Synthesis and pharmacokinetic study of a ¹¹C-labeled cholesterol 24-hydroxylase inhibitor using ‘in-loop’ [¹¹C]CO₂ fixation method

Abstract

Cholesterol 24-hydroxylase, also known as CYP46A1 (EC 1.14.13.98), is a monooxygenase Received and a member of the cytochrome P450 family. CYP46A1 is specifically expressed in the brain Received in revised form where it controls cholesterol elimination by producing 24S-hydroxylcholesterol (24-HC) as the Accepted major metabolite. Modulation of CYP46A1 activity may affect Aβ deposition and p-tau accumulation by changing 24-HC formation, which thereafter serves as potential therapeutic pathway for Alzheimer’s disease. In this work, we showcase the efficient synthesis and preliminary pharmacokinetic evaluation of a novel cholesterol 24-hydroxylase inhibitor 1 for use in positron emission tomography.

Graphical Abstract

1. Introduction

Alzheimer’s disease (AD) is a long-term neurodegenerative disorder that ranks sixth in the leading cause of all deaths in the United States and features amyloid β protein (Aβ) deposition and neurofibrillary tangles composed of aggregated phosphorylated tau protein (p-tau).[1] To date, an estimated 6% of 65-year or older people in the world are suffering from AD, which approximately accounts for 95% of AD cases. Despite tremendous research efforts towards AD, the underlying cause of AD remains elusive, and no treatment has been disclosed to stop or reverse the AD progress.

Recent studies have demonstrated that abnormalities of cholesterol homeostasis in the brain are strongly associated with several neurodegenerative diseases, such as AD, Parkinson’s disease (PD) and Huntington’s disease (HD).[2–6] As an important structural component of myelin as well as membranes of neurons and glial cells, cholesterol plays a central role in the process of dendrite outgrowth and myelination, maintenance of stability of microtubules, and synaptogenesis.[7] In the brain, cholesterol is mainly metabolized into 24S-hydroxylcholesterol (24-HC) by cholesterol 24-hydroxylase (CH24H) for elimination. CH24H, also known as CYP46A1 (EC 1.14.13.98), is a monooxygenase and a member of the cytochrome P450 family. It was reported that CYP46A1 polymorphism in the brain, which is deeply linked with increased Aβ and p-tau expression, could be a risk factor in AD pathology.[8–10]. Both inhibition and activation of CYP46A1 could have a therapeutic potential,[11] and CYP46A1 activity increases may suppress neuronal cell death, Aβ deposition, and p-tau accumulation.[6, 12, 13] Mast et al have demonstrated that a series of marketed drugs could be repurposed as potent CYP46A1 modulators including tranylcypromine, thioperamide, voriconazole, clotrimazole, and fluvoxamine.[14–17] Some of these compounds were co-crystallized with CYP46A1 to elucidate their molecular interactions.

Positron emission tomography (PET) is a sensitive, quantitative, and non-invasive nuclear imaging tool to provide diagnostic information about the biological processes under disease conditions.[18–20] PET studies of CYP46A1 would allow to better understand in vivo biochemistry and physiology of cholesterol homeostasis, help to accelerate the translation of CYP46A1 modulators in clinical trials, as well as offer an indepth insight into AD. Recently, we developed a facile “in loop” [¹¹C]CO₂ fixation technology, which directly takes advantage of an automated apparatus to enable ¹¹C-carbonylation in high efficiency.[21–24] With this technology, we and other groups have successfully synthesized a series of PET tracers, such as [¹¹C]SL25.1188, [¹¹C]JNJ1661010 and [¹¹C]MAGL-2–11.[23, 25–27] As our combined interest in the CYP46A1-targeted PET ligand development and application of ‘in-loop’ [¹¹C]CO₂ fixation, we present herein the facile synthesis of ¹¹C-labeled compound 1, which was recently patented as a potent CYP46A1 inhibitor (Figure 1).[28, 29] Pharmacological and physicochemical evaluation of compound 1, as well as pharmacokinetic profiling via in vivo PET imaging and whole-body biodistribution experiments of [¹¹C]1 are reported in this work.

Figure 1.

Chemical structure of CYP46A1 inhibitor 1 and [¹¹C]1

2.1. Chemistry

The synthetic route of CYP46A1 inhibitor 1 is shown in Scheme 1. In brief, we used commercially available tert-butyl 4(4-phenylpyrimidin-5-yl)piperazine-1-carboxylate 2 as the starting material. Trifluoroacetic acid (TFA)-promoted removal of the tert-butyloxycarbonyl (Boc) group proceeded smoothly to give free amine 3 in 99% yield, which also served as the labeling precursor for follow-up [¹¹C]CO₂ fixation. Subsequent installation of benzylaminocarbonyl group into 3 was readily achieved with the combination of benzylamine, triphosgene and N,N-diisopropylethylamine (DIPEA) in THF. In all, the desired CYP46A1 inhibitor 1 was synthesized in 52% yield.

Scheme 1.

Synthesis of CYP46A1 inhibitor 1. (i) TFA, CH₂Cl₂, rt, 8 h, 99% yield; (ii) benzylamine, triphosgene, DIPEA, THF, rt, 30 min, 52% yield. TFA = trifluoroacetic acid, DIPEA = N,N-diisopropylethylamine, THF = tetrahydrofuran.

2.2. Pharmacology and physiochemical property

We select compound 1 from a patent literature describing therapeutic drug molecules towards CYP46A1 for the treatment of neurodegenerative diseases.[28, 29] The compound showed 95% inhibitory activity towards CYP46A1 at 1 μM concentration and possessed a carbamide (urea) moiety which is amenable for radiolabeling using our ‘in-loop’ [¹¹C]CO₂ fixation.[21–24] Subsequent in silico calculation also indicated a favorable value of 60.3 Å for topological polar surface area (tPSA). Multiparameter optimization (MPO) has been widely utilized in the CNS drug discovery to evaluate brain permeability.[30] Our test compound 1 showed a MPO score of 4.9, which together with other prediction indicated that 1 may have high likelihood to penetrate the blood-brain barrier. Experimentally, compound 1 was subjected to potency evaluation towards CYP46A1 in vitro with our previously developed method.[16, 31] Briefly, purified CYP46A1 (d(2–50)CYP46A1–4His) was reconstituted with varying concentrations of compound 1, rat NADPH cytochrome P450 oxidoreductase and cholesterol. The hydroxylation reaction was started by nicotinamide adenine dinucleotide phosphate (NADPH) at 37 ^oC, and maintained at the same temperature for 30 minutes. Enzymatic reaction was terminated by addition of dichloromethane. The hydroxylation products extracted with dichloromethane were then subjected to GC-MS (gas chromatography-mass spectrometry) analysis with 24-OTMS d₇cholesterol as internal standard. As depicted in Figure 2, the Ki of compound 1 was determined to be 7.3 ± 0.1 μM. We postulated the discrepancy of binding affinity of compound 1 could be derived from different measurement methods. In all, although it is certain that a more potent compound is necessary to demonstrate high binding potential in PET study, a radiochemical application of our [¹¹C]CO₂ fixation method based on compound [¹¹C]1 were carried out to provide preliminary pharmacokinetic profiling as an entry point for further design CYP46A1 ligands.

Figure 2.

The concentration-response curve of 1 for inhibition of cholesterol 24-hydroxylation by CYP46A1.

2.3. Radiosynthesis and lipophilicity measurement

In virtue of the urea group, radiosynthesis of compound 1 could be directly carried out via our previously developed ‘in-loop’ [¹¹C]CO₂ fixation strategy.[23, 24, 26] As shown in Figure 3, [¹¹C]CO₂ was first reacted with benzylamine to form a carbomate intermediate [¹¹C]5 with BEMP as the [¹¹C]CO₂ trapping reagent. POCl₃-mediated dehydration of [¹¹C]5 provided the isocyanide species [¹¹C]6, which, upon quenched with secondary amine 3, yielded the desired radioligand [¹¹C]1. It was found that the efficiency of this reaction is closely associated with the amount of benzylamine. Using 4.6 μmol of benzylamine, the desired product [¹¹C]1 was obtained in a poor radiochemical conversion (4% RCC), whereas the side symmetric product [¹¹C]7 was formed in 80% RCC (entry 1, Table 1). Decreasing the amount of benzylamine to 0.92 μmol successfully suppressed the formation of side product [¹¹C]7, and the RCC of [¹¹C]1 increased to 45% (entry 2). Further decrease of benzylamine loading by 50% led to an increased RCC to 61% (entry 3). For comparison, we also performed a vial-based [¹¹C]CO₂ reaction for the radiosynthesis of [¹¹C]1. Unfortunately, merely trace product was obtained (entry 4), which further highlighted the advantage of our ‘in-loop’ [¹¹C]CO₂ fixation method. Ultimately, [¹¹C]1 was isolated in 21% decay-corrected radiochemical yield (RCY) with excellent radiochemical purity (>99%) and high molar activity (>37 GBq/μmol). Excellent in vitro formulation stability was observed for [¹¹C]1, as evidenced by no radiolysis 90 min post formulation in saline containing 5% ethanol (Figure 4). Lipophilicity of a specific drug is an important index to evaluate its permeability of the blood-brain barrier (BBB) with a favorable range from 1.0 to 3.5.[32–34] The lipophilicity of [¹¹C]1 (logD) was measured as 3.05 ± 0.02 via the ‘shake flask’ method, which is the liquid-liquid partition between phosphate buffered saline (PBS) and n-octanol.[35]

Figure 3.

Schematic diagram of the “in loop” [¹¹C]CO₂-fixation module

Table 1.

Optimization of reaction parameters for radiosynthesis of tracer [¹¹C]1.^a

Entry	4 (μmol)	3 (μmol)	[11C]CO2 trapping (%)	RCC of [11C]l (%)	RCC of [11C]7 (%)
1	4.6	8.4	>99	4	80
2	0.92	18.5	>99	45	35
3	0.46	18.5	>99	61 (21)b	12
4c	0.92	8.3	90	trace	trace

^aReaction conditions: loop A: benzylamine 4 and BEMP (2.5 μL, 8.6 μmol) in DMF (40 μL); loop B: 3 in DMF (50 μL); loop C: 0.2% POCl₃ (v/v, 1.1 μmol) in MeCN (100 μL) and additional 800 μL MeCN.

^bDecay-corrected radiochemical yield (RCY).

^cReaction was conducted in a vial.

Figure 4.

Stability of radioligand [¹¹C]1 in saline containing 5% of ethanol.

2.4. PET imagine studies in mice

With [¹¹C]1 in hand, we then performed dynamic PET acquisitions in CD-1 mice for 60 min. Representative PET images in the brain (0–60 min summed) and whole-body as well as time-activity curves (TACs) are shown in Figure 5 and Figure S1 in the Supporting Information. [¹¹C]1 demonstrated limited brain uptake with a maximum standard uptake value (SUV) of 0.42 at 0.5 min, followed by a steady washout. Pretreatment of 2-phenylcyclopropan-1-amine (3 mg/kg), a known CYP46A1 inhibitor,[14] reduced the uptake of [¹¹C]1 (~31% decrease of AUC, area under curve). These experiments indicated a marginal level of in vivo binding specificity for [¹¹C]1.

Figure 5.

(A) Representative PET images of [¹¹C]1 in mouse brain (0–60 min summed); (B) Time–activity curves of [¹¹C]1 in mouse brain.

2.5. Whole-body ex vivo biodistribution studies

To further probe the uptake, biodistribution and elimination of [¹¹C]1, whole-body ex vivo biodistribution was carried out in mice at four time points (5, 15, 30 and 60 min) post intravenous (IV) injection of the radioligand. As shown in Figure 6 and Table S1 in the Supporting Information, tracer [¹¹C]1 exhibited limited brain uptake, which is in line with the results from PET imaging studies. High radioactivity was accumulated in several organs including lungs, pancreas, small intestine, kidneys and by last washout. The radioactivity in small intestine reached a climax at 30 min followed by fast elimination, which together with high residue radioactivity in the small intestine and liver, suggested the urinary and hepatobiliary elimination pathway for [¹¹C]1.

Figure 6.

Whole-body ex vivo biodistribution studies of [¹¹C]1 in mice at four different time points (5, 15, 30 and 60 min) post tracer injection. Data are expressed as % ID/g (mean ± SD; n = 4). Asterisks indicate statistical significance. * p < 0.05, ** p ≤ 0.01, *** p ≤ 0.001, and **** p ≤ 0.0001.

3. Conclusion

In this work, we have efficiently synthesized a novel CYP46A1 PET tracer [¹¹C]1 in high radiochemical yield (21%), excellent molar activity (>37 GBq/μmol) and radiochemical purity (> 99%) using ‘in-loop’ [¹¹C]CO₂ fixation method. By PET imaging and ex vivo whole body distribution studies, we provided a pharmacokinetic profile (uptake, retention and clearance) of [¹¹C]1 in major organs, of which the brain exhibited low uptake (0.4 SUV) and marginal binding specificity in vivo. Further improvement of binding affinity and ADME properties are underway to develop new CYP46A1 ligands.