Purpose: The in vivo binding parameters of the novel imidazopyridine TSPO ligand [(18)F]PBR102 were assessed and compared with those of [(18)F]PBR111 in a rodent model of neuroinflammation. The validity of the key assumptions of the simplified reference tissue model (SRTM) for estimation of binding potential (BP) was determined, with validation against a two-tissue compartment model (2TC).
Methods: Acute neuroinflammation was assessed 7 days after unilateral stereotaxic administration of (R,S)-α-amino-3-hydroxy-5-methyl-4-isoxazolopropionique (AMPA) in anaesthetized adult Wistar rats. Anaesthetized rats were implanted with a femoral arterial cannula then injected with a low mass of [(18)F]PBR102 or [(18)F]PBR111 and dynamic images were acquired over 60 min using an INVEON PET/CT camera. Another population of rats underwent the same PET protocol after pretreatment with a presaturating mass of the same unlabelled tracer (1 mg/kg) to assess the validity of the reference region for SRTM analysis. Arterial blood was sampled during imaging, allowing pharmacokinetic determination of radiotracer concentrations. Plasma activity concentration-time curves were corrected for unchanged tracer based on metabolic characterization experiments in a separate cohort of Wistar rats. The stability of neuroinflammation in both imaging cohorts was assessed by [(125)I] CLINDE TSPO quantitative autoradiography, OX42/GFAP immunohistochemistry, Fluoro-Jade C histology, and elemental mapping using microparticle-induced x-ray emission spectroscopy. The BP of each ligand were assessed in the two cohorts of lesioned animals using both SRTM and a 2TC with arterial parent compound concentration, coupled with the results from the presaturation cohort for comparison and validation of the SRTM.
Results: The BPs of [(18)F]PBR102 [(18)F]PBR111 were equivalent, with improved signal-to-noise ratio and sensitivity compared with [(11)C]PK11195. The presaturation study showed differences in the volume of distribution between the ipsilateral striatum and the striatum contralateral to the injury (0.7) indicating that an assumption of the SRTM was not met. The modelling indicated that the BPs were consistent for both ligands. Between the SRTM and 2TC model, the BPs were highly correlated, but there was a bias in BP.
Conclusion: [(18)F]PBR102 and [(18)F]PBR111 have equivalent binding properties in vivo, displaying significantly greater BPs with lower signal-to-noise ratio than [(11)C]PK11195. While an assumption of the SRTM was not met, this modelling approach was validated against 2TC modelling for both ligands, facilitating future use in longitudinal PET imaging of neuroinflammation.