Quantification of Muscarinic Cholinergic Receptors with [11C]NMPB and Positron Emission Tomography: Method Development and Differentiation of Tracer Delivery from Receptor Binding

Quantification of human brain muscarinic cholinergic receptors was investigated with the use of [11C]N-methyl-4-piperidyl benzylate (NMPB) and positron emission tomography (PET). Whole-brain uptake of NMPB at 90 to 110 minutes after intravenous injection was approximately 10% of the administered dose. The initial cerebral distribution of NMPB corresponded to the pattern of cerebral perfusion; however, at progressively longer postinjection intervals, regional distinctions consistent with muscarinic receptor binding were evident: activity at 90 to 110 minutes postinjection was highest in the striatum and cerebral cortex, intermediate in the thalamus and pons, and lowest in the cerebellum. After the development of a chromatographic system for isolation of authentic [11C]NMPB in plasma, tracer kinetic modeling was used to estimate receptor binding from the cerebral and arterial plasma tracer time-courses. Ligand transport rate and receptor-binding estimates were obtained with the use of compartmental models and analytical methods of varying complexity, including a two-parameter pixel-by-pixel-weighted integral approach and regional least-squares curve-fitting analyses employing both two-and three-compartment model configurations. In test—retest experiments, precision of the methods and their abilities to distinguish altered ligand delivery from binding in occipital cortex during an audiovisual presentation were evaluated. Visual stimulation increased the occipital blood-to-brain NMPB transport rate by 25% to 46% in estimates arising from the various approaches. Weighted integral analyses resulted in lowest apparent transport changes and in a concomitant trend toward apparent binding increases during visual activation. The regional least-squares procedures were superior to the pixel-by-pixel method in isolating the effects of altered tracer delivery from receptor-binding estimates, indicating larger transport effects and unaltered binding. Precision was best (less than 10% test—retest differences) for the weighted integral analyses and was somewhat lower in the least-squares analyses (10–25% differences). The authors conclude that pixel-by-pixel-weighted integral analyses of NMPB distribution introduce transport biases into receptor-binding estimates. Similar confounding effects also are predicted in noncompartmental analyses of delayed radiotracer distribution. The use of regional nonlinear least-squares fitting to two- and three-compartment models, although more labor intensive, provides accurate distinction of receptor-binding estimates from tracer delivery with acceptable precision in both intra- and intersubject comparisons.