Development of Novel Brain-Penetrant Radioligands for PET Imaging of GABA Transporter-1

Objectives: γ-Aminobutyric acid (GABA) is the predominant inhibitory neurotransmitter in the central nervous system and modulated by GABA transporters (GAT) with 3 major subtypes, GAT-1 to 3. Dysregulation of GABAergic transmission has been implicated in numerous neurologic and psychiatric diseases. Over the years in vivo investigation of GABA system involvement in diseases has been made possible by the availability of PET radioligands for the GABA_A receptors such as [¹¹C]flumazenil. However, development of PET ligands for the major GABA transporter in the brain, GAT-1, has not been successful due to limited brain permeability of available compounds. Here we describe the design, synthesis and evaluation of novel, brain-penetrant GAT-1 radiotracers.

Methods: A library of candidate ligands was designed and synthesized using tiagabine as the lead structure. The affinity for GAT-1 was assessed by the ability of compounds to inhibit the binding of [³H]tiagabine to rat brain sections using an autoradiographic binding assay. In addition, lipophilicity (LogD) and passive membrane permeability (PAMPA assay) were determined experimentally. The efflux ratio by the human P-glycoprotein (Pgp) transporter was measured for the most promising candidates. Four compounds were selected as priority candidates for radiolabeling and in vivo evaluation based on their high affinity for GAT-1 (IC₅₀≤100 nM), favorable lipophilicity (LogD=1-3), good passive permeability and minimal liability for Pgp efflux. Two radioligands, [¹⁸F]GATT-34 and [¹⁸F]GATT-44 were synthesized from their respective aryltrimethyltin precursors via Cu(II)-catalyzed ¹⁸F-fluorination and subsequent hydrolysis of the ester group. Baseline and tiagabine blocking scans of up to 180 min were conducted on the Focus-220 scanner in rhesus macaques to evaluate the distribution, brain uptake, and binding specificity of the radioligands. Regional volume of distribution (V_T, mL/cm³) was estimated with kinetic modeling using the plasma input function generated by arterial sampling and metabolite analysis for each scan. The non-displaceable volume of distribution (V_ND) derived from the tiagabine blocking scan was used to calculate regional binding potential (BP_ND, = (V_T- V_ND)/V_ND).

Results: Both radioligands were prepared in >96% radiochemical purity and mean molar activity of 110.3 MBq/nmol ([¹⁸F]GATT-34, n=4) and 233.8 MBq/nmol ([¹⁸F]GATT-44, n=5) at the end of synthesis. In rhesus monkeys both tracers appeared to be metabolically stable, with parent fraction of >75% in plasma at 180 min post-injection. Plasma free fraction was 10.9±0.4% for [¹⁸F]GATT-34 (n=4) and 14.8±1.2% for [¹⁸F]GATT-44 (n=5). PET imaging results showed that both radioligands entered the brain readily and distributed heterogeneously, with higher uptake in the cortex, moderate in striatum and cerebellum, and lower in hippocampus and thalamus. Time-activity curves were well fitted by the 1-tissue compartment (1TC) model to derive reliable V_T values (Table 1). Binding of the radiotracers in the brain appeared to be specific to GAT-1, as pre-treatment of the animals with tiagabine (0.5 mg/kg, i.v.) significantly reduced the specific binding of both [¹⁸F]GATT-34 and [¹⁸F]GATT-44 across the brain. Regional BP_ND values calculated with V_ND estimated from the blocking scans were higher for [¹⁸F]GATT-44 than [¹⁸F]GATT-34 (Table 1).

Conclusions: We have synthesized two novel radioligands [¹⁸F]GATT-34 and [¹⁸F]GAT-44 and evaluated their pharmacokinetic and binding properties in rhesus monkeys. Both radiotracers were demonstrated to enter the brain and bind specifically to GAT-1, with [¹⁸F]GAT-44 displaying higher specific binding signals. Taken together, [¹⁸F]GAT-44 appears to be a promising PET radiotracer for GAT-1. This is the first report of brain-penetrant radioligands for successful imaging and quantification of GAT-1 in the non-human primate brain. Acknowledgements: Supported by NIH grant U01MH107803.