An in vitro assay for predicting successful imaging radiotracers.

Purpose: To develop an in vitro binding assay able to predict whether a radiolabel is likely to be a useful clinical tracer for positron emission tomography (PET).

Procedures: Rodent and rhesus brain sections were incubated with radioligands, most of which are tritiated or iodinated versions of known clinical PET radiotracers, and assayed for binding to brain receptors for a 20-minute period using a no-wash protocol (n=>/=3).

Results: Radiolabeled flumazenil (RO-151788), WAY100635, N-methylscopolamine, N-methylspiperone, raclopride, citalopram, (1-)2,5-dimethoxy-4-iodophenyl-2-aminopropane (DOI), paroxetine, and 4-(2'-methoxyphenyl)-1-[2'-[N-(2"-pyridinyl)-p-flurobenzamido]ethyl]piperazine (MPPF) were assessed for binding to either rhesus caudate putamen, and/or frontal cortex, or rat whole brain sections. Specific binding for these compounds ranged from 0 to 94% by 20 minutes. Those with %-specific binding less than 10% have also been shown to not be effective as in vivo PET radiotracers. In addition, successful PET radiotracers incubated in tissue sections with target receptor either absent or present in low density behaved poorly in this assay, as expected, as did radiolabels previously shown to possess high non-specific binding.

Conclusions: An in vitro binding assay using rodent and rhesus brain sections has been developed that, within the currently assayed radiotracers, is able to rapidly predict whether a radiolabeled compound is a useful clinical PET radiotracer. This method suggests significant potential for the rapid in vitro evaluation of potential in vivo PET radiotracers.