Quantitative assessment of translocator protein (TSPO) in the non-human primate brain and clinical translation of [18F]LW223 as a TSPO-targeted PET radioligand

Abstract

Objectives: Translocator protein 18 kDa (TSPO) positron emission tomography (PET) can be harnessed for the non-invasive detection of macrophage-driven inflammation. [¹⁸F]LW223, a newly reported TSPO PET tracer which was insensitive to rs6971 polymorphism, showed favorable performance characteristics in a recent imaging study involving a rat myocardial infarction model. To enable quantitative neuroimaging with [¹⁸F]LW223, we conducted kinetic analysis in the non-human primate (NHP) brain. Further, we sought to assess the utility of [¹⁸F]LW223-based TSPO imaging in a first-in-human study.

Methods: Radiosynthesis of [¹⁸F]LW223 was accomplished on an automated module, whereas molar activities, stability in formulation, lipophilicity and unbound free fraction (f_u) of the probe were measured. Brain penetration and target specificity of [¹⁸F]LW223 in NHPs were corroborated by PET-MR imaging under baseline and pre-blocking conditions using the validated TSPO inhibitor, (R)-PK11195, at doses ranging from 5 to 10 mg/kg. Kinetic modeling was performed using one-tissue compartment model (1TCM), two-tissue compartment model (2TCM) and Logan graphical analyses, using dynamic PET data acquisition, arterial blood collection and metabolic stability testing. Clinical PET scans were performed in two healthy volunteers (HVs). Regional brain standard uptake value ratio (SUVr) was assessed for different time intervals.

Results: [¹⁸F]LW223 was synthesized in non-decay corrected radiochemical yields (n.d.c. RCYs) of 33.3 ± 6.5% with molar activities ranging from 1.8 ± 0.7 Ci/µmol (n = 11). [¹⁸F]LW223 was stable in formulation for up to 4 h and LogD_7.4 of 2.31 ± 0.13 (n = 6) and f_u of 5.80 ± 1.42% (n = 6) were determined. [¹⁸F]LW223 exhibited good brain penetration in NHPs, with a peak SUV value of ca. 1.79 in the whole brain. Pre-treatment with (R)-PK11195 substantially accelerated the washout and attenuated the area under the time-activity curve, indicating in vivo specificity of [¹⁸F]LW223 towards TSPO. Kinetic modeling demonstrated that 2TCM was the most suitable model for [¹⁸F]LW223-based neuroimaging. Global transfer rate constants (K₁) and total volumes of distribution (V_T) were found to be 0.10 ± 0.01 mL/cm³/min and 2.30 ± 0.17 mL/cm³, respectively. Dynamic PET data analyses across distinct time windows revealed that the V_T values were relatively stable after 60 min post-injection. In a preliminary clinical study with two healthy volunteers, [¹⁸F]LW223 exhibited good brain uptake and considerable tracer retention across all analyzed brain regions. Of note, an excellent correlation between SUVr with V_T was obtained when assessing the time interval from 20 to 40 min post tracer injection (SUVr_{(20-40 min)}, R² = 0.94, p < 0.0001), suggesting this time window may be suitable to estimate specific binding to TSPO in human brain.

Conclusion: Our findings indicate that [¹⁸F]LW223 is suitable for quantitative TSPO-targeted PET imaging in higher species. Employing state-of-the-art kinetic modeling, we found that [¹⁸F]LW223 was effective in mapping TSPO throughout the NHP brain, with best model fits obtained from 2TCM and Logan graphical analyses. Overall, our results indicate that [¹⁸F]LW223 exhibits favorable tracer performance characteristics in higher species, and this novel imaging tool may hold promise to provide effective neuroinflammation imaging in patients with neurological disease.

Keywords: Kinetic modeling; Neuroinflammation; Positron emission tomography (PET); Translational molecular imaging; Translocator protein 18 kDa (TSPO).