In vivo binding of antipsychotics to D3 and D2 receptors: a PET study in baboons with [11C]-(+)-PHNO

Abstract

Measuring the in vivo occupancy of antipsychotic drugs at dopamine D(2) and D(3) receptors separately has been difficult because of the lack of selective radiotracers. The recently developed [(11)C]-(+)-PHNO is D(3)-preferring, allowing estimates of the relative D(2) and D(3) binding of antipsychotic drugs. We used positron emission tomography (PET) imaging in baboons with [(11)C]-(+)-PHNO to examine the binding of clozapine and haloperidol to D(2) and D(3) receptors. Four animals were scanned with dynamically acquired PET and arterial plasma input functions. Test and retest scans were acquired in single scanning sessions for three subjects to assess the reproducibility of [(11)C]-(+)-PHNO scans. Four additional scans were acquired in each of three subjects following single doses of antipsychotic drugs (clozapine 0.5534 mg/kg, haloperidol 0.0109 mg/kg, two administrations per drug per subject) and compared with baseline scans. The percent change in binding (ΔBP(ND)) following challenges with antipsychotic drugs was measured. A regression model, based on published values of regional D(2) and D(3) fractions of [(11)C]-(+)-PHNO BP(ND) in six brain regions, was used to infer occupancy at D(2) and D(3) receptors. BP(ND) following antipsychotic challenge decreased in all regions. Estimated D(2) : D(3) selectivity was 2.38 for haloperidol and 5.25 for clozapine, similar to published in vitro values for haloperidol (3.03), but slightly higher for clozapine (2.82). These data suggest that acute doses of clozapine and haloperidol bind to D(3) receptors in vivo, and that the lack of D(3) occupancy by antipsychotics observed in some recent imaging studies may be because of other phenomena.

Figure 1

(a) Averaged arterial plasma concentration of [¹¹C]-(+)-PHNO under baseline (▾, n=8 scans), clozapine challenge (▴, n=6 scans) and haloperidol challenge (•, n=6 scans) conditions. Continuous curves (—) represent the average of the modeled fits as described in the Methods section. Data are normalized to injected activity. The graph shows that [¹¹C]-(+)-PHNO concentration in arterial plasma was unaffected by the drug challenges. (b) Averaged time activity curves from all studies under baseline (▾, n=8 scans), clozapine challenge (▴, n=6 scans) and haloperidol challenge (•, n=6 scans) conditions. Continuous curves (—) represent the average of the two-tissue compartment model (2TC) fits. Error bars are the SD across the measured data. Data are normalized to injected activity and body mass and are shown for a region with similar contributions from dopamine-2 (D₂) and dopamine-3 (D₃) binding (globus pallidus, (GP)), mostly D₂ binding (putamen) and mostly D₃ binding (substantia nigra/ventral tegmental area (SN/VTA)), as well as the cerebellum.

Figure 2

[¹¹C]-(+)-PHNO BP_ND maps in one baboon at the level of the GP across the three conditions. Images are the average of three baseline scans and two post-drug scans for each antipsychotic, and were generated at each voxel with the SRTM algorithm.

Figure 3

Scatter plots of regions of interest (ROI) analysis using SRTM. The left panel shows SRTM-derived BP_ND using the full 120 min of data plotted against 2TC-derived BP_ND. The graph shows that there is a similar proportionality between methods across moderate-binding regions, but not in high-binding regions. In the right panel, SRTM-derived data using 90-min truncated data are plotted against SRTM-derived data using 120 min of data. The graph shows the measures are virtually identical in the moderate-binding regions, but that the decrease in high-binding regions (relative to 2TC) is more pronounced with the truncated data. The solid line in each graph is the line of identity.