Pharmacokinetic evaluation of the 5-HT₆ radiotracer [¹⁸F]2FNQ1P in rhesus macaque

Objectives: The 5-HT₆ class of serotonin neuroreceptors is implicated in various psychiatric disorders, memory, cognitive processing and regulation of food intake. [¹⁸F]2FNQ1P was previously reported as the first 5-HT₆ PET radiotracer with low in vitro affinity for the highly homologous 5-HT_2A receptor. Evaluation in cynomolgous macaques included blocking with a structurally distinct 5-HT₆ ligand, however those studies did not include blood sampling. The goal of the current work in rhesus macaque is to evaluate [¹⁸F]2FNQ1P with arterial sampling and metabolite analysis for full pharmacokinetic characterization and validation of a reference tissue, and to perform blocking studies to confirm 5-HT₆ selectivity is maintained in vivo.

Methods: Dynamic PET imaging with arterial blood sampling was performed in rhesus macaque. Collection of listmode data began concurrent with slow bolus injection of 250 MBq of [¹⁸F]2FNQ1P. Arterial samples were processed to measure concentrations in whole blood and plasma. Selected plasma samples underwent radiometabolite analysis with column-switching radio-HPLC. An arterial sample drawn immediately prior to tracer administration was incubated with [¹⁸F]2FNQ1P ex vivo and plasma protein binding was determined by the ultracentrifugation method. Reconstructed PET images were aligned to a stereotaxic template where volumes of interest were delineated. Regional time activity curves were extracted and the metabolite corrected arterial plasma input function was used to perform kinetic analysis with compartmental models and graphical analysis techniques.

Results: [¹⁸F]2FNQ1P was metabolized relatively slowly with ~70% parent fraction at 60 min. Free fraction in plasma was 23%. [¹⁸F]2FNQ1P in gray matter peaked quickly (SUV >3 at ∼6 minutes). Compartmental modeling resulted in good fits and with the two-tissue model (2T) determined best in most regions according to the Akaike information criterion. K₁ values were ~0.25-0.3 ml/min/cc and V_T values ranged from ~4.9 to 7.2 ml/cc with rank order of binding consistent with expected 5-HT₆ distribution: putamen > caudate > thalamus > cortical gray matter > hippocampus > cerebellum. Assuming the cerebellum to be a valid reference tissue, maximum regional BP_ND values were ~0.5 in the putamen but were robustly estimated. Analysis of truncated data sets yielded very similar results for 60, 90, or 120 min of data. Logan plots linearized quickly (t[asterisk]≤30 min) and provided V_T values having extremely strong correlation with compartmental modeling results (R²=0.99) though with modest underestimation (V_T[Logan]=0.93[asterisk]V_T[2T]+0.18); BP_ND estimates showed similarly strong correlation but with lower bias (BP_ND[Logan]=0.97[asterisk]BP_ND [2T]+0.01).

Conclusions: [¹⁸F]2FNQ1P data can be readily quantified in rhesus macaque using standard kinetic analysis methods with arterial input functions. Ongoing blocking studies are being conducted to test the validity of the cerebellum as a reference tissue and confirm in vivo selectivity for 5-HT₆ receptors. If verified, [¹⁸F]2FNQ1P appears to have considerable potential for non-invasive imaging of 5-HT₆ receptors in health and disease conditions. Acknowledgements: NIH R01MH100350, P41EB022544, and S10OD018035

• Download figure
• Open in new tab
• Download powerpoint