1.5 Radiotracers for CB1 receptors

1 Ligand Development

1.5 Radiotracers for CB₁ receptors

Only quite recently have the actions of marijuana begun to be understood at the biochemical level.⁶⁸ In the endocannabinoid system, two sub-types of G-protein-coupled receptor, CB₁ and CB₂, are recognized as mediating effects arising from exposure to cannabomimetic ligands. CB₁ receptors are abundant in brain and located at presynaptic nerve terminals. They represent an interesting target for the treatment of several psychiatric (e.g. anxiety, addiction and depression) and neurodegenerative disorders (e.g. Huntington’s disease and Tourette’s syndrome). However, the role of the CB₁ receptors in these disorders is poorly understood. PET coupled to an effective radioligand for CB₁ receptors would provide one important means for understanding the many such disorders linked to this receptor.

Molecular modification of the 1,5-diarylpyrazole core of SR 141716A (rimonabant), a potent (antagonist binding affinity K_B = 6.8 nM) and selective CB₁ receptor antagonist, currently under clinical trial, has been the focus for prior ligand and radiotracer development. The most notable of these ligands, AM 251 (K_B = 3.8 nM), is amenable to labeling with iodine-123, making it potentially useful for SPECT imaging. Studies with [¹²³I]AM 251 have however been disappointing because of its poor brain penetration, which may be attributed to its very high lipophilicity (cLogP, 7.36). This conclusion prompted the development of AM 281 (K_B = 8.9 nM), an analog of AM 251 with much lower lipophilicity (cLogP, 5.66). AM281 has also been labeled with iodine-123 and, although studied in Tourette’s syndrome patients, gives very low signal.⁶⁹ Several other candidate radiotracers based on rimonabant have been investigated by other groups but none has proven to be useful.

Initially, we set out to synthesize and characterize further ligands based on the 1,5-diarylpyrazole template with a view to discovering improved radiotracers for PET imaging. In the prepared series of 19 compounds, K_B values at CB₁ receptors ranged from 11 to > 16,000 nM, CB₁ vs CB₂ selectivities from 0.6 to 773 and cLogPs from 3.61 to 6.25. An interesting new ligand, emerged with more appealing properties (K_B, 11 nM; CB1 selectivity, > 773; cLogP, 5.85; 48) than any predecessor for labeling with carbon-11 and evaluation.⁷⁰ A lower affinity chloro analog (K_B, 25 nM; CB₁ selectivity, > 84; cLogP, 5.69; 49) of this compound had been suggested by others to be a promising radioligand, based on in vitro autoradiography findings against human brain tissue.⁷¹ Both ligands were labeled with carbon- 11 and evaluated in cynomolgus⁷² and rhesus monkeys. Each radiotracer gave low brain uptake of radioactivity in baseline experiments (100-150 %SUV). Use of a CB₁-selective ligand from Lilly (Antagonist LY1) in pre-block experiments resulted in increased brain uptake of radioactivity, followed by fast washout of radioactivity from all brain regions, suggestive of peripheral blockade of CB₁ receptors and indicating appreciable receptor-specific signal. However, neither radiotracer was suitable for further development due to unfavorable kinetics.

In 2004, Solvay Pharmaceuticals disclosed less lipophilic high affinity CB₁ receptor antagonists.⁷³ We prepared several analogs in a search for ligands with improved properties and discovered two new interesting racemic ligands, a methoxy compound (CB₁ K_B, 8 nM; CB₂ K_B, > 3875 nM; cLogP, 4.56; 50) and a nitrile compound (CB₁ K_B, 3 nM; CB₂ K_B, > 3875 nM; cLogP, 3.81; 51). The former was labeled by ¹¹C-methylation and evaluated in rhesus monkey, but proved unpromising.⁷⁴ The latter was labeled with [¹¹C]cyanide and evaluated in cynomolgus monkey. This radiotracer gave high uptake of radioactivity in brain (4% injected dose at 25 min). In pre-blocking and displacement experiments with Antagonist LY1, radioactivities in brain and ROIs were reduced to the same low level (~ 1.5% of injected dose in whole brain) at 90 min after radiotracer injection. Ratios of radioactivity in CB₁ receptor-rich putamen to that in pons reached about 1.75. The data however indicate that pons also contains CB₁ receptors and is therefore an invalid reference region i.e. the receptorspecific signal is underestimated. This radiotracer therefore appears quite promising.

Under the CRADA with Eli Lilly & Co, we identified two ligands (with structures that must be kept confidential under our agreement) for labeling and evaluation, Ligand A (CB₁ K_B, 1 nM; CB₂ K_B, > 7300 nM; cLogP, 5.14) and Ligand B (CB₁ K_B, 0.35 nM; CB₂ K_B, 280 nM; cLogP, 5.73). Ligand A was labeled with carbon-11 in a single step with [11C]carbon monoxide as labeling agent, and also with fluorine-18 in three steps. The 11C-labeled tracer gave very low brain uptake of radioactivity in rat and mouse (~ 40% SUV). Pretreatment of rats with CsA dramatically increased the brain radioactivity uptake (to ~180% SUV in striatum at 10 min). This was shown to be due to a greater plasma input of radiotracer into brain and not to a blockade of Pgp (i.e. the radiotracer is not a Pgp substrate). Under the effect of CsA, rat brain radioactivity was highly (~ 50%) displaceable by Antagonist LY1. The brain uptake of radioactivity in CB₁ receptor knockout mouse was almost negligible (~ 5 %SUV) compared to that in wild type mouse, so indicating the latter uptake to be predominantly CB₁ receptor-specific binding.

In rhesus monkey, this radiotracer gave somewhat higher brain uptake of radioactivity (~ 120 %SUV) early after injection (~ 10 min) followed by a steady increase to 160 %SUV at 120 min. In a self-block experiment there was the same early initial uptake of radioactivity followed by washout of radioactivity from all ROIs (e.g. cerebellum, striatum, neocortex), indicative of receptor-specific signal in these ROIs in the baseline experiment. Self displacement was also demonstrated with whole brain radioactivity reduced from 105 %SUV to ~ 33 %SUV. Displacement of radioactivity from ‘pons‚ was observed, showing once again this cannot be used as a reference CB₁- receptor devoid region. A transient equilibrium of specific binding was not reached, suggesting a need for a longer-lived fluorine-18 label.

Ligand B was labeled efficiently by single-step ¹¹C-methylation. In a baseline experiment in rhesus monkey, this radiotracer gave a high early uptake of radioactivity (510 %SUV in cerebellum at ~ 10 min) which was mostly retained up to 120 min (Panel A, Fig. 5). In a preblock experiment with Antagonist LY1 all ROIs reduced to ~ 130 % SUV at 120 min, so providing evidence for a strong receptor-specific signal in the baseline experiment (Panel B, Fig. 5). A second CB₁ receptor-selective agent from Lilly (Antagonist LY2) gave the same pre-block effect, and also induced displacement of radioactivity from all ROIs (Panel C, Fig. 5). The distribution of radioactivity in brain accords with the expected distribution of CB1 receptors (Fig. 6). This radiotracer rapidly produced radiometabolites in monkey plasma, but it was possible to fit the PET data to a 2-TCM. Acquisition of PET data over 110 min was required to derive accurate measures of total V. Since, in monkey, the kinetics are quite slow, a longer-lived version of this radiotracer may be preferred for precise quantitation, especially in human subjects. Ligand B was found not to be a substrate for Pgp in rat or mice by comparison of radiotracer behaviors in wild type and Pgp knockout mice. PET experiments comparing brain uptake in wild type and CB₁ knockout mice confirmed the CB₁ receptor selectivity of this radiotracer, which is concluded to be very promising.

Research in this area was in collaboration with Prof. C. Halldin (Karolinska Institutet), with chemistry mostly conducted under a PhD studentship (Mr. S. Donohue) within the NIH-Karolinska Institutet Graduate Training Partnership in Neuroscience. Studies on Ligand A and B were performed under a CRADA with Eli Lilly & Co.