Integrated analysis of Human PET data across multiple brain regions and receptors to make inferences from limited data.
Bart Laurijssens
Clinical Pharmacology Modelling and Simulation, GSK, Greenford, United Kingdom
Objectives: To estimate the concentration - receptor occupancy relationship for the main target brain receptor (R1) for centrally acting drugs using human PET data. The data was limited in that for R1 maximal displacement was observed for all dose levels studied, and the extent displacement differed between two relevant brain regions.
Methods: Data consisted of 1 baseline scan + 2 scans at steady state for each subject 5 hrs after dosing. There were 8 subjects divided over 3 dose levels (10 fold range). For each scan, data was available for 3 brain regions and 2 target receptors (R1 and R2): frontal cortex (FCTX, R2), caudate (CDT,R1) and putamen (PTM, R1), as well at the average plasma concentration during the scan. A model integrating data in all brain regions and both receptors was fitted to the measured Volume of distribution (Vd) ratio (Vdt/Vdr, which is equal to the Binding Potential + 1). Several hypotheses for the difference in displacement between caudate and putamen were tested. Although the IC50 for R1 could not be estimated accurately (no data below maximal displacement), a likely upper limit could be determined using log-likelihood profiling.
Results: The plasma IC50s were estimated to be: R2 (FCTX) 69 ng/ml (95% CI: 55, 90), and R1 (CDT and PTM) < 2.6 ng/ml. The analysis also suggested about 29% unexplained non-specific binding in PTM.
Conclusions: The final integrated model allowed estimation of the concentration -receptor occupancy relationship for R2 (FCTX) (not the main objective), as well as the likely upper limit of the IC50 for our main receptor of interest. From this the minimal R1 occupancy for a given dose could be estimated. The results were consistent with previous results and predictions. The modelling also allowed identification of the most likely explanation of those tested, additional unexplained non-specific binding in PTM, for the discrepancy between CDT and PTM data.
