Evaluation of a novel positron emission tomographic radioligand to measure cyclooxygenase-2 in a model of neuroinflammation in rhesus macaque

Background: The cyclooxygenase (COX) isozymes, COX-1 and COX-2, mediate inflammatory responses, and are targets for nonsteroidal anti-inflammatory drugs (NSAIDs). To study the distribution and function of these COX isozymes in vivo, our laboratory developed two novel PET radioligands: ¹¹C-PS13 and ¹¹C-MC1. PS13 demonstrated potent selectivity for COX-1 (IC₅₀ = 1 nM) compared to COX-2 (IC₅₀ > 1,000 nM). Conversely, MC1 was potent and selective for COX-2 (IC₅₀ = 1 nM) compared to COX-1 (IC₅₀ > 1,000 nM). Previously, we showed that ¹¹C-PS13, but not ¹¹C-MC1 shows specific uptake in normal monkey brain. Whether ¹¹C-MC1 exhibits specific binding in conditions that elicit significant densities of COX-2 has yet to be studied. This study sought to examine whether ¹¹C-MC1 could image COX-2 in a model of neuroinflammation via intracerebral injection of lipopolysaccharide (LPS). As a control, we measured COX-1 using ¹¹C-PS13 before and after LPS.

Methods: To elicit an inflammatory response in monkey brain, LPS (from Escherichia coli O26:B6), was injected into the right putamen of monkeys (n=4). Prior to injection, a T1-weighted MRI was obtained to guide the insertion of the needle. In total, 10 µg of LPS was injected at a concentration of 1 µg/µL and an infusion rate of 0.5 µL/min. Dynamic brain PET scans were acquired for two hours both pre- and post-LPS injection. Approximately 220 MBq of radioligand was injected intravenously into the monkey before each scan. Blocking studies were also conducted with non-radioactive PS13 or MC1 (0.3-1 mg/kg) to confirm the specific uptake of the radioligands in the brain. Full quantitation with arterial sampling was done to measure volume of distribution (V_T) before and after LPS injection.

Results: In this LPS model of neuroinflammation, ¹¹C-MC1 measured specific binding to COX-2. After the first injection, a widespread, 60% increase of ¹¹C-MC1 uptake was observed in the brain that was displaced by non-radioactive MC1. Following the second injection, there was an even more dramatic increase (>200%) in ¹¹C-MC1 uptake at the site of injection (Figure 1). As a comparator, we found no increase in COX-1 after LPS injection. ¹¹C-PBR28, a radioligand for translocator protein (TSPO), showed increased uptake, as expected, at the site of injection. Conclusions: ¹¹C-MC1 is useful for imaging COX-2 upregulation in rhesus macaque brain from basal levels to 60% after a single injection and >200% after a second injection. We show that in monkey brain, COX-2 but not COX-1 is upregulated after inflammation, which is consistent with the general notion that expression of COX-1 is constitutive and COX-2 is inducible. Of interest, COX-2 was primarily in neurons. Ongoing first in-human studies will measure COX-1 in healthy conditions, and COX-2 in inflammatory disorders, and can also be used to evaluate the selectivity of NSAIDs, and delivery to brain for the two COX isozymes.Figure 1. The effect of inflammogen on ¹¹C-MC1 uptake in a monkey brain shown as volume of distribution (V_T). In normal brain (pre-LPS), there was no detectable uptake. After the second injection of LPS into right putamen, ¹¹C-MC1 uptake increased by >200% in the area of lesion, and blocked by non-radioactive MC1 (1 mg/kg i.v.).

• Download figure
• Open in new tab
• Download powerpoint