumption of low specific binding in the healthy brain. We blocked [11C]PBR28 binding in healthy volunteers with a TSPO ligand (XBD173) to estimate VND. This was used to estimate BPND in subjects with different genotypes at the rs6971 SNP in the TSPO gene (high, low or mixed affinity binders (HAB, LAB, MAB))(2).
Methods Healthy volunteers (10 HAB, 6 MAB, 2 LAB) received a single [11C]PBR28 PET scan. 3 HAB were re-scanned following an oral dose of XBD173 (90, 45, 25 mg). 5 HAB and 3 MAB received a repeat baseline scan 3 months later, to evaluate within-subject reproducibility. Regional VT (derived from a 2TCM) was the primary outcome parameter. VND was estimated from the occupancy data (modified Lassen plot).
Results We found no evidence of a specific signal in LAB. Mean regional VT was 3.1-3.9 (MAB) and 4.9-6.6 (HAB). XBD173 reduced regional VT by 60% at the 90mg dose (ED50 ~14mg in HAB). [11C]PBR28 VND was estimated at 1.7 (95% CI 0.75, 2.16), implying a regional BPND of 1.6-2.9 (HABs) and 0.8-1.3 (MABs). Thus, the regional specific binding ratio of MABs:HABs was 0.42-0.48, consistent with our data from competition assays in vitro (3), which imply a ratio of 0.50. Within-subject reproducibility of [11C]PBR28 VT was modest (23-34% across ROI), though better for HAB alone (15-27%).
Conclusions This is the first TSPO blocking study in man and shows a substantial specific component in [11C]PBR28 VT in healthy brain. The validity of pseudo reference region approaches for modeling TSPO PET should therefore be considered carefully. In vivo estimates of [11C]PBR28 BPND correlated very well with the effect of the rs6971 SNP predicted by in vitro competition binding assays. Precise TSPO quantification in vivo may therefore require blocking studies.