Original Article
European Journal of Nuclear Medicine and Molecular Imaging
pp 1-9
- Rawaha AhmadAffiliated withDepartment of Imaging & Pathology, Nuclear Medicine and Molecular Imaging, KU Leuven and University Hospitals LeuvenDivision of Nuclear Medicine, University Hospital Leuven Email author
- , Andrey Postnov
- , Guy Bormans
- , Jan Versijpt
- , Mathieu Vandenbulcke
- , Koen Van Laere
Keywords
Cannabinoid type 2 receptor CB2R, Neuroinflammation, Alzheimer’s disease, PET imaging
Abstract
Purpose
The cannabinoid type 2 receptor (CB2R) is expressed by immune cells such as monocytes and macrophages. In the brain, CB2R is primarily found on microglia. CB2R upregulation has been reported in animal models of Alzheimer’s disease, with a preferential localization near amyloid beta (Aβ) plaques, and in patients post mortem. We performed in vivo brain imaging and kinetic modelling of the CB2R tracer [11C]NE40 in healthy controls (HC) and in patients with Alzheimer’s disease (AD) to investigate whether higher CB2R availability regionally colocalized to Aβ deposits is present in vivo.
Methods
Dynamic 90-min [11C]NE40 PET scans were performed in eight HC and nine AD patients with full kinetic modelling using arterial sampling and metabolite correction and partial volume correction. All AD patients received a static [11C]PIB scan 40 min after injection. In four HC, a retest scan with [11C]NE40 PET was performed within 9 weeks to investigate test–retest characteristics.
Results
[11C]NE40 was metabolized quickly leading to 50 % of intact tracer 20 min after injection and 20 % at 90 min. A two-tissue kinetic model fitted most of the time–activity curves best; both binding potential (BPND) and distribution volume (V T) parameters could be used. Brain uptake was generally low with an average K 1 value of 0.07 ml/min/ml tissue. V T and BPND were in the range of 0.7 – 1.8 and 0.6 – 1.6, respectively. Test values in HC were about 30 % for V T and BPND. AD patients showed overall significantly lower CB2R binding. No relationship was found between regional or global amyloid load and CB2R availability.
Conclusion
Kinetic modelling of [11C]NE40 is possible with a two-tissue reversible model. In contrast to preclinical and post-mortem data, [11C]NE40 PET shows lower CB2R availability in vivo in AD patients, with no relationship to Aβ plaques. A possible explanation for these findings is that [11C]NE40 binds to CB2R with lower affinity and/or selectivity than to CB1R.
