Positron Emission Tomography Imaging of the Endocannabinoid System: Opportunities and Challenges in Radiotracer Development

Abstract

The endocannabinoid system (ECS) is involved in a wide range of biological functions and is comprised of cannabinoid receptors and enzymes responsible for endocannabinoid synthesis and degradation. Over the past two decades, significant advances towards developing drugs and positron emission tomography (PET) tracers targeting different components of the ECS have been made. Herein, we summarized the recent development of PET tracers for imaging cannabinoid receptors 1 (CB1R) and 2 (CB2R) as well as the key enzymes monoacylglycerol lipase (MAGL) and fatty acid amide hydrolase (FAAH), particularly focusing on PET neuroimaging applications. State-of-the-art PET tracers for the ECS will be reviewed including their chemical design, pharmacological properties, radiolabeling, as well as preclinical and human PET imaging. In addition, this review addresses the current challenges for ECS PET biomarker development and highlights the important role of PET ligands to study disease pathophysiology as well as to facilitate drug discovery.

Graphical Abstract

1. INTRODUCTION

The endocannabinoid system (ECS) constitutes a lipid-based signaling system in the human central nervous system (CNS), which includes primarily two cannabinoid receptors (CB1R and CB2R), their endogenous ligands, and enzymes responsible for ligand synthesis and degradation. In the brain, 2-arachidonoylglycerol (2-AG) and N-arachidonoylethanolamine (AEA), the most well-known endogenous ligands for cannabinoid receptors, are produced by the postsynaptic membrane as “on-demand” to stimulate adjacent CB1 receptors. Scheme 1 depicts a general model for the ECS signaling in the CNS. This model, whereby post-synaptic endocannabinoid synthesis in response to a presynaptic signal acts presynaptically to depress the release of the presynaptic neurotransmitter, has been suggested to provide a protective negative feedback mechanism against over-activity of the presynaptic nerves.¹ Peripheral roles of the endocannabinoid system in the regulation of processes as diverse as gut function and bone turnover are also well-described.² These endogenous cannabinoids are mainly synthesized by membrane phospholipids on the postsynaptic membrane.³ Several catabolic pathways responsible for the termination of endocannabinoid signaling have been described.⁴ Among all hydrolytic enzymes, fatty acid amide hydrolase (FAAH) is the main metabolic enzyme of AEA, while in the brain monoacylglycerol lipase (MAGL) is primary metabolic enzyme of 2-AG.^4–6 Nonetheless, other serine hydrolases α/β hydrolase domain-6 (ABHD6) and α/β hydrolase domain-12 (ABHD12) may be more important for endocannabinoid metabolism in microglial cells.⁷ Additional catabolic pathways, such as oxidation catalyzed by cyclooxygenase-2 and by some members of the cytochrome P450 (CYP) family of enzymes have also been reported and yield biologically active products,^{4, 8} but are outside the scope of the present article.

Scheme 1.

A general model for the ECS signaling in the nervous system

The ECS regulates important physiological processes such as pain, emotion, stress, exercise, and cognition functions.^{2, 9, 10} Accordingly, cannabinoid receptors have become promising targets for the management of numerous pathologies including, but not limited to, chronic pain, epilepsy, anxiety, depression, Alzheimer disease (AD), Parkinson’s syndrome (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS), drug dependence, osteoporosis and liver, kidney, intestine, cardiovascular dysfunction (CVD).^11–15,16 Cannabinoids, particularly phytocannabinoid preparations, also have a wide off-label use in the treatment of chronic pain, despite modest evidence of efficacy in rigorous randomized clinical trials and a high incidence of unwanted effects.¹⁷

Positron emission tomography (PET) is a non-invasive and highly sensitive molecular imaging technique that allows in vivo quantification of biochemical and pharmacological processes under healthy and diseased conditions.¹⁸ Using radioactive PET tracers, in vivo imaging can be achieved at the molecular level to obtain quantitative information on target engagement and the degree of receptor occupancy at a given drug dose. PET scanning equipment and PET probes are just like the relationship between pistol and bullet. The bullet can accurately target in vivo biomarkers and provide real-time information on tracer accumulation in the area of interest, which plays a crucial role in precision medicine and quantitative pharmacokinetic research.^{19, 20} Such information is invaluable and harbours enormous potential to accelerate drug development,^{21, 22} especially in Phase 0 clinical trials.^23–26 Over the past decades, while many PET probes targeting endogenous cannabinoid receptors (CB1R and CB2R) and hydrolases (FAAH and MAGL) have substantially advanced our understanding of the ECS,²⁷ the vast majority of reported reviews are confined to one specific ECS component such as CB1R²⁸ or CB2R.^{29, 30} The last review of CNS PET did not cover the radiotracers targeting cannabinoid receptors and MAGL.³¹ In contrast, this review provides a comprehensive summary of PET-based ECS imaging, thereby covering CB1R-, CB2R-, FAAH- and MAGL-targeted probes, as well as presenting challenges in drug discovery, radioligand development, and medical imaging applications. For each ECS component, representative PET probes will be introduced with their pharmacological properties, radiolabeling techniques, and PET imaging in rodents/nonhuman primates (NHPs)/human subjects.

2. RECEPTOR TARGETING PET TRACERS

Both CB1R and CB2R are members of the G-protein-coupled receptor (GPCR) superfamily containing seven transmembrane structures.³² CB1R density is high in the human brain, especially in the substantia nigra, globus pallidus, putamen, hippocampus and cerebellum.^33–35 CB1R is also abundantly expressed in peripheral nerves and tissues, though, in a region-specific manner.³⁶ For example, in the gastrointestinal system, CB1R can regulate intestinal peristalsis and the secretion of gastric acid and neurotransmitter, which in turn affect appetite and digestion.³⁷ In the cardiovascular system, CB1R is significantly upregulated under pathological conditions and promotes disease progression.^{38, 39} CB1R expression has also been detected in adipose tissue, bone, skin, as well as several cancer cells.^{36, 40}

Although CB2R and CB1R share 44% structural homology, their distribution patterns in the mammalian system are fundamentally different.⁴¹ Indeed, CB2R is predominantly found in peripheral organs harboring large numbers of immune cells such as the spleen and tonsils. Further, moderate expression can be found in the cardiovascular system, bone marrow and reproductive organs.^{41, 42,43} Recently, the crystal structures of CB1R and CB2R have been revealed.^{44, 45} Hua et al.⁴⁶ uncovered the structures of the inactivated, intermediate and activated states of CB1R and CB2R, respectively, which provides more accurate molecule models and theoretical supports for drug and ligand discovery.

Due to the diverse regulatory effects by chronic therapeutic exposure, and the presence of genetic splice variants, cannabinoid receptors have unique pharmacological and physiological effects and have been considered as an attractive drug target for over 20 years. CB1R plays an important role in normal physiology (cognition, appetite, memory),⁴⁷ and is closely related to several neuropsychiatric⁴⁸ and metabolic diseases.^{49, 50} CB1R agonists and inverse agonists have been extensively developed, but are accompanied by side effects impairing cognitive function and the ECS reward system.^{51, 52} A well-known example is rimonabant (SR141716A), a potent and selective CB1R inverse agonist developed by Sanofi-Synthelabo.⁵³ It was found to improve lipid and glucose metabolism and became one of the most promising weight-loss drugs. However, rimonabant was later linked to rare but severe cases of depression and suicide, resulting in its withdrawal from the market in 2008.⁵⁴ Currently approved drugs by the FDA include tetrahydrocannabinol, cannabidiol, and their analogs or compositions, such as Marinol^® (dronabinol),^{55, 56} Cesamet^® (nabilone)⁵⁷, Sativex^® (nabiximols)⁵⁸, and Epidiolex^® (cannabidiol) ⁵⁹. CB1 negative allosteric modulators, dronabinol, and nabilone are mainly used to treat nausea, vomiting, and weight loss when traditional treatments are ineffective, and for the treatment of spasticity and pain in MS as combination therapy (nabiximols). Cannabidiol is used to treat epilepsy associated with Lennox-Gastaut Syndrome (LGS) and Dravet Syndrome (DS). In order to avoid CNS side effects, drugs that act on peripheral CB1R alone have been developed and demonstrated promising results. NEO1940 (ART27.13), a peripheral CB1R and CB2R agonist for the treatment of cancer-related anorexia and weight loss, has entered a phase II clinical trial.⁶⁰ CRB-4001(JD-5037) is a peripheral CB1R inverse agonist in a phase I clinical trial, which was intended for the treatment of non-alcoholic steatohepatitis.^{61, 62} CB2R regulation is disrupted in peripheral inflammation, neurological diseases, cancer, and atherosclerosis.^{63, 64} Of note, CB2R activation is associated with anti‐inflammatory and protective effects in the CNS and periphery.⁶⁵ In several neurodegenerative disease models such as AD or PD, treatment with selective CB2R agonists are effective in attenuating pro-inflammatory cytokine release and migratory activity of immune system cells.^{66, 67} Thus the effects of CB2R agonists may offer a new therapeutic approach for the treatment of neurodegenerative disorders, taking advantage of the notion that CB2R activation does not elicit the psychoactive side effects frequently observed with brain penetrating CB1R agonists.^{68, 69}

CB2 agonist JBT-101 (Lenabasum) is currently in a phase III clinical trial for treating diffuse skin systemic sclerosis and dermatomyositis, and phase II clinical trial for treating cystic fibrosis, systemic sclerosis and systemic lupus erythematosus.⁶¹ CB2 agonist APD371 (Olorinab) entered phase II trials for the treatment of gastrointestinal pain.⁷⁰ As mentioned above, the development of PET radiotracers for cannabinoid receptors facilitates our understanding of the ECS under normal physiological and disease states and has played an important role in the design and evaluation promising drugs that target the receptors. Several research groups have made substantial efforts towards the development of CB1R or CB2R PET radioligands, highlighted applications in CNS diseases.

Given the diversity of ECS-targeted drug discovery, the clinical availability of suitable PET probes is crucial to guide drug development as well as to identify and validate novel disease areas, where the ECS can be exploited as a pharmacological target. As such, several research groups have made substantial efforts towards the development of CB1R or CB2R PET radioligands.

2.1. PET Tracers for Imaging CB1R

Endogenous CB1R ligands have moderate affinity and are highly lipophilic, so they are not suitable for PET or SPECT (Single Photon Emission Computed Tomography) tracer development. By reducing their lipophilicity and retaining their high affinity, researchers synthesized a series of compounds for imaging CB1R in the CNS. Based on selective CB1 antagonists, PET tracers have been widely developed and selected structures are shown in Figure 1. A summary of the relevant parameters for selected CB1R specific PET tracers is presented in Table 1.

Figure 1.

Representative chemical structures for CB1R-specific PET tracers

Table 1.

Radiolabeling and PET imaging application of CB1R specific tracers

Radioligand (RCYs)	PET imaging study
	subject	brain uptake
[11c]5 (16±5%a)	Rodents:obesity (myocardial imaging) NHP Human: SZ; obesity (myocardial imaging); alcohol dependence	Mouse: 6.6 %ID/g (striatum, 15 min) Baboon: BP: 1.3 (putamen) Human: VT: 1.35±0.35 Human: SUVmax = 1.75±0.35 (putamen) Human: VT: 0.72±0.07 (putamen)
[11C]6 (2.5%±1.1 %b)	Rodents, NHP Human: SZ	Monkey: VT: 24.3 (mL/m3, striatum) Human: VT: 29.1±17.4 (putamen) Human: SUVmax= 3.5±0.5 (putamen)
[11C]7 (16.5%b)	Rodents, NHP	Rat: high BBB permeability, high specificity to CB1R Monkey: SUVmax = 3.3 (30 min)
[18F]8 (7.92±2.16%b)	Rodents, NHP	Rat: high BBB permeability, high specificly to CB1R Monkey: SUVmax = 2.75±0.75 (15 min)
[18F]9 (5.92%±1.34%b)	Rodents, NHP	Rat: high BBB permeability, high specificity to CB1IR Monkey: SUVmax = 5.75±0.75 (20 min)
[18F]10 (7.93±2.48%b)	Rodents: AD; BAT NHP Human: chronic cannabis smokers; alcohol-dependent; SZ; Psychosis	Mouse: 1.5 %ID/mL (BAT) Monkey: SUVmax = 5.5±1.0 (20 min) Monkey: VT 35.4 (mL/m3, striatum) Human: SUVmax = 3.5±0.5 (putamen) Human: VT 24.3±7.2 (putamen)
[18F]11 (-)	Rodents: HD; PD NHP Human: anorexia and bulimia nervosa; episodic migraine; epilepsy; HD	Rat, Monkey: high BBB permeability, high specificity to CB1R Human: SUVmax = 15±0.2 (putamen) Human: VT 25.1±13.3 (putamen)
[11C]12 (50%a)	Rodents, NHP, Human	Monkey: VT: 2.36 (mL/m3, striatum) Human: SUVmax = 2.35±0.65 (putamen) Human: VT: 3.15±0.70 (putamen)
[11C]15 (7.35±4.25%b)	NHP	Monkey: SUVmax = 2.2±0.5 (120 min)
[18F]16 (3.55±2.05%b)	NHP	Monkey: SUVmax = 1.8 (240 min)
[11C]7 (17±8%)	Rodents: BAT	Mouse: 2.8 %ID/mL (BAT)

^anon-decay corrected RCYs

^bdecay corrected RCYs

The data of PET imaging are listed only in the control group.

Initially, a series of PET tracers ([¹⁸F]1-[¹¹C]4) based on rimonabant were developed for CB1R imaging and preliminary biological evaluations were carried out. However, these tracers had either high lipophilicity (logD > 4), resulting in nonspecific binding, or poor brain permeability that is not suitable for CNS imaging.^71–73 To improve the binding affinity of rimonabant (K_i hCB1R = 46.2 nM, cLogD = 6.0), Fan et al. synthesized OMAR/JHU5528 (K_i hCB1R = 11 ± 7 nM, cLogD = 4.3) and its analogs by introducing cyano groups. OMAR is a CB1R antagonist, demonstrating higher specificity than earlier CB1R antagonists with less lipophilic and significantly increased B_max/K_i ratio to 91, indicating great potential for successful in vivo imaging.^{74, 75} Further PET evaluations in mouse and baboons demonstrated that [¹¹C]OMAR ([¹¹C]5) readily entered the brain, specifically and selectively visualizing cerebral CB1 receptors. In the biodistribution experiment, the radioactive accumulation reached its maximum value (6.6 %ID/g in the striatum) at 15 min; radioactivity in the baboon brain reached its peak at about 10 min and the ratio of CB1-rich regions to CB1-poor regions was up to 2.5 and the highest binding potential (BP) value was 1.3 in the putamen.⁷⁶ Wong et al.^{77, 78} presented the first human studies by using [¹¹C]5, which successfully displayed the distribution of CB1R in the healthy controls and SZ patients. Maximum standard uptake value (SUV_max) in controls reached 1.4–2.1 at about 20 min post-injection and average volume of distributions (V_T) ranged from 1.0–1.7. Compared with control subjects, regional total V_Ts of [¹¹C]5 were higher in SZ patients, especially in the pons (0.93 vs 0.72, P < 0.05). Another study comparing the availability of CB1R between male SZ patients and healthy controls highlighted that CB1R availability of male SZ patients was moderately and globally reduced, and the use of antipsychotics and tobacco increased CB1R availability in patients⁷⁹. In vitro studies have shown that the ECS plays a vital role in alcohol dependence.^80–82 Neumeister et al. used [¹¹C]5 to carry out the first in vivo experiment to demonstrate this correlation. V_T in alcohol-dependent patients administered [¹¹C]5 was approximately 20% (P = 0.023) higher than that of healthy controls.⁸³

Compound [¹¹C]5 was not only used to image CB1R in the CNS, but was also successfully employed to assess peripheral cardiovascular CB1R expression. Valenta et al. demonstrated that the retention rate of [¹¹C]5 in the myocardium of obese mice was significantly higher than that of normal-weight mice.^{84, 85} These results may provide a theoretical basis for further clinical testing of CB1R targeted molecular imaging in cardiac metabolic diseases. However, due to the moderate lipophilicity of [¹¹C]5, it tends to accumulate in the liver, which may hamper accurate quantitative assessment.⁸⁶

Similarly, based on the structure of SR141716, Yasuno et al. reported the synthesis of [¹¹C]MePPEP ([¹¹C]6) (K_i hCB1R = 0.11 nM, K_b = 0.57 nM, LogD_7.4 = 4.8, K_b represents an in vitor efficacy constant that is calculated using the following equation: K_b = IC₅₀/[1 + [Agonist]/EC₅₀], which is derived from Cheng-Prusoff relationship.⁸⁷) by the reaction of the O-desmethyl precursor with [¹¹C]CH₃I. Compared to PET imaging of [¹¹C]5 in baboon brain, the uptake kinetics of the two tracers were similar, but the peak uptake of [¹¹C]6 was higher and SUV_max almost reached 6.0 within 10–20 minutes post injection.⁸⁸ Subsequently, Donohue et al. synthesized [¹¹C]6 and its four analogs [¹¹F]FMePPEP ([¹¹C]7, K_b = 0.22 nM, Log D_7.4 = 5.7), [¹⁸F]FEPEP ([¹⁸F]8, K_b = 0.42 nM, Log D_7.4 = 5.8), [¹⁸F]FMPEP ([¹⁸F]9, K_b = 0.19 nM, Log D_7.4 = 5.7) and [¹⁸F]FMPEP-d₂ ([¹⁸F]10, K_i hCB1R = 0.11 nM, LogD = 4.24),⁸⁹ and demonstrated that [¹¹C]6-7 and [¹⁸F]8-9 crossed the blood-brain barrier adequately and exhibited high specificity to CB1R in rat brain tissues. Terry et al. further evaluated these four imaging agents ([¹¹C]7, [¹⁸F]8–10) in monkeys and SUV_max in the brain were 3.3 by 30 min, 2–3.5 at 15 min, 5–6.5 at 20 min, 4.5–6.5 by 20 min, respectively. And then they advanced the most promising agent [¹⁸F]10 to humans.⁹⁰ [¹⁸F]10 uptake in the human brain was similar to [¹¹C]6, and the peak SUV in the putamen ranged from 3.0 to 4.0, however, the retest variability of [¹¹C]6 in plasma was relatively large (58%), as compared to [¹⁸F]10 (16%), and [¹⁸F]10 had a good retest variability of V_T (14%), which had previously been reported by Terry et al.⁹¹ In subsequent studies, [¹¹C]6 was used as a CB1R imaging agent to study the anxiogenic effects of cannabis in humans, which proved that this effect was closely related to the activation of CB1Rs on the amygdala.⁹² [¹⁸F]10 has been used to demonstrate a reversible downregulation of CB1Rs in chronic cannabis smokers,^{93, 94} and in alcohol-dependent patients,⁹⁵ [¹⁸F]10 demonstrated reduced CB1R binding that did not recover after 2–4 weeks of abstinence. It is worth noting that in CB1R imaging of alcohol-dependent patients with [¹⁸F]10, the results are opposite to the [¹¹C]5 imaging study aforementioned. Differences in experimental design or patient selection may have contributed to these divergent findings. In addition, [¹⁸F]10 has been used to study mental illnesses. Borgan et al. designed a cross-sectional, case-controlled study in patients with first-episode psychosis, demonstrating that CB1 receptors are viable therapeutic targets in psychosis.⁹⁶ Besides CNS-targeted CB1R imaging, [¹⁸F]10 has been used to quantitatively monitor the density of CB1R in brown adipose tissue (BAT). Eriksson et al. and Lahesmaa et al. confirmed that [¹⁸F]10 showed specicific binding to BAT in rodents and humans, respectively.^{97, 98}

Using taranabant as a lead compound, Merck research laboratories synthesized [¹⁸F]MK-9470 ([¹⁸F]11, IC_50(human) CB1R = 0.7 nM, IC_50(human) CB2R = 44 nM), a new potent CB1R inverse agonist, with a high affinity for CB1R.⁹⁹ In vivo brain imaging and autoradiography distribution studies in rhesus monkeys have demonstrated reversible binding to CB1R with reasonable binding specificity. The occupancy of CB1R in the human brain has low Test-retest (TRT) variability (ca. 7%), as determined by the area under curve (AUC) of the respective time-activity curve. However, in vivo radioactive distribution of CB1R is inconsistent with ex vivo findings in the postmortem research, an inconsistency which was suggested to be related to the different properties (agonists and inverse agonists) of radioligands.¹⁰⁰ Subsequently, Van et al. analyzed the distribution of [¹⁸F]11 in 50 healthy individuals.¹⁰¹ In women, as the age increased, radiotracer-receptor binding increased in the basal ganglia, the lateral iliac crest, and the limbic system, especially in the hippocampus. Men do not have these age-related characteristic changes, but men show higher [¹⁸F]11 binding in the limbic system and cortical-thalamic-cortical loop cluster.

Compound [¹⁸F]11 has been used in imaging studies of various neurological diseases: Anorexia nervosa patients demonstrated increased [¹⁸F]11 uptake (24.5%, P = 0.0003) in the brain compared to healthy controls; ¹⁰² a study investigating [¹⁸F]11 in female with episodic migraine showed that there was an increased interictal binding (average gray matter difference+16%, P = 0.009), indicating a low level of competing endogenous cannabinoids in this area;¹⁰³ in patients with epilepsy of temporal lobe due to hippocampal sclerosis, [¹⁸F]11 uptake increased in the ipsilateral temporal lobe.¹⁰⁴ In the premanifest phase of the HD, the PBA-HD (Problem Behavior Assessment for HD) scores inversely correlated with CB1R binding in prefrontal regions and cingulate cortex.¹⁰⁵ Imaging studies of [¹⁸F]11 performed in the brains of PD patients showed that the reduced availability of CB1R is associated with various cognitive impairments¹⁰⁶. Although the CB1R inverse agonist PET tracers highlighted above have been used in clinical research to study receptor availability in several neurological disorders, rimonabant and taranabant development was discontinued in 2008 due to neurological side effects.¹⁰⁷

Donohue et al. developed a new selective CB1 receptor antagonist, [¹¹C]SD5024 ([¹¹C]12, K_i hCB1R = 0.47 nM, LogD_7.4 = 3.79), which demonstrated high specific binding in the cynomolgus monkey brain.¹⁰⁸ The brain uptake reached a peak SUV of 2.0 at 24 min and declined to 1.5 at 90 min. Tsujikawa et al. evaluated this radioligand in vitro and in vivo and compared it with the following ligands: [¹¹C]6 (K_i hCB1R = 0.11 nM, LogD_7.4 = 4.77), [¹¹C]5 (K_i hCB1R = 2.05 nM), [¹⁸F]11 (K_i hCB1R = 0.10 nM) and [¹⁸F]10 (K_i hCB1R = 0.11 nM, LogD_7.4 = 4.24).¹⁰⁹ [¹¹C]12 demonstrated optimal lipophilicity and precise measurement accuracy in terms of intersubject variability (22%) for V_T. It is worth noting that Hjorth et al.¹¹⁰ employed [¹¹C]12 in NHP imaging to obtain the pharmacokinetic profile, calculate the occupancy rate and explore the optimal dose of three novel selective CB1R antagonists developed by AstraZeneca (AZ).¹¹¹ Compound [¹¹C]13, a similar structure to [¹¹C]12, was reported by Donohue et al. in an abstract, and it showed high specificity in the monkey brain,¹¹² however, no further studies have been reported.

WIN 55212–2 is a cannabinoid receptor agonist that belongs to the aminoalkylindoles (AAI) family. This type of compounds has been radiolabeled for PET or SPECT imaging, but exhibited poor brain permeability and specificity.¹¹³ An important example of this compound class is [¹⁸F]14.¹¹⁴ Following these studies, Allen et al. synthesized a new array of aryl sulfonyl-substituted indoles as CB1R inverse agonists.¹¹⁵ Finnema et al. and Donohue et al. radiolabeled PipISB (K_b = 1.5 nM, cLogD = 5.1) with ¹⁸F and ¹¹C, to afford [¹¹C]15 and [¹⁸F]16, respectively, and further evaluated them in rhesus and cynomolgus monkeys. [¹¹C]15 and [¹⁸F]16 crossed the blood–brain barrier (BBB) and exhibited CB1R specific and displaceable binding in monkeys.^{116, 117}

Based on the previous research¹¹⁸, Yamasaki et al.¹¹⁹ modified 1-(4-chlorophenyl)-3-{3-[6-(pyrrolidin-1-yl)pyridin-2-yl]phenyl}urea (PSNCBAM-1) to synthesize compound [¹¹C]17 (K_D = 15.3 μM, LogD = 3.76) and conducted PET imaging where this radiotracer demonstrated high specificity in the mouse BAT (2.8 %ID/mL). Zanato et al.¹²⁰ labeled a fluorinated cannabis analog HU-210F with ¹⁸F ([¹⁸F]18 (K_i hCB1R = 0.98 nM, K_i hCB2R = 3.83 nM) as a potential imaging agent, but further biological evaluation has not been carried out. The latest CB1R imaging probe [¹⁸F]19, was reported by Chang et al.¹²¹ by modifying DBPR211; however, the uptake of this tracer in the mouse brain was low (< 1%).

2.2. PET Tracers for Imaging CB2R

Although CB2R expression is limited in the healthy CNS, there is a consensus that CB2R is upregulated on activated microglial cells upon neuroinflammation.^{63, 122} Thus, the CB2 receptor is regarded as an exceptionally promising target to exploit for neuroinflammation manifested in AD, PD and ALS.¹²³ Despite strenuous efforts by several research groups worldwide, CB2R-targeted PET radioligand development has only moderately progressed in recent years. Reported probes have mainly been hampered by nonspecific binding due to high lipophilicity, metabolic instability, inappropriate pharmacokinetics and lack of selectivity over CB1R or other CNS targets. The search for suitable CB2R PET radioligands is further hampered by the lack of reliable information on CB2R B_max values in the CNS as well as the lack of a highly selective CB2R antibody to confirm autoradiographic findings by histological assessment. Nonetheless, some promising examples with potential for clinical translation have emerged and will be discussed in this chapter. The structural scaffolds for CB2R PET ligands include oxoquinoline, pyridine, thiazole, triazine and oxadiazole derivatives (Figure 2). A summary of the relevant parameters for selected CB2R specific PET tracers is presented in Table 2.

Figure 2.

Representative chemical structures for CB2R-specific PET tracers

Table 2.

Radiolabeling and PET imaging application of CB2R specific tracers

Radioligand (RCYs)	PET imaging study
	subject	brain uptake
[11C]20 (32±11%, 40±12%)	Rodents, NHP Human: AD	Mouse: SUVmax =1.2 (2 min) Rat: SUVmax =1.4 (1.5 min) Monkey: SUVmax = 1.5 (1.5 min) Human:low CB2R availability in AD patients
[11C]21 (33%a)	Rodents: neuroinflammation, atherosclerosis Human: ALS (spinal card tissues), atherosclerosis (specimen )	Rat: high BBB permeability specificly to CB2R
[18F]22 ( ~6%a 12.0±2.7%a)	Rodents: HD Human: ALS (spinal cord tissues)	Rat: not sensitive to the CB2R overexpression Human: no specificly to CB2R
[11C]23 (41.7±8.7%a)	Human: ALS (spinal cord tissues)	Human: specificly to CB2R
[11C]26 (-)	Rodents: neuroinflammation	Rat: specificly to CB2R
[18F]27 (11±4%a)	Rodents Human: ALS (spinal cord tissues)	Rat: SUVmax = 2.7±0.3 (2 min. Spleen) Human: specificly to CB2R
[11C]30 (26±2%b)	Rodents: neuroinflammation, AD, cerebral ischemia	Mice:sensitive to the CB2R overexpression in the AD mice
[18F]31 (0.95±0.65%)	NHP: neuroinflammation	Monkey: SUVmax =0.85±0.15 (brain)

^anon-decay corrected RCYs

^bdecay corrected RCYs

Oxoquinoline derivatives have been exploited for their suitability as CB2R PET radioligands. As such, the most extensively studied candidate in this class of compounds, [¹¹C]NE40 (^[11C]20), was successfully synthesized via O-methylation of the respective phenolic precursor using either [¹¹C]CH₃I or [¹¹C]methyl triflate.^124,125 With a binding affinity (K_i CB2R) of 9.6 nM and a selectivity factor of ~ 100 over CB1R, [¹¹C]20 proved to be specific for CB2R in the rodent spleen, while concurrently exhibiting a fast washout from the CB2R-deficient healthy brain.¹²⁵ Moreover, preclinical evaluation of [¹¹C]20 revealed specific and reversible binding to human CB2R (hCB2R) in a rat model with local hCB₂R overexpression in the right striatum.¹²⁶ In the same study, the authors reported a high brain uptake (SUV_max = 1.5 at 90 s) in rhesus monkeys, confirming that the tracer is able to penetrate the brain in non-human primates. In a subsequent clinical study, [¹¹C]20 did not detect the anticipated CB2R upregulation in the brain of AD patients, thus raising concerns about the suitability of the tracer for CB2R imaging in AD.¹²⁷ Particularly, given the high abundancy of CB1R in the human brain, a selectivity factor of ~ 100 might have been insufficient to avoid off-target binding of [¹¹C]20 to CB1R. Along this line, no relationship was found between regional or global amyloid load and [¹¹C]20 uptakes.¹²⁷

Slavilk et al.¹²⁸ used anisidine and diethyl 2-(ethoxymethylene)malonate to synthesize a series of novel 4-oxoquinoline derivatives based on the structure of KD2 (K_i CB2R = 1.7 nM, logD_7.4 = 3.3)¹²⁹. RS-016 (21) emerged as most suitable candidate, exhibiting an improved affinity (K_i = 0.7 nM) and lipophilicity (LogD_7.4 = 2. 8).¹³⁰ Dynamic brain PET scans of [¹¹C]RS-016 ([¹¹C]21) in a murine neuroinflammation model induced by lipopolysaccharide (LPS) application showed higher CB2R expression level in different brain regions including the cortex, hippocampus, and cerebellum. Further, [¹¹C]21 was used to visualize the CB2R in human and mouse atherosclerotic lesions.¹³⁰ To overcome the limitation of the short half-life of carbon-11, Slavik et al.¹³¹ developed a radiofluorinated analog of [¹¹C]21, codenamed [¹⁸F]RS-126 ([¹⁸F]22). Although the in vitro results with [¹⁸F]22 were encouraging, no CB2R expression was detected in the brain of LPS-treated mice. The latter was attributed to the low specific activity of [¹⁸F]22. In addition, [¹⁸F]22 was rapidly metabolized in mice. Haider et al.¹³² improved the synthesis strategy to obtain higher radiochemical yields (RCYs) and molar activities of [¹⁸F]22. In the transgenic R6 /2 mouse model for HD, [¹⁸F]22 was not sensitive to the CB2R overexpression, although CB2R mRNA upregulation was detected by qPCR in the same tissue. Further, due to high nonspecific neuronal tissue binding, [¹⁸F]22 did not show specific CB2R binding in human ALS spinal cord sections. In sharp contrast, the newly developed hydroxylated derivative, [¹¹C]RS-028 ([¹¹C]23), exhibited a significantly reduced lipophilicity, thus leading to less nonspecific binding and allowing the detection of CB2R upregulation in postmortem human ALS spinal cord tissue. Notably, the spleen uptake of [¹¹C]23 was relatively low compared to its predecessors and was accompanied by a rapid washout in vivo, potentially owing to enzymatic degradation. To investigate the influence of the O-alkyl chain length and oxygen position on metabolic stability, [¹⁸F]AH-040 and [¹⁸F]AH-043 ([¹⁸F]24 and [¹⁸F]25) were synthesized. Although these probes showed improved stability to liver enzymes, their selectivity over CB1R was substantially reduced, impeding their further development.¹³²

Among the reported structural scaffolds of CB2R PET radioligands, 2,5,6-trisubstituted pyridines (Figure 2) have been thoroughly assessed and currently constitute the most promising candidates for clinical translation. Indeed, initial studies with [¹¹C]RSR-056 ([¹¹C]26, Ki hCB2R = 2.5 nM, logD7.4 = 1.94) revealed significantly increased brain uptake following LPS-induced neuroinflammation in mice. Nonetheless, it should be noted that only part of the increased [¹¹C]26 accumulation in the brain was attributed to CB2R-specificity, as evidenced by the weak signal reduction under blockade conditions.¹³³ The latter can be, at least in part, explained by inflammation-induced damage to the blood-brain barrier, which would ultimately result in an increased tracer delivery to the brain, independent of CB2R.¹³⁴ Due to the short physical half-life of carbon-11, which confines the use of [¹¹C]26 to facilities with an on-site cyclotron, the authors synthesized a series of fluorinated derivatives via structural modifications at positions 2 and 5 of the pyridine core. These efforts led to the development of the radiofluorinated analog [¹⁸F]27,¹³⁵ with a prolonged half-life of 109.8 min, a binding affinity of 6 nM and a selectivity factor of ~ 600 over CB1R. [¹⁸F]27 proved to be CB2R-specific by in vitro autoradiography experiments of the rodent spleen. Further, CB2R receptor upregulation in postmortem spinal cord tissues of ALS patients was confirmed with [¹⁸F]27. Notwithstanding the promising in vitro results obtained with [¹⁸F]27, in vivo specificity for the CB2R-rich rodent spleen was only modest. To overcome this limitation, the authors studied the influence of structural modifications at position 6 of the pyridine core, which ultimately led to the discovery of [¹⁸F]RoSMA-18 ([¹⁸F]28), a ligand with subnanomolar affinity (Ki hCB2R = 0.7 nM) and a remarkable selectivity factor of > 12,000.¹³⁶ [¹⁸F]28 revealed exceptional specificities for the CB2R-positive rodent spleen by in vitro autoradiography (71 ± 2%) and ex vivo biodistribution (86 ± 2%), which was superior to any previously reported CB2R-targeted PET probe. Indeed, the outstanding specificity was further elegantly corroborated in the CB2R knockout mouse spleen. [¹⁸F]28 was sensitive in detecting CB2R upregulation in postmortem human ALS spinal cord tissue by in vitro autoradiography. Further, PET experiments confirmed the specific and reversible CB2R binding of [¹⁸F]28 in the CB2-positive rat spleen. The authors found only intact [¹⁸F]28 in the rat brain following intravenous tracer injection. Nonetheless, in vivo defluorination was noted by substantial skull uptake, which was circumvented by replacing the hydrogen atoms in the fluoropropyl side chain with deuterium atoms to afford [¹⁸F]RoSMA-18-d₆ ([¹⁸F]29), in which case no radioactivity accumulation in the skull was observed. [¹⁸F]29 is currently considered a promising candidate for clinical translation.¹³⁶

A-836339 is a highly potent and selective CB2R agonist (K_i CB2R = 0.7 nM, K_i CB1R = 270 nM), which is based on a thiazole core structure¹³⁷. Horti et al. radiolabeled [¹¹C]A-836339 ([¹¹C]30) via its desmethyl precursor and performed PET imaging in LPS-treated CD1 mice, APP/PS1 transgenic mice (mouse model for AD) as well as wild-type control animals.¹³⁷ Whole body time-activity curves (TACs) for the control mice revealed a fast tracer washout from the brain, while CB2R specific binding was observed in the spleen. In LPS-treated mice, [¹¹C]30 uptake was significantly increased in the brain (SUV_max = 2.3), as compared to the control group (SUV_max = 0.5), after five days of LPS treatment. Similarly, higher brain uptake was found in transgenic APP/PS1 strains compared to wild-type littermates, and the regional distribution was consistent with the distribution of Aβ amyloid plaques. AM630 blocking showed 29%−36% specific binding in the cerebellum, brainstem, and cortex. Attempts to develop a radiofluorinated analog of [¹¹C]30 led to an improved CB2R affinity with Ki of 0.4 nM for CB2R and 380 nM for CB1R. Nonetheless, the rapid metabolism of this ¹⁸F analog following intravenous injection as well as the presence of radiometabolites in the brain hampered further evaluation.¹³⁸ Savonenko et al. further investigated the utility of [¹¹C]30 as a PET biomarker for neuroinflammation in AD mice.¹³⁹ The authors found an increased tracer uptake in AD mice, which was blocked by the commercially available CB2R inverse agonist AM630. In sharp contrast, Pottier et al. did not detect CB2R upregulation with [¹¹C]30 PET in two distinct rat models of cerebral ischemia.¹⁴⁰

Triazine scaffolds have been evaluated for CB2R-targeted PET imaging by Hortala et al.¹⁴¹. The authors reported on the development of [¹⁸F]31, demonstrating brain penetration in rhesus macaques and baboons. Baboon pretreatment with LPS led to V_T increase of about 30%, confirming previous observations with other CB2R PET probes in mice. In another study by Yrjola et al., the best performing triazine derivative, [¹⁸F]32 (EC₅₀ = 0.60 nM), was found to suffer from rapid metabolism and elimination, rendering its pharmacokinetic properties unsuitable for in vivo PET.¹⁴²

MA2 and MA3 are the analogs of highly potent CB2 agonist N-arylamide oxadiazole.¹⁴³ Tissue distribution of [¹¹C]33 and [¹⁸F]34 demonstrated high brain uptake at 2 min post injection. Notably, [¹⁸F]34 was found to specifically accumulate in brain regions with overexpression of the human CB2R in rats; however, no specific binding of [¹⁸F] 34 to the CB2 receptor was observed in PET scans in nonhuman primates.¹⁴⁴

3. HYDROLASE TARGETING PET TRACERS

It is known that drugs that act as direct agonists or antagonists on the endocannabinoid receptors can cause drug dependence or addiction.¹⁴⁵ Hence, as an alternative approach to achieve therapeutic effects, drugs/tool compounds that inhibit the hydrolases that metabolize AEA and 2-AG have been developed. The two main hydrolases involved in the hydrolytic metabolism of these endocannabinoids are FAAH and MAGL. FAAH is a 60kDa integral membrane hydrolytic enzyme that is highly expressed in the mammalian brain along with liver and kidney, which is responsible for terminating the endogenous cannabinoid signaling of AEA by hydrolysis of the amide bond.^{146, 147} MAGL, is a 33kDa serine hydrolase with high expression in the brain, BAT, adrenal glands, liver and kidneys, and acts as a critical degradative enzyme responsible for ca. 85% of 2-AG metabolism in the brain, generating glycerol and arachidonic acid (AA).^148–151 As shown in Scheme 2, systemic or local inhibition of FAAH or MAGL increase levels of AEA or 2-AG, while concurrently decreasing arachidonic acid level, respectively. These changes also affect endocannabinoid signaling and suppresses eicosanoid production. Increasing levels of AEA exert effects of anti-anxiety, anti-depression, and treating addiction and withdrawal symptoms.^{152, 153} Similarly, increasing the level of 2-AG in the brain has effects of anti-anxiolytic, anti-depression, anti-emetic, anti-inflammation and neuroprotection.^{154, 155} Moreover, inhibition of MAGL (genetically/pharmacologically) has therapeutic implications for some metabolic diseases.^{156, 157} Numerous studies have demonstrated that FAAH and MAGL are potential therapeutic targets for a variety of human pathologies including pain, depression, (neuro)inflammatory diseases and cancer.^{149, 155, 158–160} For several of these indications proof-of-concept studies have been performed in several animal models with the respective knockout animals.^{8, 157, 161} It is worthy of note that Scheme 2 only captures main features of FAAH and MAGL in the ECS since both FAAH and MAGL can metabolize other related lipids: in the case of FAAH, blockade of its hydrolysis activity for N-palmitoylethanolamine is an interesting pharmacological approach, given the anti-inflammatory effects of this endogenous lipid.¹⁶² In the case of MAGL, it has shown that that it promotes cancer cell growth by providing long chain fatty acids from their corresponding monoacylglycerols,^{163, 164} and to provide arachidonic acid for prostaglandin synthesis in neuroinflammation.¹⁵⁵

Scheme 2.

Pharmacological properties of FAAH and MALG inhibitors

Currently, a variety of FAAH and MAGL inhibitors have been developed, characterized in animal models of disease and assessed at various stages of clinical trials.^165–167 Initial clinical studies with FAAH inhibitors PF-04457845,¹⁶⁸ V158866¹⁶⁹ and JNJ-42165279¹⁷⁰ indicated that the compounds were well-tolerated (vide infra for a discussion with respect to FAAH inhibitor BIA10–2474¹⁷¹), although PF-04457845 was not efficacious in a study of osteoarthritic pain. Recently, Fazio et al. reviewed newly developed single- and dual-target FAAH inhibitors in the literature, and proposed that multi-target inhibition may be a promising strategy for the development of new effective inhibitors. Future drug design and medicinal chemistry could also take advantage of the recently disclosed human FAAH crystal structure.¹⁷² Among MAGL inhibitors, ABX-1431 has entered clinical phase II studies for neurological disorders. In this context, PET ligand development towards FAAH and MAGL has also made considerable progress for target engagement studies in drug discovery and studying pathophysiological changes during disease states.

3.1. FAAH Inhibitor based PET Tracers

A major clinical trial incident occurred in January 2016,¹⁷³ which led to the death of one of the participants during the study with a FAAH inhibitor codenamed BIA 10–2474. Following the incident, there has been a solid body of evidence that the drug exhibited off-target effect.^174–177 The tragic results of this trial emphasized that it is especially important to fully understand the pharmacological effects, metabolism and pharmacokinetic-pharmacodynamic (PK-PD) relationship of drugs in the body before conducting clinical trials. Using PET probes are uniquely well suited for in vivo drug evaluation to assess target engagement and correlate therapeutic dose-target occupancy relationship accurately in humans via a phase 0 study. Currently, several PET tracers for imaging FAAH have been disclosed and can be classified into two categories, namely irreversible and reversible tracers, based on different binding kinetics, as shown in Figure 3. We also summarize the pharmacological and imaging properties of FAAH PET tracers in Table 3.

Figure 3.

Representative chemical structures for FAAH-specific PET tracers

Table 3.

Radiolabeling and PET imaging application of FAAH specific tracers

Radioligand (RCYs)	PET imaging study
	subject	brain uptake
[11C]36 (4%a)	Rodents, Human: cannabis addiction, alcohol use disorders, psychiatric, borderline personality disorder	Rat: SUVmax = 2±0.4 Human: SUVmax = 4.3±0.8 (putamen)
[11C]43 (20±4.6%b)	Rodents, NHP	Rat: SUVmax ~ 1.5 (cerebellar nucleus) Monkey: SUVmax = 2.0 (occipital cortex)
[11C]44 (19.5±2.5%b)	Rodents, NHP	Rat: SUVmax = 4.55±0.57 (cortex) Baboon: SUVmax=2.1 (whole brain,55 min)
[11C]45 (15±5%b)	Rodents	Rat: SUVmax = 7.8±0.1 (cortex,40 min)
[11C]46 (4.5±1.3%)	Rodents	Rat: SUVmax = 2.8±1.6
[11C]47 (10.7±4.4%b)	Rodents	SUV ~ 0.8 (cortex, 90 min)
[11C]50 (-)	NHP, Human	Monkey: BP: 0.7±0.3, VT: 3.5±0.5 Monkey: SUVmax >1 Human: VT: 17±3
[18F]51 (13%a)	Rodents	Mice: SUVmax =0.87 (cerebellar nuclei, 1.5 min)
[18F]52 (17±6%b)	Rodents	9.2 %ID/g

^anon-decay corrected RCYs

^bdecay corrected RCYs

The data of PET imaging are listed only in the control group.

3.1.1. Irreversible FAAH Radiotracers

URB597 (IC₅₀ = 4.6 nM, cLogD = 3.6) is a classic FAAH inhibitor based on O-aryl carbamate scaffold with potentially useful activity in an animal model of anxiety.^178–181 A ¹¹C-labeled URB597 analogue, biphenyl-3-yl [¹¹C]-4-methoxyphenylcarbamate ([¹¹C]35) was the first radiolabeled FAAH inhibitor for PET imaging studies in rodents.¹⁸² Although the LogD value (2.27) was suitable for brain penetration, a continuous washout was observed after rapid peak brain uptake, which indicated no covalent binding to FAAH.¹⁸² However, it represented a reasonable starting point for the design of radiolabeled FAAH inhibitors.

CURB (IC₅₀ = 30 nM, LogD_7.4 = 2.8), also named URB694, is an active and selective irreversible inhibitor of FAAH with improved brain penetration and plasma stability, as compared to URB597.¹⁸³ As such, [¹¹C]CURB ([¹¹C]36) was the first reported successful irreversible PET tracer to image FAAH.^{184, 185} [¹¹C]36 displayed similar ex vivo biodistribution to FAAH in the rat brain, with higher brain uptake in the cortex and lower signal in the hypothalamus. The radioactivity peaked at 5 min (SUV_max = 1.6–2.4) and no significant washout was observed, consistent with the suggested irreversible binding mechanism. High dose pretreatment using URB597 showed high specificity of binding. A polar metabolite was detected in blood, which has little effect on the imaging results as it is not blood-brain barrier permeable.¹⁸⁵ Based on these promising preclinical studies, [¹¹C]36 became the first PET tracer for mapping FAAH activity in the human brain.¹⁸⁶ By performing 90-minute PET scans with arterial blood sampling on six healthy volunteers, λk₃ was identified as the optimal parameter for quantifying FAAH binding and SUV reached its peak (4.3±0.8) 30 min post-injection. For the kinetic modeling of irreversible inhibitors, λk₃ can be calculated by the equation K₁ / k₂ * k₃ and constitutes a key parameter of kinetic modeling for two-tissue compartment models. Of note, λ is the ratio between the two rate constants, K₁ and k₂, and represents the equilibrium distribution volume of the ligand in the plasma and the organ of interest – in this case the brain. Moreover, k₃ describes the influx of [¹¹C]36 from the free and nonspecific compartment to the specifically bound compartment.¹⁸⁷ Recently, [¹¹C]36 has been applied in neuroimaging application of several neurological disorders, particularly in addictive diseases, such as cannabis addiction,^{188, 189} alcohol use disorders,¹⁹⁰ psychiatric¹⁹¹ and borderline personality disorder,¹⁹² which further proves that FAAH could be a promising therapeutic target for a variety of CNS diseases.

Moreover, a series of [¹¹C-carbonyl]-radiolabeled O-aryl carbamates were synthesized ([¹¹C]37-[¹¹C]42) and evaluated as radiotracers for imaging FAAH with PET by Wilson et al.¹⁹³ All six radiotracers showed moderate to excellent brain uptake, good regional distribution (cortex > hypothalamus) and excellent specificity (80–95% blocking via pretreatments with URB597) for FAAH in the rat brain. Brain kinetics showed that [¹¹C]dihydrooxazole carbamates ([¹¹C]40–42) had much higher uptake (3.0–4.5 SUV in the cortex at 2 min) than [¹¹C]biphenyl carbamates ([¹¹C]36–39, 0.8–2.7 SUV in the cortex at 2 min).

2-Methylpyridin-3-yl-4-(5-(2-fluorophenyl)-4H-1,2,4-triazol-3-yl)piperidine-1-carboxylate (MFTC) is a high-affinity, selective FAAH ligand (IC₅₀ = 0.34 nM for human FAAH)¹⁹⁴ with moderate lipophilicity (cLogD = 3.6). In vitro studies demonstrated a similar binding affinity for FAAH as URB597 and CURB.¹⁸⁵ [¹¹C]MFTC ([¹¹C])43) was successfully synthesized using bis(2-methylpyridin-3-yl)-[¹¹C]carbonate as a new radioactive intermediate.¹⁹⁵ PET studies of this ligand in rat brain demonstrated high uptake (SUV_max ~1.5 in the cerebellar nucleeus), and the distribution of radioactivity corresponded to FAAH expression. The accumulation of [¹¹C]43 in the brain was successfully blocked by pretreatment with nonradioactive MFTC and URB597. Furthermore, PET imaging in monkeys brain demonstrated heterogeneous distribution and radioactivity accumulation was highest in the occipital cortex (SUV_max= 2.0); however, [¹¹C]43 did not show high specific binding in the striatum, which was in contrast to [¹¹C]36 and [¹¹C]MK-3168 (a reversible inhibitor of FAAH, vide infra). In addition, NHP PET imaging showed that the tracer was gradually washed out in all regions and no plateau was reached during the 90 min scan period, suggesting that total binding in the monkey brain might be affected by nonspecific binding to some extent.

¹⁸F-Labeled PET tracers based on the dihydrooxazole scaffold have also been developed. [¹⁸F]DOPP ([¹⁸F]44) was synthesized via a multi-step automated synthesis, as a potent, selective FAAH ligand with an IC₅₀ value of 0.82 nM for rat FAAH and a logD value of 2.33. In vitro assays showed that DOPP inhibited FAAH in a time-dependent manner. Ex vivo biodistribution studies demonstrated excellent brain uptake (> 4 SUV) and a distribution pattern consistent with FAAH patterns in the rat brain. Autoradiography of [¹⁸F]44 in the rat brain also showed high bound activity in FAAH rich regions. In addition, heterogeneous binding of [¹⁸F]44 was displaced by pretreatment with the FAAH inhibitors CURB or PF-04457845, indicating that the radiotracer bound specifically to FAAH in vitro.¹⁹⁶ PET/magnetic resonance imaging (MRI) studies in baboons demonstrated that [¹⁸F]44 rapidly crossed the blood-brain barrier with SUV_peak of 2.1 and that the kinetic modeling parameter λk₃ was appropirate for FAAH quantification.¹⁹⁷ As a proof-of-concept, [¹⁸F]44 was employed to evaluate the potency of BIA 10–2472 in vivo in the brain, which may be useful to provide a rational approach to dose selection in clinical trials and avoid severe adverse events.¹⁹⁸

Another ¹⁸F-labeled FAAH radiotracer, [¹⁸F]FCHC ([¹⁸F]45, IC₅₀ = 1.2 nM), was also evaluated in rodents.¹⁹⁹ Similar to previous results,¹⁹³ ([¹⁸F]45 demonstrated high brain penetration (SUV_peak = 4.6), excellent specificity and appropriate biodistribution in rodents, warranting further evaluation in higher species.

Two urea-based irreversible tracers [¹¹C]PF-04457845 ([¹¹C]46, IC₅₀ = 3.2 nM, LogD_7.4 = 3.48) and [¹⁸F]PF-9811 ([¹⁸F]47, IC₅₀ = 16 nM, LogP = 3.14) have been developed. The compounds interacted with FAAH in an analogous irreversible mechanism like other carbamate-based inhibitors.²⁰⁰Due to its high brain penetration, high selectivity and safety profile in humans,²⁰¹ PF-04457845 was chosen as a promising candidate to be labeled with ¹¹C via the [¹¹C]CO₂ fixation strategy.^{200, 202} Ex vivo biodistribution of [¹¹C]PF-04457845 ([¹¹C]46) in the rat brain was consistent with the known expression of FAAH and the highest uptake (4.4 SUV) was observed in the cortex. In addition, pretreatment with URB597 resulted in a 71–81% reduction of bound activity in all areas of the brain, demonstrating high specificity for FAAH.²⁰³ [¹⁸F]PF-9811 ([¹⁸F]47) was designed based on the scaffold of PF-04457845, in which the trifluoromethyl group was replaced by a fluoroethoxy functionality to facilitate the incorporation of a fluorine-18 radiolabel.²⁰⁴ Biodistribution experiments in rats showed good heterogeneous brain uptake, with high SUVs in the cortex, hippocampus and cerebellum (ca. 0.8 SUV at 90 min), and specific to nonspecific binding ratios of 2.3 to 2.6. Pretreatment experiments proved irreversible binding in the brain. After 90 minutes, the radioactivity of all brain regions decreased to 37%−73%, but the signal in the blood increased by a factor of 12. The authors claimed that this may have been related to irreversible binding in the brain and peripheral organs.

Dahl et al.^{205, 206} introduced a new and efficient fully automated [¹¹C]CO₂-fixation apparatus, and used this method to synthesize two FAAH radiotracers [¹¹C]JNJ1661010 ([¹¹C]48, IC₅₀ = 12 nM) and [¹¹C]49, with significant increased non-isolated RCY. This method can help to produce various ¹¹C-carbonyl compounds in a highly efficient manner to facilitate rapid screening of potential FAAH PET ligands in vivo.

3.1.2. Reversible FAAH Radiotracers

Based on the binding mechanism, tissue uptake of irreversible PET tracers is frequently governed by perfusion, which is why they are described as flow limited tracers. As such, a flow limited PET tracer with irreversible binding mechanism may reveal information that is substantially confounded by organ perfusion, rather than reflecting the density/activity of the biological target. Indeed, kinetic parameter from in vivo PET quantifications, including K₁, k₂, k₃, V_T and BP_ND can be affected by the binding mechanism of a given tracer.^{207, 208} Examples of irreversible PET tracers that were shown to be spared from significant flow limiting effects include [¹¹C]CURB¹⁸⁴ and [¹⁸F]DOPP.¹⁹⁷ Thus, recent research has focused on the development of reversible binding PET tracers for FAAH in order to obtain a comprehensive quantitative assessment of FAAH distribution, concentration, and availability in the brain. Currently, a limited number of reversible FAAH inhibitor PET probes have been disclosed based on pyrazole or α-ketoheterocyclic scaffolds.

A systematical evaluation of a library of FAAH inhibitors based on a pyrazole scaffold was conducted at Merck. Attributed to excellent potency (IC_50(human) = 1.0 ± 0.6 nM) and selectivity, reasonable lipophilicity (LogD = 3.3), and amenability for radiolabeling, [¹¹C]MK-3168 ([¹¹C]50) was prepared as the first reported reversible PET tracer for imaging FAAH. In vivo PET imaging studies were performed under baseline and blocking conditions in rhesus monkeys, demonstrating consistent regional heterogeneous distribution (accumulation in the frontal cortex / striatum / hippocampus regions), good brain uptake (SUV > 1) and high specific binding (total/ nonspecific signal ~ 2:1).²⁰⁹ According to an abstract, the two tissue compartment model well simulated the [¹¹C]50 TAC in monkey and human, and values of binding potential (BP_ND) and V_T were 0.4–1 and 3–4 in the monkey respectively and the V_T value in the human between 14 and 20. ²¹⁰ [¹¹C]50 has been applied to determine the optimal dose for FAAH inhibitior JNJ-42165279 in the human brain.²¹¹

FAAH inhibitors based on an α-ketoheterocyclic scaffold were comprehensively studied by Boger and co-workers.^{212, 213} These inhibitors exhibited reversible binding towards FAAH and other serine hydrolases. Based on the structure of OL-135 (a potent and selective FAAH inhibitor, K_i = 4.7 nM),²¹⁴ Wang et al. designed two positron-emitting analogs, ¹⁸F- and ¹¹C-labeled MPPO ([¹¹C]51 represents the ¹¹C-analog), with communal pyridinyl oxazole ring.²¹⁵ PET imaging of ¹⁸F-analog showed rapid defluorination and high bone uptake in vivo, and failed to show sufficient brain permeability. MPPO, which exhibited an excellent potency (IC₅₀ = 10 nM, K_i = 5.8 nM) to FAAH and selectivity over other enzymes,²¹⁴ was selected as a candidate as it was amenable for ¹¹C-labeling. The ex vivo biodistribution studies in mice indicated that the distribution of [¹¹C]51 was consistent with the distribution of FAAH. However, only moderate brain uptake (0.8 SUV) was observed. PET imaging studies showed low-to-moderate specific binding to FAAH. Subsequent PET imaging in P-glycoprotein (PgP) / breast cancer resistance protein (Bcrp) knockout mice indicated that this tracer was not likely a substrate for Pgp/Bcrp. It should be noted here that [¹¹C]51 showed a steady washout in the brain, indicating α-ketoheterocyclic scaffold indeed provided a roadmap for further investigation of reversible FAAH radiotracers.

Recently, Chen et al.²¹⁶ discovered an improved FAAH inhibitor with a novel heterocyclic scaffold and the corresponding PET tracer [¹¹C]FAAH-1906 ([¹¹C]52, IC_50(mice) = 11 nM, LogD = 2.9) was further studied. Ex vivo biodistribution studies demonstrated that [¹¹C]52 possessed high brain penetration (9.2 %ID/g), reversible binding kinetics and moderate specificity. Further development of this tracer and close analogs is underway to validate this class of compounds in higher species.

3.2. MAGL Inhibitor based PET Tracers

MAGL is responsible for about 85% of 2-AG hydrolysis in the brain,^{5, 217, 218} and significant efforts have been contributed to drug discovery for MAGL inhibitors. So far, most MAGL inhibitors are irreversible based on carbamate and urea-bearing scaffolds, involving the formation of a covalent bond between the Ser122 of MAGL and the carbonyl group of the inhibitor. Recently, a limited number of reversible MAGL inhibitors were disclosed with high binding affinity, and advanced to potential PET tracers for imaging quantification ([¹¹C]61-64, Figure 4). We also summarize the pharmacological and imaging properties of MAGL PET tracers in Table 4.

Figure 4.

Representative chemical structures of MAGL-specific PET tracers

Table 4.

Radiolabeling and PET imaging application of MAGL specific tracers

Radioligand (RCYs)	PET imaging study
	subject	brain uptake
[11C]53 (4.5±1.5%a, 20%b)	Rodents, NHP	Rat: SUVmax= 1.75 (cerebral cortex and striatum) Monkey: SUVmax= 0.93 (occipital cortex)
[11C]54 (19%b)	Rodents	Rat: SUVmax = 0.43 (cerebellum)
[11C]55 (40%)	Rodents, NHP	Mice, Rat: high BBB permeability, specificity to MAGL Monkey: SUVmax ~4.2, 2.5 min
[11C]56 (11%)	Rodents, NHP	Rat: SUVmax >1.5 Monkey: SUVmax =1.7 (occipital cortex)
[11C]57 (2.5±0.4%b)	Rodents	Rat: SUVmax = 2.26 (3.5 min)
[11C]58 (15.0±6.8%b)	Rodents	Rat: SUVmax >1.5
[11C]59 (25%)	Rodents, NHP	Rat: high BBB permeability, specificity to MAGL Monkey: SUVmax ~3.5 (40 min)
[18F]60 (8%a)	Rodents	Mice: SUVmax = 1.75 (2.5 min)
[18F]63(13.0%)	Rodents, BAT	BAT: 21.4 %ID/g
[18F]64(21,7±3.0%b)	Rodents, NHP	Mice: 1.0±0.1% ID/g (biod,1 min) Monkey: SUVmax >1.5

^anon-decay corrected RCYs

^bdecay corrected RCYs

3.2.1. Irreversible MAGL Radiotracers

The first reported MAGL PET probe was described by Hicks et al. in 2014²¹⁹, where carbamate based inhibitors ([¹¹C]JW642, [¹¹C]KML29) and two urea based inhibitor ([¹¹C]ML30, [¹¹C]JJKK-0048) were synthesized. However, all these inhibitors showed poor brain penetration and were not suitable to advance further PET imaging studies.

SAR127303 is a highly potent and selective carbamate-based inhibitor of MAGL developed by Sanofi, and the values of IC₅₀ in mouse and human MAGL are 3.8 and 29 nM, respectively. SAR127303 has shown therapeutic effect on relieving inflammatory pain and epilepsy symptoms in mice.²²⁰ In 2016, Wang et al.²²¹ labeled [¹¹C]SAR127303 ([¹¹C]53) with [¹¹C]CO₂ as a novel MAGL imaging agent in rodents. PET imaging of [¹¹C]53 in rats showed 48% reduction during blocking conditions. Independently, Wang et al. ²²² screened MAGL inhibitors from a series of sulfonamido-based carbamates and urea-based compounds. Among them, the most potent SAR127303 (IC₅₀ = 39.3 nM) and TZPU (IC₅₀ = 35.9 nM) were labeled with [¹¹C]COCl₂ to afford [¹¹C]53 and [¹¹C]TZPU ([¹¹C]54). [¹¹C]53 showed a high level of brain permeability (1.5 SUV for rats and ca.1 SUV for NHP) and heterogeneous regional brain distribution which was consistent with the distribution of MAGL in the brain. The uptake of radioactivity was almost four times higher than [¹¹C]54 in the rat brain.²²² To further assess PET quantification of [¹¹C]53 in the rat brain, Yamasaki et al.¹⁸⁷ successfully used an irreversible two-tissue compartment model (2-TCMi, k₄ = 0) to quantify the concentration distribution of MAGL in different brain regions. Among the kinetic parameters (k₃, K_i, and λk₃) they have obtained, the correlation between Ki value and in vitro binding of [¹¹C]53 was the highest.¹⁸⁷

In 2017, Ahamed et al.²²³ reported on a novel MAGL PET tracer, [¹¹C]MA-PB-1 ([¹¹C]55), which was modified from JW642 by replacing the phenoxy group with a methoxy group. The initial uptake of [¹¹C]55 in mouse and rat brain was comparable with that of [¹¹C]53. However, significant washout was observed after reaching peak uptake in the rat brain, and the radioactivity remained stable after 1h. One explanation could be species differences postulated by the authors. Self-blocking experiments and chasing studies with MJN110 (a structurally distinct MAGL inhibitor) confirmed [¹¹C]55 as a specific irreversible MAGL tracer. In rhesus monkey PET imaging of [¹¹C]55, initial high brain uptake reached (~4.2 SUV) at 2.5 min and a steady state (2.5 SUV) reached after 20 min. This result further underlined the irreversible binding characteristics of [¹¹C]55.

In the same year, Butler et al. disclosed a series of MAGL inhibitors bearing an azetidine scaffold to reduce the lipophilicity and increase the binding specificity.²²⁴ Subsequently, Cheng et al. further developed an array of azetidinyl carbamate-based MAGL inhibitors. Among them, the most promising MAGL-0519 was labeled with [¹¹C]COCl₂ or [¹¹C]CH₃I at two different positions and evaluated in rodents and rhesus monkeys.²²⁵ In rodent PET imaging, [¹¹C]MAGL-0519 ([¹¹C]56) had a high binding specificity and selectivity as well as a heterogeneous distribution with high brain uptake (SUV_max is up to 1.5), which was consistent with the distribution of MAGL. An unexpected clearance of [¹¹C]56 in rat brain may be caused by the metabolism of the linkage between carbonyl group of the tracer and the serine residue of MAGL in vivo. In PET imaging studies with rhesus monkey, [¹¹C]56 also had a heterogeneous distribution with peak brain uptake greater than 1.5 SUV as well as an unexpected clearance.

In 2019, Chen et al²²⁶ developed a novel array of reversible and irreversible MAGL inhibitors with a “tail switching” strategy on a piperazinyl azetidine scaffold.^227–229 Among them, three promising MAGL inhibitors were labeled and evaluated in rodents, including the irreversible inhibitor [¹¹C]MAGL-2–11 ([¹¹C]57, IC₅₀ = 0.88 nM, LogD = 1.9) and two reversible inhibitors [¹⁸F]MAGL-4–11 ([¹⁸F]61, IC₅₀ = 11.7 nM, LogD = 2.7) and [¹¹C]PAD ([¹¹C]62, IC₅₀ = 2.7 nM, LogD = 3.4). [¹¹C]57 showed good brain permeability with a maximum brain SUV of 2.26 at 3.5 min post injection and no significant washout was observed during PET scan. The distribution of [¹¹C]57 in the rat brain was heterogeneous and consistent with MAGL distribution. However, BBB permeability of two reversible tracers ([¹¹C]PAD ([¹¹C]62) and [¹⁸F]MAGL-4–11 ([¹⁸F]61)) was poor (0.3–0.4 SUV_max), which may be related to Pgp and Bcrp efflux. This hypothesis was validated in an experiment with Pgp/Bcrp knockout mice. Further studied in higher species are underway to verify if species difference exists for brain penetration.

In the same year, Mori et al.²³⁰ reported a novel irreversible MAGL inhibitor [¹¹C]58 based on azetidine scaffold with a hexafluoroisopropyl (HFIP) alcohol leaving group. [¹¹C]58 has a high MAGL potency (IC₅₀ = 0.4 nM) and a suitable lipophilicity (cLogD=3.29). [¹¹C]58 was labeled with [¹¹C]COCl₂ and evaluated in rodents. [¹¹C]58 exhibited a good brain permeability with a maximum brain uptake of 1.5 SUV and specific and irreversible binding to MAGL.

Recently, an azabicyclo[3.1.0]hexane scaffold-based irreversible MAGL inhibitor, PF-06809247 with a unique trifluoromethyl glycol leaving group was labeled with carbon-11 to assess its potential as a MAGL radioligand both in vitro and in vivo.²³¹ [¹¹C]PF-06809247 ([¹¹C]59) exhibited good MAGL potency (IC₅₀ = 13 nM, cLogD = 2.7), high permeability and high specific binding in the frontal cortex, hippocampus, and cerebellum in the brain. In NHP PET imaging, [¹¹C]59 showed high brain uptake with SUV_max ca. 3.5. Although its target occupancy and pharmacokinetics in NHP have not yet been published, [¹¹C]59 has entered clinical trials.²³² This will provide the basis for the development of MAGL targeting therapeutics and further understanding of MAGL pathophysiological changes in vivo.

In parallel, Chen et al.²³³ reported another azabicyclo[3.1.0]hexane scaffold-based MAGL PET tracer [¹⁸F]PF06795071 ([¹⁸F]60). PF06795071 is a promising MAGL inhibitor developed by McAllister et al. with excellent MAGL potency (IC₅₀ = 3 nM) and excellent serine hydrolase selectivity.²³⁴ [¹⁸F]60 was ¹⁸F-labeled via spirocyclic iodonium ylide (SCIDY) strategy^235–238 and further evaluated in rodents. By in vitro autoradiography, [¹⁸F]60 exhibited excellent binding specificity and heterogeneous regional brain uptake in rodent brain, consistent with MAGL distribution. In addition, [¹⁸F]60 readily crossed the BBB (SUV_max = 1.8 at 2.5 min) and exhibited heterogeneous regional brain uptake in rat PET imaging. PET evaluation in NHP is underway to verify its translation potential for human use.

3.2.2. Reversible MAGL Radiotracers

A limited number of reversible MAGL radiotracers have been reported as shown in Figure 4. [¹⁸F]61 and [¹¹C]62 with limited brain penetration have already been mentioned above, which could be attributed to species difference.²²⁶ Subsequently, Chen et al.²³⁹ developed a novel reversible MAGL PET probe, [¹⁸F]FEPAD ([¹⁸F]63), based on the structure of [¹¹C]62. [¹⁸F]63 showed excellent MAGL potency (IC₅₀ = 23.8 nM) and target selectivity (> 100 fold) among any other major serine hydrolases. As a proof-of-concept, [¹⁸F]63 was used to differentiate brown and white adipose tissue in the lipid network. Indeed, [¹⁸F]63 had much higher uptake in BAT (peak value 21.4 %ID/g) than white adipose tissue (WAT, 3.5 %ID/g). Recently Hattori et al.²⁴⁰ developed a novel MAGL PET tracer [¹⁸F]T-401 ([¹⁸F]64, IC_50(human) = 4.0 nM, LogD = 1.86) based on the piperazinyl pyrrolidine-2-one scaffold reported by Takeda pharmaceuticals. [¹⁸F]64 exhibited the highest uptake in the frontal cortex, followed by the striatum, and these results were consistent with the distribution of MAGL. The rhesus monkey PET imaging with [¹⁸F]64 exhibited reversible binding characteristics, and brain uptake of [¹⁸F]64 peaked immediately (SUV_max > 1.5) post injection, and subsequently decreased to 30–40% of the maximum at the end of 120 min scan.

4. PERSPECTIVES AND CONCLUSIONS

4.1. Opportunities in developing ECS PET tracers based on allosteric modulation

CB1R ligand development has proven challenging due to severe psychoactive adverse events observed with the majority of orthosteric CB1R ligands. The latter has prompted the focus towards allosteric CB1R modulators, providing an enhanced subtype-selectivity and reducing receptor downregulation as well as inter-receptor promiscuity – two detrimental effects observed with classical orthosteric CB1R ligands.²⁴¹ Both CB1R and CB2R contain several allosteric binding sites.^{242, 243} Over the past few years, a number of synthetic or natural compounds were characterized as CB1R allosteric modulators.^244–247 A detailed review on the pharmacological properties and therapeutic potential of these allosteric modulators has recently been reported by Khurana et al.²⁴⁸ Similarly, by computational modeling and molecular dynamics simulation, Dainese et al.²⁴⁹demonstrated that FAAH function can be fine-tuned by allosteric modulation. This discovery provided new concepts for the regulation of FAAH activity and the synthesis of allosteric FAAH inhibitors. Notably, the contemporary focus on allosteric modulators of the ECS has implications not only for therapeutic strategies but also for the development of ECS-targeted PET radioligands. Given that PET probes are crucial to facilitate drug development via target validation experiments, receptor occupancy and proper dose-finding studies, it is not surprising that there is a growing interest in radioligands that bind to allosteric sites of CB1R and CB2R. Indeed, allosteric binding sites are yet to be exploited in future tracer development programs and harbor a huge potential to improve our understanding of the ECS and to facilitate drug discovery on allosteric modulation.

4.2. Opportunities in developing PET tracers for important yet understudied biological targets within ECS

Other serine hydrolases, in addition to FAAH and MAGL, are involved in the biosynthesis and metabolism of endocannabinoids and other lipids. For some of these serine hydrolases, covalent and/or reversible inhibitors have been reported, which could potentially provide a starting point for developing PET tracers. For example, ABHD12 and ABHD6 have been shown to contribute to total brain 2-AG hydrolysis activity and may make more specialized contributions to this process, especially in cell types with low MAGL expression. In this regard, ABHD6 and ABHD12 both show distinct cellular and subcellular expression patterns compared to MAGL and may regulate different 2-AG pools.^{7, 250} Development of PET tracers for ABHD6 and ABHD12 may facilitate an understanding of the potentially distinct roles played by these enzymes in 2-AG metabolism and signaling. There are several highly selective inhibitors available for ABHD6 and ABHD12 such as KT182, an orally active irreversible ABHD6 inhibitor²⁵¹, and DO264 – an orally active reversible ABHD12 inhibitor.²⁵²

The integral membrane serine hydrolases diacylglycerol lipase α (DAGLα) and DAGLβ have been identified as DAG lipases involved in the biosynthesis of 2-AG.²⁵³ Different isoform-selective and non-selective DAGL inhibitors have been reported, including KT109 (DAGLβ isoform-selective, non-CNS active),²⁵⁴ and DO34 and DH376 (DAGLα/β dual, CNS active).²⁵⁵ Beyond serine hydrolases, other lipases of interest include NAPE-PLD, which converts NAPEs into NAEs and is involved in the biosynthesis of AEA. Recently, a CNS-active NAPE-PLD inhibitor, LEI-401, has been reported and characterized in mice.²⁵⁶

Development of PET-tracers based on inhibitors of key enzymes involved in eCB metabolism is an important step in the development and characterization of new drug candidates for ECS-related diseases.

5. Summary

During the past decades, a variety of CB1R radioligands have been synthesized and evaluated. Several tracers have been transitioned into clinical research studies, of which [¹¹C]5 and [¹⁸F]10 have gained promising results in mental disorders, such as schizophrenia and addiction disorder. PET radioligands for CB2R have shown promise in neuroinflammatory animal models and postmortem human tissues of ALS patients. It should be noted that CB2R tracer development has been plagued by the lack of a suitable antibody/ligand, which would allow the validation of autoradiographic findings and determination of CB2R B_max in the CNS. Now that the crystal structure of CB2R has been resolved, it is envisioned that computational modeling may aid the development of a selective antibody/ligand, thereby facilitating appropriate validation of PET probes in the pipeline.⁴⁵ Since a major clinical trial tragedy in 2016, few attempts have been made to develop new FAAH radiotracers, which represents a current unmet clinical need. Conversely, the development of PET probes for MAGL has recently received growing attention with a plethora of structurally distinct radiolabeled MAGL inhibitors evaluated in vivo. Reversible inhibitor PET probes for MAGL, albeit in the nascent stage, will continue to be of interest in the future.

Overall, the development of PET radioligands for ECS targets is a very energetic research field, which will lay the foundation for the research of diseases and the drugs related to the ECS. Nonetheless, only a few PET tracers displayed suitable lipophilicity and specificity from in vivo studies. Apart from improving the physical properties of the target compound, there are two primary areas of improvement needed. First of all, ¹⁸F-labeled tracers should be more actively developed because of its longer half-life than ¹¹C, which potentially allows for longer scan times, diverse radiolabeling procedures, and remote regional distribution. The second important improvement is to develop and apply radiotracers with diverse binding characters (reversible and irreversible binding) to obtain comprehensive biological knowledge for quantification of the receptor in human subjects.²⁵⁷ Moreover, tracer validation requires continuous improvement of animal models, particularly regarding neuroinflammatory models that would accurately reflect alterations in the human ECS. Finally, advances in novel radiochemical technology,²³⁸ including [¹¹C]CO₂ fixation²⁰² and SCIDY^{235, 237, 258} for radiofluorination, will enable the discovery and access to novel, innovative PET tracers and may ultimately improve the current standard of diagnostic and therapeutic options for patients.

ABBREVIATION USED

AA

arachidonic acid

AAI

aminoalkylindoles

ABHD6

α/β hydrolase domain-6

ABHD12

α/β hydrolase domain-12

AD

Alzheimer disease

ALS

amyotrophic lateral sclerosis

AEA

N-arachidonoylethanolamine

BAT

brown adipose tissue

BBB

blood-brain barrier

BP

binding potential

CB1R

cannabinoid receptors 1

CB2R

cannabinoid receptors 2

CNS

central nervous system

CVD

cardiovascular dysfunction

CYP

cytochrome P450

DS

Dravet Syndrome

ECS

endocannabinoid system

FAAH

fatty acid amide hydrolase

GPCR

G-protein coupled receptor

HD

Huntington’s disease

LGS

Lennox-Gastaut Syndrome

MAGL

monoacylglycerol lipase

MRI

magnetic resonance imaging

NHP

nonhuman primate

PD

Parkinson’s syndrome

PBA-HD

Problem Behavior Assessment for Huntington’s disease

PET

positron emission tomography

PgP/Bcrp

P-glycoprotein/breast cancer resistance protein

PK-PD

pharmacokinetic-pharmacodynamic

RCYs

radiochemical yields

SCIDY

spirocyclic iodonium ylide

SPECT

Single Photon Emission Computed Tomography

SUV

standard uptake value

TRT

Test–retest

V_T

volume of distribution

2-AG

2-arachidonoylglycerol

Biography

Lu Hou is a third-year graduate student in Jinan University (JNU), majoring in imaging medicine and nuclear medicine under the guidance of Dr. Hao Xu and Dr. Lu Wang. Her main research interests are the imaging of the central nervous system and the clinical evaluation and transformation of new PET probes.

Jian Rong obtained his BSc degrees in Chemistry and Biology with Professor Wen-Jing Xiao in Central China Normal University (CCNU) in 2012, followed by his Ph.D. in Organic Chemistry with Professor Jinbo Hu in Shanghai Institute of Organic Chemistry (SIOC), Chinese Academy of Sciences (CAS) in 2017. After that, he continued his postdoctoral research with Professor Steven H. Liang in Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Massachusetts General Hospital (MGH) and Harvard Medical School (HMS). His research interests include 18F-labeling methodologies and development of PET tracers for the central and peripheral nervous system.

Ahmed Haider obtained his M.S. in Pharmaceutical Sciences at ETH Zurich, Switzerland, where he subsequently earned his Ph.D. in radiochemistry and molecular imaging with Professor Simon M. Ametamey in 2018. In the same year, he moved to the Department of Nuclear Medicine at the University Hospital Zurich, working with Prof. Catherine Gebhard to advance cardiovascular gender medicine with particular focus on preclinical and clinical cardiovascular PET imaging. Recently Dr. Haider joined Professor Steven H. Liang’s research group in the Division of Nuclear Medicine and Molecular Imaging at Harvard Medical School and Massachusetts General Hospital in Boston, USA.

Daisuke Ogasawara received his B.S. in 2011 and M.S. in 2013 in Pharmaceutical Sciences from Nagoya City University. He then joined the research group of Prof. Benjamin F. Cravatt at The Scripps Research Institute, where he conducted his doctoral study focusing on development of chemical tools for serine hydrolases involved in metabolism of signaling lipids. After receiving his Ph.D. in Chemistry in 2019, he joined the research group of Prof. Stuart L. Schreiber as a JSPS Overseas Research Fellow.

Cassis Varlow received her BSc in Chemical Biology from McMaster University in 2019. She continued her education at the University of Toronto in Dr. Neil Vasdev’s lab at the Centre for Addiction and Mental Health (CAMH) Brain Health Imaging Centre as an MSc candidate, and a prospective PhD student. Her research interests include in vitro radioligand characterization to de-risk the development of successful PET radiotracers for pre-clinical and human translation, for the study of neurodegenerative diseases.

Michael A. Schafroth received his B.Sc. in Chemistry in 2010 and M.Sc. in Interdisciplinary Sciences in 2012 from ETH Zürich. He then joined the research group of Prof. Erick M. Carreira, where he worked on the design and development of asymmetric iridium-catalyzed allylic substitution reactions, applications of stereodivergent methods to the synthesis of cannabinoids, and development of novel tools to study interactions of small molecules with the cell surface proteome. After receiving his Ph.D. in Chemistry in 2017, he joined the research group of Prof. Ben Cravatt at Scripps Research as a SNSF sponsored postdoctoral research associate. He currently works on the synthesis of small molecule inhibitors for serine hydrolases and is developing novel chemoprotoemic probes to modulate protein function in different biological arenas.

Lingjing Mu received her PhD in Chemistry from Nankai University, China in 1996. She worked as a postdoctoral fellow at the University of Basel and Center for Radiopharmaceutical Sciences (CRS) in Switzerland. In 2004, she was appointed as the team leader for PET chemistry at CRS. She is the Research Head of Radiopharmacy since 2010 and has working places at both University Hospital and ETH Zurich. Her research interests focus on PET tracer development and the development of new ¹⁸F-radiolabelling methods as well as the translation of new PET tracers into the clinic.

Jiefeng Gan is a third-year graduate student in JNU with the major of nuclear medicine under the supervisor of Dr. Lu Wang. Her research focus on preliminary evaluation of novel PET tracers in rodents and non-human primates.

Hao Xu received his PhD in nuclear medicine from Humboldt University of Berlin in 1996. Dr. Xu is currently the director of the Nuclear Medicine Department and PET/CT-MRI Center in the First Affiliated Hospital of Jinan University (JNUH), and the executive vice director of the Institute of Molecular and Functional Imaging in JNU. Dr. Xu’s research interests are clinical nuclear medicine and molecular imaging of cancer and neurological diseases.

Christopher J. Fowler obtained his B.A. at Cambridge University in 1975, followed by his Ph.D. in Pharmacology at the same University in 1978. Following post-doctoral posts in Sweden, Ireland and France, he was a researcher at Astra Läkemedel (now part of AstraZeneca) from 1982 to 1994. He was thereafter appointed Professor of Pharmaology at Umeå University in Sweden and took up his appointment there in 1995. Dr. Fowler’s current research interests concern the pharmacology of the endocannabinoid system. Dr. Fowler was recognized as a highly-cited researcher 2019 by the Web of Science group.

Ming-Rong Zhang received his B.S. at China Pharmaceutical University in 1985, followed by his Ph.D. in Medicinal Chemistry with Professor Masatoshi Yamato in Okayama University in 1993. Then he joined the Molecular Probe Program in National Institute of Radiological Sciences (NIRS) of Japan in 1998 and was promoted to the leader of Radiochemistry Team in 2006 and the director of Molecular Probe Program of NIRS in 2012. Dr. Zhang is currently the director of Department of Advanced Nuclear Medicine Sciences, NIRS, QST of Japan. Dr. Zhang’s research interests include radiochemistry, radiotracer and radiotherapy development, as well as clinical translation.

Neil Vasdev received his PhD in Chemistry from McMaster University, followed by a NSERC postdoc at Berkeley National Labs in 2003. He is the Director of the CAMH Brain Health Imaging Centre and the Azrieli Centre for Neuro-Radiochemistry. He is an endowed Chaired Professor at the University of Toronto & Canada Research Chair in Radiochemistry and Nuclear Medicine. He began his faculty career in Toronto in 2004 and was the Director of Radiochemistry at MGH and an Associate Professor at HMS (2011–2017). His research interests include applying novel radiochemistry methods with ¹¹C and ¹⁸F to human PET imaging.

Simon M. Ametamey obtained his M.S. in Chemistry from the Technical University of Merseburg, Germany, in 1985. He earned doctorate degree in Organic Chemistry from the University of Zurich, Switzerland, under the supervision of with Prof. H.H. Heimgartner. Following, a postdoctoral fellowship with Prof. K. Bernhard at Hoffmann La Roche (Basel, Switzerland), he joined the Center for Radiopharmaceutical Sciences of ETH, PSI and USZ in Villigen, Switzerland, in 1991. In 1995, he became the group leader and was appointed Titularprofessor in 2006. Since then, he has co-headed the Animal Imaging Center-PET at the Institute of Pharmaceutical Sciences at ETH Zurich.

Benjamin F. Cravatt obtained his undergraduate education at Stanford University, receiving a B.S. in the Biological Sciences and a B.A. in History. He then received a Ph.D. from The Scripps Research Institute (TSRI) in 1996. Dr. Cravatt is a Professor and the Norton B. Gilula Chair of Chemical Biology in the Department of Chemistry at The Scripps Research Institute. His research group is interested in developing chemical proteomic technologies that enable protein and drug discovery on a global scale and applying these methods to characterize biochemical pathways that play important roles in human physiology and disease, especially as pertains to the nervous system and cancer.

Lu Wang received his B.S. in pharmacy from the Ocean University of China in 2010, followed by his Ph.D. in Chemical Genomics from Peking University in 2016. Dr. Wang was a senior visiting scholar at HMS/MGH (2014–2016) and has been a visiting research fellow at NIRS (2017-now). Dr. Wang is currently the Vice Director of the Center of Cyclotron and PET Radiopharmaceuticals (CCPR) in JNUH. Dr. Wang’s research interests include radiochemistry, PET tracer development and clinical translational studies.

Steven H. Liang obtained his B.S. at Tianjin University in 2003, followed by his Ph.D. in Chemistry with Professor Marco Ciufolini in the University of British Columbia in 2010. Then he started as a NSERC fellow with Professor E. J. Corey at Harvard University. In 2012, Dr. Liang accepted a junior faculty position at HMS and MGH. Dr. Liang is currently the Director of Radiochemistry and Biomarker Development, Nuclear Medicine and Molecular Imaging at MGH, and Associate Professor of Radiology at HMS. Dr. Liang’s research interests include radiochemistry, imaging biomarker and radiotherapy development, and clinical translation.