Improved Synthesis of [¹¹C]COU and [¹¹C]PHXY, Evaluation of Neurotoxicity, and Imaging of MAOs in Rodent Heart

Abstract

The radiotracers [¹¹C]COU and [¹¹C]PHXY are potential PET imaging agents for in vivo studies of monoamine oxidases (MAOs), as previously shown in rodent and primate brain. One-pot, automated methods for the radiosynthesis of [¹¹C]PHXY and [¹¹C]COU were developed to provide reliable and improved radiochemical yields. Although derived from the structure of the neurotoxin MPTP, COU did not exhibit in vivo neurotoxicity to dopaminergic nerve terminals in the mouse brain as assayed by losses of VMAT2 radioligand binding. PET imaging studies in rats demonstrated that both [¹¹C]COU and [¹¹C]PHXY exhibit retention in cardiac tissues that can be blocked by pretreatment with the MAO inhibitors deprenyl (MAO-B) and pargyline (MAO-A and -B). In addition to prior neuroimaging applications, [¹¹C]COU and [¹¹C]PHXY are thus also of interest for studies of MAO enzymatic activity and imaging of sympathetic nerve density in heart.

In previous work, we developed and investigated the use of 1-[¹¹C]methyl-4-phenoxy-1,2,3,6-tetrahydropyridine ([¹¹C]PHXY) and 4-methyl-7-((1-[¹¹C]methyl-1,2,3,6-tetrahydropyridin-4-yl)oxy)-2H-chromen-2-one ([¹¹C]COU) as brain penetrant PET imaging agents.^1,2 Both radiotracers are nontoxic ether analogues of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) and are selectively oxidized by MAO-A or -B to yield a common radiolabeled metabolite 2,3-dihydro-1-[¹¹C]methyl-4(1H)-pyridinone.³⁻⁵ In rodent and nonhuman primate (NHP) studies, it was demonstrated that, after i.v. injection, [¹¹C]PHXY and [¹¹C]COU can freely cross the blood–brain barrier (BBB), but inside the brain they are metabolically cleaved, the resultant radiolabeled metabolite is trapped, and that provides a quantifiable amount of signal for estimating MAO enzymatic activity. Studies in NHPs showed the advantage of using a trapped metabolite approach, as regional SUV values (4.8 in striatum, 4.3 in thalamus, 3.8 in cerebellum, and 2.8 in cortex) were higher than values obtained using reversible binding ligands for MAO.⁶ We had three goals for further advancing these PET imaging agents: (1) develop a one-pot, high yielding, and automated method for the radiosynthesis of [¹¹C]PHXY and [¹¹C]COU; (2) demonstrate that PHXY and COU do not exhibit neurotoxicity as is observed with MPTP; and (3) provide preliminary evidence for metabolic trapping in the rodent heart and thus potential as imaging agents for sympathetic nerve density.

Monoamine oxidases (MAOs) are flavoenzymes found on the outer mitochondrial membrane where they catalyze the oxidative deamination of catecholamines including the neurotransmitters epinephrine, norepinephrine, and dopamine. MAO has important roles in both the central and peripheral nervous system.^7,8 MAO expression in astrogliosis makes it a potential marker for neuroinflammation associated with neurodegenerative disease like Alzheimer’s disease, while in the sympathetic nervous system MAO breaks down catechol neurotransmitters that have key roles in regulation of cardiac function. In addition, the high expression of mitochondria in the heart suggests imaging MAO could elucidate information on the pathogenesis of ischemia-reperfusion injury, heart failure, and other cardiovascular diseases involving sympathetic neurons. MAO imaging agents thus present an alternative to previously developed radiotracers for the sympathetic nervous system that have focused on the norepinephrine transporter.^9,10

The potential applications of MAO imaging have prompted the synthesis and evaluation of a number of PET radiotracers based on reversible and irreversible MAO inhibitors.⁶ Classical radiotracers like [¹¹C]L-deprenyl-D2 irreversibly inhibit MAO by binding to it covalently, while [¹¹C]PHXY and [¹¹C]COU, as metabolically trapped imaging agents, provide a different mechanism for radioactivity localization.⁶ While limited work with [¹¹C]L-deprenyl-D2 has attempted to image MAO in the heart,¹¹ most efforts with both classes of MAO radiotracer have concentrated on imaging applications in neurology and psychiatry.⁶ The potential for MAO radiotracers to serve as imaging biomarkers of sympathetic nerve density in cardiac applications as such remains unexplored, and in this work we therefore report preliminary cardiac PET imaging with [¹¹C]PHXY and [¹¹C]COU. Additionally, we also describe improved syntheses and demonstrate lack of toxicity, both of which are important milestones in our efforts to translate these new MAO radiotracers into clinical studies.

To improve radiochemical yield and automate the synthesis of [¹¹C]PHXY and [¹¹C]COU using a commercial synthesis module, we needed to develop a one-step methylation procedure. Our previously reported route required two steps: methylation of a pyridine, followed by the reduction of the [¹¹C]methyl-pyridinium intermediate with sodium borohydride to form the desired product (Figure 1A). This method allowed the quick evaluation of multiple 4-aryloxy substituted tetrahydropyridine substrates, but it utilized a purpose built system and multiple steps from [¹¹C]methyl triflate ([¹¹C]MeOTf). In addition, radiochemical yields (RCYs) were low for both [¹¹C]PHXY and [¹¹C]COU (0.8–1.7%), limiting utility of the approach. To address these challenges, the generation of desmethyl 1,2,3,6-tetrahydropyridine precursors for [¹¹C]PHXY and [¹¹C]COU was undertaken to develop a straightforward one-step labeling procedure that provides the radiotracers in high RCY (13.4–15.5%) (Figure 1B).

Figure 1.

Strategies for the synthesis of [¹¹C]PHXY and [¹¹C]COU.

The desmethyl tetrahydropyridine precursor for PHXY, 2, could be prepared via reduction of the previously utilized precursor 4-phenoxypyridine 1 using a mixture of lithium aluminum hydride and aluminum trichloride in diethyl ether at room temperature (rt), Figure 2A. The reaction was complete after 18 h in satisfactory yield (54%).¹² Precursor 2 was easily methylated with [¹¹C]MeOTf at rt for 3 min using a commercial synthesis module, and, after purification utilizing an injectable buffer solution (10% EtOH, 10 mM NaOAc, pH 4.4), doses suitable for animal experiments were obtained. The radiochemical yield for [¹¹C]PHXY increased to 13.4% (non-decay-corrected, 102 ± 30 mCi, n = 4) compared to 0.8% (non-decay-corrected, 7.3 ± 4.2 mCi, n = 8) obtained using our prior 2-step process.

Figure 2.

Synthesis of [¹¹C]PHXY (A) and [¹¹C]COU (B).

Unfortunately, the simple reduction of the pyridine precursor 4, readily accessed from 7-hydroxy-4-methylcoumarin 3, did not provide the required desmethyl tetrahydropyridine precursor for [¹¹C]COU due to undesired side reactions with the lactone portion of the coumarin. To prevent this, pyridine 4 was N-alkylated with 3-bromoprop-1-ene overnight (o/n) at 140 °C to give the allyl-pyridinium intermediate which, upon reduction with sodium borohydride at 0 °C for 1 h, yielded allyl protected tetrahydropyridine 5 (Figure 2B). The allyl group of 5 was then removed by palladium-catalyzed reduction with 1,3-dimethylbarbituric acid (DMBA) at 35 °C for 3 h to give the desired precursor 6.¹³ Alkylation of 6 was accomplished using [¹¹C]MeOTf in DMSO at rt for 3 min, and [¹¹C]COU was purified by HPLC (50% acetonitrile/50 mM NaHCO₃, pH 10). The product was reformulated in ethanol and saline to provide a dose suitable for injection. The one-step synthesis gave [¹¹C]COU in 15.5% RCY (non-decay-corrected, 140 ± 19.6 mCi, n = 6), which was a significant improvement compared to the previous 1.7% RCY (15.6 ± 3.8 mCi, n = 10) obtained with the original two-step procedure. This method of protecting the ring nitrogen might be useful for radiolabeling other complex substituted tetrahydropyridines, as it appears to be milder than the direct tetrahydropyridine reduction employed to prepare [¹¹C]PHXY precursor 3.

COU and PHXY are structurally related to the neurotoxin MPTP, but incorporate an oxygen atom as an ether linkage between the tetrahydropyridine ring and 4-aryl substituents. MPTP is oxidized by monoamine oxidase to form the neurotoxic metabolite 1-methyl-4-phenylpyridinium ion (MPP⁺), as seen in Figure 3A. The 4-aryloxy-N-methyl-tetrahydropyridines such as PHXY and COU are also oxidized by MAO, forming the pyridinium species 7; however, this structure undergoes spontaneous hydrolysis, cleaving the ether-linkage to form two nontoxic products (Figure 3B).⁴

Figure 3.

(A) MPTP oxidation by MAO to produce neurotoxic MPP⁺. (B) MAO oxidation and hydrolysis of 4-aryloxytetrahydropyridines to form nontoxic metabolites.

The structural similarity of COU to MPTP, and the higher brain uptake ([¹¹C]COU: 1.10%ID/g vs [¹¹C]MPTP: 0.03%ID/g),¹⁴ prompted us to evaluate the potential CNS toxicity of COU in a mouse model. For that study, we employed a second radiotracer [¹¹C]dihydrotetrabenazine ([¹¹C]DTBZ) that specifically binds to vesicular monoamine transporter type 2 (VMAT2) of brain monoaminergic nerve terminals. Losses of in vivo VMAT2 radioligand binding have previously been used to demonstrate the toxicity of MPTP in mouse and nonhuman primate brains.^15,16 Groups of male C57BL/6 mice were injected with saline (N = 9, controls) or unlabeled COU (86 mg/kg, 0.32 mmol/kg, N = 9). At 3 (N = 5) and 14 days (N = 4), subgroups of control animals were injected with [¹¹C]DTBZ, and the regional distributions of brain radioactivity were determined at 20 min by ex vivo dissection. Similarly, the subgroups of COU-treated mice were studied for [¹¹C]DTBZ binding at 4 (N = 5) and 15 (N = 4) days. This protocol and timing were chosen to reasonably match the prior study using an equivalent molar injection of MPTP (55 mg/kg, 0.32 mmol/kg).¹⁶

The specific binding of [¹¹C]DTBZ to the VMAT2 was estimated as [(%ID/g region of interest)/(%ID/g cerebellum) – 1]. There were no significant differences (P > 0.05) in [¹¹C]DTBZ binding between control and COU-treated mice at either 4 or 15 days (Supporting Information Chart 1). This is in contrast to the previous demonstration that MPTP treatment produces very large (>60%) losses of VMAT2 radioligand binding in the striatum of the mouse (Figure 4). These results support a conclusion that, as shown previously,⁴ 4-aryloxy substituted N-methyl-tetrahydropyridines do not form the neurotoxic N-methylpyridinium species found after MAO oxidation of MPTP.

Figure 4.

Specific binding of [¹¹C]DTBZ in striatum for saline (controls), COU, and MPTP treated C57BL/6 mice. Significant reductions were only observed after MPTP treatment (MPTP data from ref (15)).

Lastly, our previous studies demonstrated MAO-dependent trapping of radioactivity in rodent and primate brain after injections of radiolabeled 4-aryloxy-N-methyltetrahydropyridines, but potential applications in other tissues have not been explored. We wished to investigate their use as cardiac imaging agents because, in the heart, MAO is important both as part of the normal metabolic route for catecholamines of the sympathetic nervous system and as a normal component of the mitochondria-rich cells of the heart muscle. Heart MAO increases in expression during aging, leading to concomitant increases in the amount of oxidative deamination of catecholamines, and contributes to mitochondrial oxidative stress and cardiovascular disease.^8,9

Preclinical PET imaging studies in Sprague–Dawley rats provided clear images of radioactivity uptake and retention in the rat heart after i.v. injection of [¹¹C]PHXY or [¹¹C]COU (Figure 5A). Radioactivity uptake into the ventricle was rapid, peaking at approximately 90 s, with maximum radioactivity uptake significantly higher for both radiotracers than for [¹³N]NH3 (SUV = ∼2.5),¹⁷ the established radiotracer for cardiac blood flow.¹⁸ In control studies, the pharmacokinetic curves then rapidly decreased by 750 s to lower plateau values of approximately 60–70% of the peak. These curves are very similar to those obtained in the brain (high uptake, some washout, and a plateau at later times¹), suggesting that [¹¹C]PHXY or [¹¹C]COU were functioning as MAO-dependent metabolic trapping agents in the heart. To demonstrate that the trapping of radioactivity was MAO dependent, PET imaging was repeated for each radiotracer after pretreatment (10 mg/kg I.P. 90 min prior to tracer injection) with the MAO inhibitors pargyline (MAO-A+B) or deprenyl (MAO-B).¹⁹ Neither MAO inhibitor significantly altered the timing or magnitude of the initial radioactivity uptake peak. However, both inhibitors produced increased washout of radioactivity until the curves reached lower plateau values (Figure 5B) representing 60–84% reductions in radiotracer trapping. These preliminary studies support the idea that both [¹¹C]PHXY and [¹¹C]COU function as metabolic trapping agents in the rat heart, and the almost complete blocking of [¹¹C]COU trapping by deprenyl is suggestive that the metabolism of that radiotracer may represent primarily MAO-B.

Figure 5.

(A) [¹¹C]COU and [¹¹C]PHXY rat heart PET images of baseline and deprenyl and pargyline blocking for each compound. (B) Standard uptake curves over time (s) of [¹¹C]COU and [¹¹C]PHXY in baseline studies, and after deprenyl and pargyline blocking.

In conclusion, this work has demonstrated a new high yield one step radiosynthesis for [¹¹C]COU and [¹¹C]PHXY, confirmed that 4-aryloxy-N-methyltetrahydropyridines such as COU have no neurotoxicity to monoaminergic nerve terminals, and showed that trapping and retention of the radiotracers in the rat cardiac ventricle is MAO dependent. Future PET imaging studies are planned and will include pharmacokinetic and MAO blocking studies in nonhuman primates, so we can better understand the potential of these new radiotracers for understanding the roles of MAO in cardiac function and disease.

Glossary

Abbreviations

BBB

blood–brain barrier

COU

4-methyl-7-((1-methyl-1,2,3,6-tetrahydropyridin-4-yl)oxy)-2H-chromen-2-one

DMBA

1,3-dimethylbarbituric acid

DTBZ

dihydrotetrabenazine

MAO

monoamine oxidase

MPTP

1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine

NHP

nonhuman primate

PET

positron emission tomography

PHXY

1-methyl-4-phenoxy-1,2,3,6-tetrahydropyridine

RCY

radiochemical yield

VMAT2

vesicular monoamine transporter type 2.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.0c00419.

• Experimental details and NMR, mass spectrometry and HPLC analysis data (PDF)

Author Contributions

The manuscript was written through contributions of all authors, and all authors have given approval to the final version of the manuscript.

We thank NIH (R21NS075553) and the University of Michigan (Energy Institute/Michigan Memorial Phoenix Project; Undergraduate Research Opportunity Program) for financial support.

The authors declare no competing financial interest.