¹¹C-MC1 has adequate sensitivity to measure low density cyclooxygenase 2 (COX-2) in healthy human brain

Introduction: Cyclooxygenase enzymes (COX) are important targets for neuroinflammation. COX-1 is constitutively present at such high density in the brain and periphery that it can measured with PET imaging using a COX-1 selective radioligand (doi: 10.1021/acschemneuro.8b00103). In contrast, COX-2 is present at low density in healthy tissue but can be markedly upregulated by inflammation. ¹¹C-MC1 is a selective and potent radioligand for COX-2, and it detected COX-2 in nonhuman primates after intracerebral injection of an inflammogen. We have found that in patients with rheumatoid arthritis, ¹¹C-MC1 detected upregulated COX-2 in symptomatic joints (doi: 10.1186/s12974-020-01804-6). The purpose of this study was to determine if ¹¹C-MC1 could measure COX-2 in healthy human brain.

Methods: Five healthy participants were injected with ¹¹C-MC1 (730 ± 37 MBq) and had two brain PET scans for 120 minutes with concurrent arterial sampling. The baseline scan was followed by one after blockade with celecoxib (600 mg p.o.), a preferential COX-2 inhibitor. Binding to the enzyme was calculated as total distribution volume corrected for free parent fraction in plasma (V_T/f_p) using a two-tissue compartment model (2TCM), and the receptor occupancy was determined using the Lassen plot.

Results: After injection of ¹¹C-MC1, the concentration of brain radioactivity peaked at 4.0 SUV at ~2.8 minutes and declined to 16% of the peak at 120 minutes. The brain time-activity curve was reasonably well-fitted to a 2TCM and provided stable values of V_T after 40 minutes (i.e. within 10% of terminal values at 120 min). However, the measured values of V_T after 60 minutes were higher than the fitted values, which might reflect accumulation of small amount of radiometabolites in brain. Nondisplaceable uptake (V_ND), calculated from the Lassen plot, was 80-90% of total uptake, and specific uptake (V_S) was only 10-20% of V_T. Based on the Lassen plot, celecoxib occupied 86% of available COX-2 in the brain. The regions with the highest amount of specific binding were occipital cortex, insula, and prefrontal cortex. Regional brain distribution of specific binding correlated with that of regional mRNA transcripts of the COX-2 gene obtained from the Allen Brain Atlas.

Conclusions: The uptake of ¹¹C-MC1 in healthy human brain reflects specific binding to COX-2 based on blockade by celecoxib and by the PET distribution of the enzyme correlating with the density of the gene transcript. Based on these preliminary results, ¹¹C-MC1 has adequate sensitivity to measure the low-density of COX-2 in a healthy human brain. However, the specific binding was only 10-20% of total uptake, and greater confidence in these results will likely require studying neuroinflammatory disorders with higher density of COX-2. Acknowledgments: Intramural Research Program at the National Institute of Mental Health