Classics in Neuroimaging: Development of PET Imaging Agents for Imaging Monoamine Oxidases

Abstract

In this Viewpoint, we highlight the history of positron emission tomography (PET) imaging agent development to quantify changes in monoamine oxidase (MAO)-A and -B enzyme expression or activity. MAO-A and-B have been central for our understanding of the dopaminergic pathway with applications in neuropsychiatric medicine of high socioeconomic burden, spanning from aging, addictions and mood disorders; as well as, uses in neurodegenerative diseases and disorders with MAO-B expression in astrogliosis, and putative applications in imaging the sympathetic neurons of the peripheral nervous system. This exploration has sparked unique radiochemical innovations including the clinical translation of irreversible propargylamine-based suicide inhibitors, application of deuterium-substitution to slow down metabolism, development of trapped metabolite imaging agents and unique ¹¹C-carbonylation chemistry to translate novel high-affinity reversibly-binding inhibitors for this important target.

Graphical Abstract

Monoamine neurotransmitters including catecholamines (dopamine, epinephrine, and norepinephrine), indolamines (serotonin), and imidazoleamines (histamine), as well as trace amines (e.g. phenethylamine and tyramine), have been a heavily studied family of small molecules due to their critical roles in the central nervous system (CNS). The metabolism of monoamines is key to their importance in various disease states. Monoamine oxidase (MAO) is an outer-mitochondrial membrane bound, flavin-containing enzyme which oxidizes monoamines; there are two isozymes, termed MAO-A and MAO-B, which are encoded by independent genes, have structural differences, and exhibit different selectivities towards the various monoamine neurotransmitters.

Impaired monoaminergic signaling underlies a large number of behavioral, neurodegenerative and psychiatric disorders in humans. Specifically, inhibition of MAO-A is utilized in the management of depression, and inhibition of MAO-B is one of the therapeutic approaches to treat Parkinson’s disease (Fowler et al., 2015; Moriguchi et al., 2018). As far as MAO-B expression in brain immune cells is concerned, MAO-B is concentrated in the astrocytes, with very low content of the enzyme in other glial elements such as microglia. An increasing body of evidence suggests upregulation of MAO-B in reactive astrocytes that is activated during neuroinflammatory processes, thereby suggesting the potential for measuring alterations in MAO-B as a marker of astrogliosis in neurodegenerative diseases such as Alzheimer’s disease (AD) (Carter et al., 2012; Narayanaswami et al., 2018).

The clinical relevance of MAO to numerous neurological disorders has fostered significant efforts in developing both carbon-11 and fluorine-18 labeled radiotracers for in vivo non-invasive positron emission tomography (PET) neuroimaging studies to investigate alterations of MAO in the human brain (Fowler et al., 2015). Similar to all targeted radiotracers for PET imaging and quantification, the challenge to develop radiotracers for MAO has been in the design of a chemical compound whose kinetics and distribution reflect specificity for a single biochemical process and whose distribution parallels the known regional concentration of the specific MAO isoform in different brain regions and in peripheral organs. The two enzymes, MAO-A and -B have long been the target of drug development, with numerous compounds developed and marketed for an array of neurological and psychiatric diseases. This body of work has provided many chemical structures that have served as starting points for the development of PET imaging agents, and include irreversible inhibitors (e.g. propargylamines), reversible inhibitors (e.g. harmine, and SL2511.88), and substrates that form trapped metabolites for imaging (MPTP and Cou) (Fig. 1). Each approach has presented its own challenges and limitations to the radiochemistry community, who have endeavored to develop MAO PET imaging agents useful for understanding MAO pathophysiology for improved patient management, and for support of drug development. It is important to note that inter-species differences exist in the expression of the two MAO isoforms; MAO-B predominates in the human brain, whereas MAO-A predominates in the rat brain. In addition, in the rat heart, postjunctional extraneuronal MAO is found unlike in primates and humans. These differences should be taken into account during preclinical evaluation of a potential MAO imaging agent and when translating PET radiotracers to clinical studies.

Figure 1:

Selected radioligands for PET imaging of MAO-A and MAO-B.

This Viewpoint aims to highlight the key advancements in the field of radiochemistry and PET imaging that have been instrumental in understanding the role of MAOs in the CNS. Herein, we will follow the history of development of MAO PET radiotracers, beginning with irreversible inhibitors and progressing through the use of reversible inhibitors and substrates. Our article aims to serve to contribute towards a series of Viewpoints and Minireviews: Classics in Neuroimaging.

Irreversible Inhibitors

Early approaches to isoform-selective MAO PET agents concentrated on irreversible (suicide) inhibitors, wherein, the PET isotope labeled suicide inhibitor will result in covalent labeling of the catalytically active enzyme residue and, provide good pharmacokinetic properties and brain uptake when administered in vivo and imaged by PET. The substituted N-methyl propargylanines, clorgyline (N-[3-(2,4-dichloro- phenoxy)propyl]-N-methyl-2-propynylamlne), and L-deprenyl(l-N,α- dimethyl-N-2-propynyl phenathylamine), have shown to be selective irreversible inhibitors of MAO-A and MAO-B, respectively. The first MAO radiotracers to be used as PET imaging agents in humans were the propargylamines [¹¹C]clorgyline (1) and [¹¹C]L-deprenyl (2). Initial PET studies in mice showed the specificity of [¹¹C]clorgyline for MAO-A and [¹¹C]L-deprenyl for MAO-B. Fowler and coworkers demonstrated the rapid brain uptake of each in humans and confirmed the irreversibility as demonstrated by reaching a plateau in the brain tissue time activity curves (Fowler et al., 2015). Moreover, the distribution of radioactivity paralleled MAO immunoreactivity in human brain tissue sections and was highest in the corpus striatum, thalamus, and brainstem. These radiotracers were confirmed as being MAO inhibitors through a blocking study with phenelzine, a non-selective MAO inhibitor.

While [¹¹C]clorgyline selectively binds to MAO-A in the human brain, it was not retained in the brain of baboon or rhesus monkey suggesting species differences in the susceptibility of MAO-A to inhibition by clorgyline. Stereoselectivity of MAO-B for L-deprenyl was confirmed by comparison of the brain uptake and retention of the ¹¹C-labeled inactive (D-) and active (L-) enantiomers of deprenyl. Results showed rapid clearance of the inactive enantiomer and retention of the active enantiomer within MAO B-rich brain regions. Pharmacokinetics of brain uptake and retention in human studies showed that the agents were likely flow limited due to the rapid rate of irreversible inhibition. To improve the pharmacokinetics, deuterated analogues were prepared in an attempt to slow the rate of trapping using the kinetic isotope effect (KIE). The primary KIE can be utilized to slow the rate of reaction observed provided the abstraction of the greater mass deuterium vs. proton is the rate limiting step of the process. In the case of the propargylamines, this is the case and the slower to cleave C–D bonds vs. C-H reduced the rate of irreversible inhibition. To that end, deuterium labeled isotopologs ([¹¹C]clorgyline-D₂ and [¹¹C]l-deprenyl-D₂ ([¹¹C]DED) were successfully radiolabeled by a collaborative effort led by Brookhaven Laboratory and University of Uppsala (Fowler et al., 2015).While this approach worked well for the imaging of MAO-B by [¹¹C]DED; for [¹¹C]clorgyline-D₂, the reduction in the trapping rate resulted in a lower ratio of specific-to-nonspecific signal in the regions of interest, and did not improve its utility for MAO-A imaging. Specifically, comparative studies of [¹¹C]clorgyline and [¹¹C]clorgyline-D₂ in the human brain revealed significant non-MAO-A binding of [¹¹C]clorgyline-D₂ in white matter and cerebellum of the human brain confirming that the relative strength of the signal due to MAO-A binding is diminished by deuterium substitution as the slower rate of irreversible binding allows more of the PET agent to wash out of the compartment. As a result, [¹¹C]clorgyline is superior to [¹¹C]clorgyline-D₂ for human brain studies as the higher rate of covalent attachment is required to attain signal above the background observed due to non-specific binding in the white matter. On the other hand, studies comparing [¹¹C]l-deprenyl and [¹¹C]DED in the human brain showed a robust deuterium isotope effect and provided a means of selectively controlling the rate of trapping of tracer in brain and enhancing sensitivity of binding to changes in MAO-B. Therefore, [¹¹C]DED is the preferred MAO-B propargylamine agent for human brain imaging studies.

[¹¹C]DED has been used to investigate the role of MAO-B in various neurological disorders, with a particular interest in neurodegeneration. The co-localization of MAO-B expressing astrocytes and amyloid plaques has been shown in human brain sections, indicating the possibility of MAO-B involvement in AD. Additionally, [¹¹C]DED binding on human tissue was determined to be highest in the earlier Braak stages. PET imaging using [¹¹C]Pittsburgh compound B ([¹¹C]PiB) has identified amyloid plaque load in AD and mild cognitive impaired patients (MCI). Carter and coworkers performed a multi-tracer trial using PET imaging with [¹¹C]DED in combination with [¹⁸F]fludeoxyglucose ([¹⁸F]FDG, which demonstrates glucose metabolism) and [¹¹C]PiB in a small cohort of age-matched normal, MCI and AD subjects (Carter et al., 2012). Based on the amyloid imaging results, the MCI subjects were then divided into PiB-positive ([¹¹C]PiB⁺) and PiB-negative ([¹¹C]PiB^-) subgroups, and the [¹¹C]DED imaging studies were then used to evaluate if MAO-B imaging might be indicative of astrocytosis as an early phenomenon in AD development. Increased [¹¹C]DED binding in the frontal and parietal cortices were found in both PiB⁺ and PiB^- MCI groups, as well as the AD cohort, but with higher and more widespread [¹¹C]DED binding in the PiB⁺ MCI group. Despite the small cohort sizes, the authors argued that these results are evidence of astrocytosis as an early step in AD progression. Follow up studies conducted by the same research group employed the same radioligands and reported initially high and then declining astrocytosis in autosomal dominant AD carriers, suggesting that astrocyte activation is implicated in the early stages of AD pathology.

Taken together, suicide inhibitors such as [¹¹C]clorgyline and [¹¹C]DED have been employed in seminal PET imaging studies to investigate the role of MAO-A and MAO-B in the living human brain. Nevertheless, these irreversibly binding suicide inhibitors are accompanied by significant technical challenges, which include: (i) a fast rate of irreversible binding making quantification of binding difficult, and (ii) formation of PET-label containing brain-penetrating metabolites, specifically (R)-methamphetamine and (R)-amphetamine that have high affinities towards monoamine transporters. Since such limitations can confound image analysis, other types of PET imaging agents, such as reversibly-binding high-affinity inhibitors were developed as an alternate approach.

Reversible Inhibitors

Inhibitors of MAO-A have been recognized for utility in the treatment of psychiatric disorders. As an example, it was theorized that endogenous levels of monoamine neurotransmitters were lower in subjects with major depressive disorder, perhaps explaining why inhibitors of the neuronal membrane monoamine transporters for serotonin and norepinephrine (SSRIs) or inhibitors of MAO-A (MAOIs) are successful drug treatments. To evaluate the potential differences in MAO-A in depression, a new MAO-A radiotracer was needed, and to address this radiochemists turned to the reversible inhibitors of MAO-A. Harmine, a potent and selective MAO-A inhibitor (k_d=2.0 nM) served as a template for development of the reversibly binding MAO-A selective radiotracer [¹¹C]harmine (3) which was synthesized by a simple [¹¹C]methylation of an O-desmethyl precursor. Comparison of [¹¹C]harmine in normal controls and drug-free depressed subjects showed an average 34% increase in specific binding across multiple brain regions, with significant increases in prefrontal cortex, temporal cortex, posterior cingulate, and thalamus. The authors propose that such elevated levels of MAO-A, which metabolizes all three of the major monoamine neurotransmitters (serotonin, norepinephrine and dopamine), might be a contributing factor in depression along with possible changes in monoaminergic receptors and the neuronal membrane monoamine transporters (these can also be studied using PET imaging and appropriate radiotracers). The ability to image and quantify changes in MAO-A offers the potential to improve our understanding of the role of the enzyme in other psychiatric disorders, making the continued development of MAO-A radiotracers (e.g., [¹¹C]befloxatone, [¹⁸F]fluoroethyl-harmol) an important area of radiopharmaceutical chemistry.Elevated MAO-A total distribution volume (V_T), measured with [¹¹C]harmine PET, particularly in the prefrontal and anterior cingulate cortex, has also been demonstrated in a number of illnesses associated with depressed mood and in high risk states for major depressive episodes (MDE). Some of these illnesses include: early withdrawal from alcohol dependence, early withdrawal from heavy cigarette smoking, and borderline personality disorders. While all of these conditions are highly co-morbid with MDE, and predispose to MDE, there are other conditions which frequently precede MDE; for example, early postpartum and perimenopause which are associated with elevated MAO-A levels across grey matter regions. Overall, [¹¹C]harmine appears to be a promising PET radiotracer for quantitative evaluation of MAO-A densities in several psychiatric and neurologic illnesses.

In addition to the reversible MAO-A PET agents that have been evaluated in humans, a reversible carbamate MAO-B reversible inhibitor [¹¹C]SL25.1188 (4) was prepared and evaluated (Fig. 2). Initial reports by Saba and co-workers described the synthesis of [¹¹C]SL25.1188 via [¹¹C]phosgene ([¹¹C]COCl₂), a carbon-11 reactant prepared from [¹¹C]CO₂, to give access to the labeled carbamate. [¹¹C]SL25.1188 demonstrated favorable properties in preclinical studies (reversible binding, high brain uptake, and slow metabolism) for imaging MAO-B in the brains of non-human primates (NHP). However, the preparation of this radiotracer via [¹¹C]COCl₂ was a primary drawback as it involves highly specialized apparatus requiring extensive upkeep, technical expertise, and replacement of key components between production runs. These technical challenges limit the use of [¹¹C]COCl₂ to only a few laboratories worldwide. This limitation was overcome by optimizing the radiosynthesis of [¹¹C]SL25.1188 such that it could be accomplished via an intramolecular cyclization reaction in an automated one-pot procedure directly from [¹¹C]CO₂, thereby precluding the use of [¹¹C]COCl₂. Our laboratory in Toronto has successfully translated [¹¹C]SL2511.88 for first-in-human clinical research studies, validating it as the first reversibly binding MAO-B PET imaging agent for human use. [¹¹C]SL2511.88 is presently being explored in several patient populations in clinical PET research studies including most recently in major depressive disorder (Moriguchi et al., 2018). In this study, we found that 50% of subjects with MDE exhibited MAO-B V_T values in the prefrontal cortex (PFC) that exceeded those of healthy subjects and greater MAO-B V_T is an index of MAO-B overexpression. These results may contribute towards understanding the pathologies of mitochondrial dysfunction, and impact how we target SSRI treatments.

Figure 2:

Radiotracer uptake of MAO-A and MAO-B PET tracers in human brain. (Adapted from Meyer, J. H. (2017) Novel Phenotypes Detectable with PET in Mood Disorders. PET Clinics, 12 (3), 361 – 371).

Carbon-11-labeled radiotracers (20.4 min half-life) enable PET studies at 2-hour intervals on the same day in the same individual. On the other hand, ¹⁸F-labeled tracers have a relatively longer half life (109.8 min) and can be distributed to facilities that lack their own production unit thereby facilitating multi-center trials. N-[3-(2′,4′-dichlorophenoxy)-2-¹⁸F-fluoropropyl]-N-methylpropargylamine ([¹⁸F]fluoroclorgyline) was developed as a potential ¹⁸F-labeled radiotracer for MAO-A (Fowler et al., 2015). In vitro measures showed high affinity for MAO-A (39 nM) and selectivity for MAO-A versus MAO-B. Low specific activity labeled compounds have also been developed from clorgyline by derivatizing either the 2-chloro- or the 4-chloro-substituent with ¹⁸F via fluorodestannylation using [¹⁸F]AcOF. However, this approach poses a methodological drawback of using low specific activity [¹⁸F]F₂ as a starting point for the synthesis, which are not commonly used by PET radiochemistry facilities.

[¹⁸F]fluorodeprenyl, which is derived by the radiofluorination of L-deprenyl on the side chain, showed highest specific binding in striatum, intermediate uptake in thalamus and cortex, and lowest binding in cerebellum of NHP. To address the drawback associated with the fast rate of the irreversible binding mechanism, [¹⁸F]fluorodeprenyl-D₂ was developed which demonstrated improved MAO-B quantitation through KIE, favorable pharmacokinetic properties with relatively fast washout from NHP brain and improved sensitivity for MAO-B. These effects were recapitulated in a comparison of [¹⁸F]fluororasagiline-D₂ vs. [¹⁸F]fluororasagiline. Nonetheless, the brain-penetrating radioactive metabolites of propargyl compounds are problematic for quantitation as they were for the cabon-11 agents. To date, no human studies with the ¹⁸F-labeled MAO tracers have been reported. We and our collaborators have been working towards optimization and preclinical evaluation of fluorine-18 labeled reversible MAO-B tracers. To that end, development of a fluorine-labeled derivative of SL25.1188 and its subsequent preclinical evaluation in NHP is underway at our laboratories. Further innovation in this area should make MAO-B PET imaging with reversibly binding radiotracers more accessible for clinical research.

Metabolic Trapping Mechanism

Finally, as a third option for radiotracer development for MAOs, our program at the University of Michigan has investigated the use of metabolic trapping for imaging MAO enzymatic activity (Drake et al., 2018). In this approach, a substrate that is freely diffusible across the blood-brain barrier is designed such that the product of MAO oxidation is sufficiently polar to be retained within brain tissues at the sites of enzyme action (Fig. 3). The development of a radiotracer substrate is an approach that has been used successfully and most importantly in the development of 2-[¹⁸F]FDG, which is a glucose analog that is trapped tissues after phosphorylation by hexokinase.

Figure 3:

Trapped Metabolite Imaging Agents for MAO

The radiotracer design of MAO substrates was based on 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), well known for forming a toxic metabolite upon oxidation by MAO. In the 1990’s, the Castagnoli lab developed 4-aryloxy derivatives of MPTP: these substrates, instead of forming MPP+ (or a related toxic metabolite), are hydrolyzed to two benign metabolites immediately after MAO-mediated oxidation. A series of 4-aryloxy-MPTP derivatives have been radiolabeled with carbon-11, investigated for selectivity and suitability for human translation, including [¹¹C]1-methyl-4-phenoxy-1,2,3,6-tetrahydropyridine ([¹¹C]PHXY) and [¹¹C]4-methyl-7-(pyridin-4-yloxy)-2H-chromen-2-one-1,2,3,6-tetrahydropyridine ([¹¹C]Cou). In the NHP brain, in vivo trapping of [¹¹C]PHXY was more sensitive to MAO-A inhibition, and [¹¹C]Cou was more sensitive to MAO-B inhibition. Given the encouraging isozyme selectivity of [¹¹C]PHXY and [¹¹C]Cou in the monkey brain, attempts were then made to improve the in vivo pharmacokinetics through application of the KIE, as previous studies with deuterated MPTP had demonstrated a KIE comparable to that observed for the propargylamines (e.g., deprenyl). However, incorporation of deuterium into the tetrahydropyridine ring of [¹¹C]Cou failed to influence either the in vitro rates of reaction with MAO or the in vivo rate of brain trapping, suggesting that the abstraction of the tetrahydropyridine protons is not the rate limiting step for MAO oxidation of [¹¹C]Cou and that any KIE was masked by a slower step. Further development and evaluation of these radiotracers is underway, with [¹¹C]Cou of interest for the measure of astrogliosis and [¹¹C]PHXY for use in psychiatric illnesses and the measurement of sympathetic neurons in the peripheral nervous system.

Throughout this Viewpoint, we have highlighted key advances in the field of radiochemistry that have been instrumental in understanding the role of MAO in the human brain. While some of the tracers discussed herein have been translated to clinical studies, there is continued interest in the development of new optimal MAO radiotracers that could be beneficial for imaging pathological changes in a broad spectrum of neurological diseases. We look forward to this new generation of MAO-B PET radiochemistry developments which offer insights to probe this target via new mechanisms, such as: the mechanistically trapped [¹¹C]Cou; reversibly binding [¹¹C]SL2511.88 which uses [¹¹C]CO₂ fixation chemistry for human radiopharmaceutical production; new ¹⁸F-tracers; and applications of these probes for neuroinflammatory processes and associated pathologies including depression.