Synthesis, 3 H-labelling and in vitro evaluation of a substituted dipiperidine alcohol as a potential ligand for chemokine receptor 2.

The immune system is implicated in the pathology of neurodegenerative disorders. The C-C chemokine receptor 2 (CCR2) is one of the key targets involved in the activation of the immune system. A suitable ligand for CCR2 could be a useful tool to study immune activation in central nervous system (CNS) disorders. Herein, we describe the synthesis, tritium radiolabelling, and preliminary in vitro evaluation in post-mortem human brain tissue of a known potent small molecule antagonist for CCR2. The preparation of a tritium-labelled analogue for the autoradiography (ARG) study gave rise to an intriguing and unexpected side reaction profile through a novel amination of ethanol and methanol in the presence of tritium. After successful preparation of the tritiated radioligand, in vitro ARG measurements on human brain sections revealed nonspecific binding properties of the selected antagonist in the CNS.

INTRODUCTION

White blood cells or leukocytes are an integral part of the immune system and protect the host organism by patrolling for infectious pathogens and tissue damage. This action is guided by a family of small signalling proteins called chemokines (chemotactic cytokines)1, 2 and their corresponding receptors. Chemokines regulate innate as well as adaptive immune responses by coordinating development, differentiation, anatomic distribution, effector functions, and trafficking of leukocytes.3 In addition, chemokines cover tasks ranging from lymphocyte development and homing4, 5 to neuronal communication.6, 7 Their physiological effect is derived from interaction with specific G protein–coupled receptors (GPCRs)—transmembrane receptors that are sometimes also referred to as serpentine receptors because the single polypeptide chain winds through the membrane seven times—called chemokine receptors. To date, approximately 50 human chemokines and 20 receptors have been identified, which have traditionally been divided into four families (CXC, CC, C, and CX3C) on the basis of the pattern of amino acid residues near the –NH2 terminus of the ligands (C stands for cysteine, and X/X3 represents one or three noncysteine amino acids respectively).8 C‐C chemokine receptor 2 (CCR2) is expressed on dendritic cells, endothelial cells and T cells. It is the primary chemokine receptor on inflammatory monocytes and it mediates migration towards chemokines such as CCL2, its primary ligand. However, CCR2 has been implicated in a variety of diseases including atherosclerosis,9, 10, 11 asthma12, 13 and autoimmune14, 15, 16 and metabolic diseases17 as well as neurological diseases such as neuropathic pain18 and Alzheimer's disease (AD).19 Various pharmaceutical companies have tried to identify CCR2 antagonists to therapeutically target one or several of these implications.20, 21, 22 Hurdles in their development have been plentiful, including discrepancies in the activity obtained in the rodent compared with the human receptor, ineffective dosing, and selectivity issues, as antagonists often exhibit high affinities for other chemokine receptors, especially CCR5 and CCR1, as well as ion channels, including hERG.22, 23, 24 Nonetheless, the CCR2/CCL2 axis remains an attractive target with several clinical trials currently ongoing.

https://clinicaltrials.gov/ct2/show/NCT02345408, https://adisinsight.springer.com/drugs/800026843

Parkinson's disease (PD) is one of the most common neurodegenerative disorders and is becoming of increasing relevance to the ageing populations of the developed world.25, 26 While a vast number of investigations and discussions around the involvement of CCR2 in inflammatory conditions such as atherosclerosis have been published over the past two decades, its potential importance in PD has primarily been highlighted in more recent years.27, 28, 29, 30 The evident upregulation of CCR2 in PD and the lack of clarity regarding its precise involvement in the pathogenesis of the disease combined with the increasing importance of PD in modern society warrant further studies into CCR2 as a potential targetable biomarker for imaging purposes.31, 32

Moreover, among a range of cytokines, chemokines, and growth factors examined in a recent study of multiple‐system atrophy cerebellar‐type (MSA‐C) patients by Yamasaki et al, only CCL2 had a significant negative correlation with disease duration.33

With regard to AD, an aggressive transgenic model of AD demonstrated that a deficiency in CCR2 accelerated disease progression and increased mortality.19 However, a recently published 2‐year study by Lee et al indicated that increased‐plasma CCL2 levels could be associated with increased severity as well as more rapid cognitive decline.34 These and several other scientific reports highlight the complex role of CCR2 in AD, warranting investigations into this potentially important therapeutic target in the central nervous system (CNS).35

Therefore, we set out to study a potential radioligand for imaging CCR2 in vitro with the possibility of translating said ligand for in vivo molecular imaging with positron emission tomography (PET). In vivo PET imaging allows addressing questions central to drug development, for instance, through bio‐distribution studies to assess whether the target tissue is reached in sufficient drug quantities for pharmacological efficacy, which is crucial for CNS disorders. Furthermore, provided with the adequate parameters, target occupancy relative to drug plasma concentration can be established, which in turn can be used to direct dose selection. As a result, PET imaging can reduce the attrition of new drugs, particularly in late development stages. The cost of bringing a new drug from its synthesis to the clinical market is approximately 1 billion USD,36, 37, 38, 39 and the time required for clinical development of CNS drug compounds is particularly long—at approximately 8.1 years (10 years if the approval phase is included).39 Therefore, a radioligand for an important target such as CCR2 could be of significant value to neuroscience research and aid the development of new medicines for the CNS.

RESULTS AND DISCUSSION

Antagonist selection

Efforts by various pharmaceutical companies have yielded a wide range of small molecule antagonists with several showing nanomolar or even sub‐nanomolar binding affinities for CCR2. A few promising drug candidates are shown in Figure 1. Compound 1, also known as AZ12567889, has been reported by AstraZeneca to stand out because of its highly potent rodent affinity as well as good pharmacokinetic properties including CNS penetration.40 Furthermore, it has been used in a preclinical model of neuropathic pain to reverse hyperalgesia.41 Compounds 2 (INCB8761/PF‐4136309)42 and 3 (INCB3284)43 both exhibited potent hCCR2 activity, combined with high selectivity over other chemokine receptors and GPCRs and reasonable oral bioavailability. Johnsson & Johnsson disclosed the development of a number of substituted dipiperidine alcohols as potent CCR2 antagonists.44 The key compound, with an hCCR2 IC50 of just 2.4 (±2.0) nM, showed high selectivity against CCR1, CCR3, CCR4, CCR5, CCR6, CCR7, and CCR8.45

Figure 1

Selected drug candidates with IC50 values for the C‐C chemokine receptor 2 below 10 nM

While it was primarily considered for the therapy of inflammatory conditions outside the CNS—showing significant in vivo efficacy in adjuvant‐induced arthritis, collagen‐induced arthritis and allergic asthma models—it also appeared as a promising candidate for use in the human brain. Compound 4 adhered rather well to the Lipinski's rule of 5,46 and based on the sum of oxygen and nitrogen atoms, it appeared to have a better chance of passing the blood‐brain barrier (BBB) than other candidates. Furthermore, it offered a superior binding affinity for CCR2 as well as the option to introduce both [18]F and [11]C into the molecule without having to alter the structure. The molecular structures of the other three candidates solely allow for the use of [11]C, since they do not feature aromatic fluorine but rather trifluoromethyl groups. While these are amenable to labelling with [18]F, the specific activity that can be achieved remains lower than desirable for studying receptors or other low‐density targets, despite recent improvements.47

To further investigate the potential of compound 4, it was synthesised according to the procedure of Xia et al45 (see Section 4).

Its antagonist and agonist potential against hCCR2 and hCCR5 were measured, confirming both the IC50 values for hCCR2 and the selectivity over CCR5 with which it shares 75% homology.48 Based on these results, 4 was chosen as a tritiation target in order to perform autoradiography (ARG) on human brain sections.

Iodination and deuteration

The indole moiety was deemed most suitable for labelling via a facile tritio‐dehalogenation protocol. As the preparation of the CCR2 antagonist 4 was performed on a small scale, only very limited amounts of it were immediately available. However, the multi‐step synthesis of 4 proceeds via intermediate 5 (shown in Scheme 1), which was present in abundance and ready to use. Furthermore, it has been reported in the literature that under standard reaction conditions for tritio‐dehalogenations, using Pd/C and tritium gas, C═C double bonds may undergo saturation or isomerisation,49, 50 even at low tritium pressures.51 Intermediate 5, unlike 4, does not contain a double bond and, therefore, did not appear at risk to undergo an unwanted reduction. Based on the original synthesis by Xia et al,45 a straightforward route to [3H]4 was envisioned, in which the alkene moiety would be introduced after incorporating tritium. This option seemed to minimise the inherent risk and could be directly implemented without the need to synthesise more of 4.

SCHEME 1

Synthesis of the iodinated precursor 6 for subsequent tritium labelling. The yield of just 14% is explained by the challenging separation of the regio‐isomers formed, only one of which was used further

As expected, the iodination of 5 using N‐iodosuccinimide (NIS) yielded a mixture of regio‐isomers with the iodination occurring at several positions on the indole ring. While challenging, it was possible to separate the major isomer 6 (see Scheme 1) using mass‐directed supercritical fluid chromatography (SFC‐MS). The structural configuration of the major isomer was determined by NMR, in particular nuclear Overhauser effect spectroscopy (NOESY) and rotating frame nuclear Overhauser effect spectroscopy (ROESY). It may be worth noting that reducing the ethyl ester before the iodination was not pursued as, in that case, the separation of the regio‐isomers after the iodination was unsuccessful (see Section 4, compound 14).

To swiftly confirm the feasibility of the reaction sequence from 6 to [3H]4 (outlined in Scheme 2), it was first tried with deuterium. The first step using deuterium gas and 10% Pd/C proceeded with a good conversion rate of about 80%. The lithium aluminium hydride reduction appeared challenging because of the small reaction scale; nevertheless, the desired product (4‐d 1) could be formed successfully, as evidenced by LCMS. This encouraged us to pursue this path further and repeat the sequence using tritium.

SCHEME 2

Envisioned synthesis sequence from 6 to [3H]4. This route was first tried with deuterium, successfully yielding crude 4‐d 1

Tritium radiolabelling

Interestingly, it was found that after the first step, namely, the tritio‐dehalogenation of 6, the major product was not the desired compound. That would have had an MS ES+ peak at 372 m/z; yet, instead, a peak with a distinct pattern showing 398, 400, 402, 404, 406, and 408 m/z was observed. Another noticeably smaller peak exhibited a similar MS pattern with 384, 386, 388, and 390 m/z. The approximate ratio by high‐performance liquid chromatography (HPLC) was 15:1 (by UV at 280 nm), assuming that, at 280 nm, the indole remains the primary source of absorption. Clearly, an unexpected side reaction was taking place, calling for further investigation into the process. The two main reaction products were isolated by preparative HPLC; and 1H NMR and 3H NMR spectra were taken for both samples.

On the basis of the combined evidence from LCMS and NMR measurements, we propose that a very small fraction of the solvent was oxidised from ethanol to ethanal. This aldehyde then reacted with the secondary amine of the piperidine moiety of 6 to form the iminium ion as outlined in Figure 2. Subsequently, this iminium ion functionality was reduced in the presence of tritium gas yielding the N‐ethylated derivative (cf Figure 3) of the expected product (7).

Figure 2

Proposed mechanism adapted from Corma et al57 for the formation of the iminium ion species, which may then undergo double‐bond migration prior to reduction in the presence of 3H(g) and 10% Pd/C, yielding the mixture of radioactive compounds displayed in Figure 3

Figure 3

Mixture of products obtained from the observed side reaction with the m/z peaks on the mass spectrum corresponding to the respective number of tritium atoms incorporated into the structure. aThis refers to the number of tritium atoms on the ethyl side chain and does not include the R part. bThe m/z value of 398 corresponds to the nonradioactive compound where R also contains no tritium

At this point, it might be worth noting that the initial deuterium trial reaction, which was performed thrice to present a meaningful comparison, consistently showed 80% to 85% (by UV) conversion to the desired product. However, LCMS analysis indicated that 15% to 20% was converted to what appeared to be the N‐ethylated side product (no methanol was used in this instance). We are not certain of the reason for this considerable difference in reactivity. Perhaps, it was due to an unexplained isotopic effect or the fact that the reaction conditions could not be perfectly reproduced in the tritium experiment.

It was first observed that alcohols are oxidised by Pd(II), or at least that Pd(II) is reduced to Pd(0) in the presence of alcohol, when Berzelius noted that metal precipitate was formed out of a heated alcoholic solution of Pd(II) in 1828.52 Pd/C generally contains a mixture of Pd(0) and Pd(II) species, and Pd/C with varying reduction degrees is commercially available.53 Furthermore, palladium on charcoal has extensively been used for the aerobic oxidation of alcohols.54 Since the catalyst was in ethanol and exposed to air prior to the freeze‐thawing cycles, this could have taken place.

Alternatively, Pd(II) could have been irreversibly reduced at a later, anaerobic stage. Moreover, the addition of alcohols to amines, via oxidation followed by a reductive amination, has been reported using a range of catalysts, including ruthenium55 and more importantly palladium.56, 57 The mechanism of “chain walking”, first introduced by Möhring et al,58 is a process involving β‐hydride elimination usually followed by immediate reinsertion, thus altering the regio‐chemistry observed.59 Its relevance in the context of this side reaction is that it could explain the mixture of products observed. Importantly, Culf et al have previously reported that such double‐bond‐migration processes can yield products of uncertain tritium distribution when using 10% Pd/C to reduce an alkene in methanol.60 Furthermore, there is debate about the precise mechanism involved in the amination of alcohols in the presence of palladium catalysts, and it is possible that an intermolecular transfer hydrogenation from another amine takes place in the process,61 contributing to the range of products observed.

The pattern in the mass spectrum and the ratio of integrals in the 3H NMR spectrum can be rationalised by considering the incorporation of several 3H atoms as outlined in Figure 3. Both non‐aromatic 3H peaks in the spectrum were broadened, which is in line with the different isotopic shifts of the possible species. The integral ratio between the two aliphatic peaks at 2.58 and 1.14 ppm in CD3OD was approximately 1:1.33, slightly below the 1:1.5 that would be expected if incorporation at the –CH3 vs the –CH2– position was completely random (Figure 3).

The second, and much more minor, side product can be explained in a similar fashion. During the lyophilisation process, ethanol was evaporated, followed by the addition of methanol and evaporation under nitrogen flow, which was repeated twice before passing a suspension in methanol through a syringe filter. It appears that during this process, a small amount of methanal was formed, which led to the N‐methylated product in a similar manner as before. The 3H NMR spectrum only shows one aliphatic peak at 2.33 (CD3OD) as expected, and the observed mass distribution corresponds to 0, 1, 2, or 3 tritium atoms being incorporated on the methyl group. On the basis of these investigations, we attempted to use a different solvent, which would not be able to react with the secondary amine. Acetonitrile was chosen, and, as expected, neither of the two previous side products could be detected. However, the conversion was inefficient with only 700 MBq measured for the crude reaction mixture and 200 MBq after purification through preparative HPLC, with a theoretical radiochemical yield of 5.1 GBq. The LiAlH4 reduction was attempted on the purified intermediate, yet the yield was again very low and insufficient to carry on further. This may be due to the fact that some of the desired products seemed to stick to the aluminium salts formed. Thus, it appeared that a change in solvent would not remedy the problem.

Hence, we decided to try the direct path to [3H]4, to investigate whether short reaction times and low tritium pressure would be sufficient to keep the alkene functionality intact. To this end, we synthesised another batch of 4, to prepare an iodinated precursor for radiolabelling.

The indole moiety of 4 was iodinated in an analogous manner before using N‐iodosuccinimide (NIS) and trifluoroacetic acid (TFA) in dichloromethane (DCM), yielding compound 10. When the reaction was tried with deuterium at a pressure of approximately 50 mbar and reaction times of both 15 minutes and 1 hour, fortunately, conversion to the desired product was achieved without affecting the C═C double bond. Hence, this reaction was carried out with tritium as shown in Scheme 3, successfully yielding [3H]4.

SCHEME 3

The successful radio‐synthesis of [3H]4, using 10% Pd/Ca2CO3, a low tritium pressure of approximately 50 mbar, and a reaction time of 1 h to keep the alkene functionality intact

Autoradiography

With the radioligand in hand, it was possible to perform in vitro ARG studies on fresh frozen human brain slices to establish whether specific binding takes place in PD tissue. In addition to the PD sample, an MSA case, 2 AD cases (AD1 and AD2), and a control was examined (demographics are shown in Table 1).

Table 1

Demographics of the human brain tissue from the Netherlands Brain Bank included in this study for the autoradiography experiment

Sample	Diagnosis	Brain Region	Age, y	Gender	PMD
PD	Parkinson's disease	Cingulate gyrus	82	M	06:05
MSA	Multiple system atrophy	Cerebellum	69	F	04:30
CT	Control	Cingulate gyrus	60	F	07:30
AD1	Alzheimer's disease	Frontal cortex	88	F	05:35
AD2	Alzheimer's disease	Frontal cortex	90	F	20:50

Note. Includes the sample, diagnosis, brain region, age (y), gender (M/F), and PMD.

Abbreviations: MSA, multiple‐system atrophy; PMD, post‐mortem delay.

As shown in Figure 4, in all cases the binding of [3H]4 at both 1 and 3 nM of concentration was not blocked by excess of unlabelled 4 compound (10 μM), suggesting a nonspecific binding of the radioligand (ie, to cell membranes and other high‐capacity binding sites). One explanation for this result could be the relatively high logP value of around 4 leading to presence of the antagonist in lipid tissue. Another possible reason could be the ubiquitous expression of CCR2. Thus, more experiments are currently ongoing in order to develop a less lipophilic and more specific radioligand for CCR2.

Figure 4

Autoradiograms showing [3H]4 total and nonspecific (NS) binding at 3 nM (A) and 1 nM (B) in human brain sections from a Parkinson's disease case (PD), a multiple‐system atrophy (MSA) case, a nondemented control (CT), and a couple of Alzheimer's disease cases (AD1 and AD2). Displacement experiments with 4 (10 μM) did not show significant blocking, indicating low specific binding levels regardless of whether a 1 (C) or 3 nM (D) of concentration was used

CONCLUSION

In summary, CCR2 was envisioned as a potential target for ARG and eventually PET studies on neurodegenerative diseases. Based on previous development efforts, a promising small molecule antagonist (4) was selected and synthesised, and its binding affinity was confirmed. The initial attempt to create a tritium‐labelled analogue of 4 via intermediate 5 gave rise to an interesting and unique side reaction profile. The use of Pd/C in ethanol as well as other alcohols is very common not only for tritium labelling reactions but also in organic chemistry in general. Therefore, we believe that this work can act as a cautionary tale to other chemists, in particular radiochemists, who may encounter this type of side reaction. This would be especially relevant to anyone working with compounds featuring amines that could be subject to this type of alkylation. After, eventually, preparing [3H]4 and learning that direct labelling presented a simpler solution than initially thought, the radioligand was used for in vitro ARG measurements on human brain slices. Unfortunately, the total binding obtained was not blocked by an excess of the nonlabelled “cold” compound (10 μM), suggesting that most of the signal corresponds to nonspecific binding. Thus, further work is needed to establish whether CCR2 antagonists are viable in this context or whether the ubiquitous expression of this receptor generally precludes any impactful measurements.

EXPERIMENTAL SECTION

General

All chemicals were purchased from Combi Blocks, Sigma‐Aldrich, or its subsidiaries. Tritium was obtained from RC Tritec AG. Anhydrous solvents were obtained from Aldrich and were used without further purification. Reactions were magnetically stirred unless otherwise noted. Tritium reactions were performed on an RC Tritec Tritium manifold.

1H NMR and 13C NMR spectra were acquired on a Bruker Avance III spectrometer running at a proton frequency of 500.1 MHz and fitted with a cryogenic probe or on Bruker Avance Nanobay spectrometers operating at 400 MHz. Chemical shifts are reported in ppm (δ) relative to tetramethylsilane (TMS) with the solvent resonance as internal standard (5.32 and 53.84 ppm for CD2Cl2, 7.26 and 77.16 ppm for CDCl3, 3.31 and 49.00 ppm for CD3OD, and 2.50 and 39.52 ppm for DMSO‐d 6).62 The signals derived from the 1H NMR spectra are reported with chemical shifts, multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, p = pentet, sxt = sextet, dt = doublet of triplets, dq = doublet of quartets, ddd = double doublet of doublets, ddt = double doublet of triplets, td = triplet of doublets, tt = triplet of triplets, qd = quartet of doublets, or m = multiplet), coupling constant (Hz), and integration. Data for 13C NMR are reported with chemical shifts and coupling to 19F, where observed. Flash column chromatography was carried out using pre‐packed silica gel columns supplied by Biotage and using a Biotage automated flash systems with UV detection.

HPLC was performed using a system composed of a Gilson 322 Pump equipped with a Gilson UV/VIS‐152 detector with an Xbridge Prep C‐18 10 μm OBD 19 × 250 mm column. Separation of enantiomers was performed on a Nova SuperSep150 and analysed by Waters Acquity UPC2. In the radioactive reaction sequence, HPLC was performed using a system composed of a Gilson 322 Pump equipped with a Gilson Ready 2000 detector with an XbridgeTM Prep C‐18 5 μm OBDTM 19 × 100 mm column.

LCMS were acquired on a Waters Acquity UPLC using a BEH C18 column (50 mm × 2.1 mm, 1.7‐μm particles) with a gradient of 10% to 90% over 4 min with MeCN–NH4/NH4CO3 or MeCN–formic acid and electrospray ionisation.

Precursor synthesis

‐Butyl 4‐(1‐bromo‐2‐ethoxy‐2‐oxoethyl) piperidine‐1‐carboxylate (11)

An oven‐dried, two‐neck, round‐bottom flask equipped with a dropping funnel was flushed with nitrogen and cooled to −78°C. Lithium bis(trimethylsilyl)amide in THF (1M, 100 mL, 100 mmol) was added, followed by tert‐butyl 4‐(2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (15.1 g, 55.5 mmol) in dry THF (55 mL). The orange solution was stirred at −78°C for 3 hours under a blanket of N2. Then chlorotrimethylsilane (12.69 mL, 100 mmol) was added dropwise at −78°C, and the mixture was stirred for another hour. Subsequently, dibromine (3.44 mL, 66.6 mmol) was added dropwise, and the dropping funnel was then washed with a little dry THF (approximately 5 mL). The mixture was stirred for another 2 hours; then it was warmed to 0°C and stirred for another hour under N2. The mixture was diluted with ethyl acetate (approximately 200 mL) and washed with sat. NaHCO3 solution (120 mL) and then with water (150 mL). The organic phases were dried over a phase separator cartridge and concentrated under reduced pressure to afford a yellow oil. The residue was purified by passing it through a Biotage samplet cartridge (Biotage SNAP Samplet KP‐SIL 34 g, 6/cs), washing with a 50:50 ethyl acetate:n‐heptane mixture. Note: This rapid purification step was chosen since it appeared as if the product degraded over time. The solvent was removed on the rotary evaporator. A yellow oil, tert‐butyl 4‐(1‐bromo‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (11) (22.0 g, approximately 50% pure by UV), was obtained that was used directly in the next step.

LC/MS (M + 1): 350.2 (82%), 352.2 (100%).

13C NMR (101 MHz, CD3Cl): δ (ppm) 168.9, 154.5, 79.5, 61.9, 51.6, 49.1, 39.8, 28.4, 22.7, 14.0.

‐Butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (5)

Under air, crude tert‐butyl 4‐(1‐bromo‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (11) (11.0 g, 31.41 mmol) (the actual amount of reactant is not truly known as it was used directly without thorough purification to avoid degradation), 3‐(piperidin‐4‐yl)‐1H‐indole (10.1 g, 50.25 mmol), and triethylamine (8.75 mL, 62.81 mmol) were combined in MeCN (80 mL) and refluxed overnight. The solvent was removed on the rotavapor, and the orange/brown solid was purified via automated flash column chromatography (98:2 → 80:20, DCM:MeOH; 340 g of SNAP column) to afford an orange oil. This was then further purified again via automated flash column chromatography (88:12 → 0:100, n‐heptane:ethyl acetate), to afford tert‐butyl 4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (3.25 g) as an off‐white solid. This racemate was separated by chiral SFC (25% EtOH/DEA 100/0.5 in CO2, 120 bar, 40°C, CelluCoat 5 μM 250 × 30 mm column, 100 mL/min), which afforded tert‐butyl (R)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (1.39 g, 2.95 mmol) and tert‐butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (5) (1.39 g, 2.95 mmol, 9.5%) as off‐white solids.

LC/MS (M + 1): 470.5 (100%).

1H NMR (400 MHz, CD2Cl2): δ (ppm) 8.28 (s, 1H), 7.61 (d, J = 7.9 Hz, 1H), 7.35 (d, J = 8.0 Hz, 1H), 7.15 (t, J = 7.5 Hz, 1H), 7.06 (t, J = 7.4 Hz, 1H), 6.98 (s, 1H), 4.21 (tt, J = 12.3, 5.4 Hz, 2H), 4.02 to 4.14 (m, 2H), 2.67 to 3.02 (m, 6H), 2.60 (t, J = 10.8 Hz, 1H), 2.34 (t, J = 11.1 Hz, 1H), 1.92 to 2.06 (m, 4H), 1.80 (td, J = 12.3, 12.2, 3.7 Hz, 1H), 1.67 (dd, J = 12.3, 3.2 Hz, 1H), 1.48 to 1.54 (m, 1H), 1.45 (s, 9H), 1.31 (d, J = 7.1 Hz, 3H), 1.13 (ddt, J = 17.4, 11.6, 5.0 Hz, 2H).

13C NMR (101 MHz, CD2Cl2): δ (ppm) 171.1, 155.1, 136.9, 127.1, 122.1, 121.8, 120.1, 119.4, 119.2, 111.5, 79.4, 73.2, 60.2, 47.3, 44.9, 44.4, 35.1, 34.3, 34.2, 34.0, 29.9, 29.3, 28.6, 27.1, 14.9.

‐Butyl (R)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate.

[α]20 D + 35° (c = 1.0, acetonitrile).

‐Butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (5).

[α]20 D − 36° (c = 1.0, acetonitrile).

‐Butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidine‐1‐carboxylate (12)

An oven‐dried, round‐bottom flask was evacuated and backfilled with nitrogen three times. It was cooled to 0°C, and lithium aluminium hydride in THF (1M, 958 μL, 0.96 mmol) was added, followed by dropwise addition of tert‐butyl 4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (5) (300 mg, 0.64 mmol) in dry THF (18 mL) under N2. The mixture was stirred at 0°C for 1 hour before being allowed to warm to room temperature (RT). After being stirred for another hour at RT, water (40 μL) was slowly added to quench, followed by 4M NaOH (40 μL) and, once again, water (120 μL). The mixture was stirred for 30 minutes and then filtered through a celite pad to remove the solids, and the layers were separated. The aqueous phase was washed twice with ethyl acetate. The organic layers were combined, washed with brine, and dried over a phase separator cartridge. The solvent was removed on the rotary evaporator to afford tert‐butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidine‐1‐carboxylate (12) (270 mg, 99%) as a yellow solid.

LC/MS (M + 1): 428.4 (100%).

1H NMR (400 MHz, CD3Cl): δ (ppm) 8.25 (s, 1H), 7.63 (d, J = 7.9 Hz, 1H), 7.35 (d, J = 8.1 Hz, 1H), 7.18 (t, J = 7.5 Hz, 1H), 7.10 (t, J = 7.4 Hz, 1H), 6.95 (d, J = 1.8 Hz, 1H), 3.58 (dd, J = 9.5, 4.5 Hz, 1H), 3.28 (td, J = 10.2, 7.5 Hz, 1H), 3.03 to 3.16 (m, 1H), 2.82 to 3.01 (m, 4H), 2.69 (d, J = 14.9 Hz, 1H), 2.60 (ddd, J = 15.2, 8.7, 4.8 Hz, 1H), 2.43 to 2.56 (m, 1H), 2.27 (s, 1H), 2.08 (d, J = 6.4 Hz, 2H), 1.85 to 1.95 (m, 2H), 1.59 to 1.82 (m, 3H), 1.53 (td, J = 12.2, 3.5 Hz, 1H), 1.48 (s, 9H), 1.44 (s, 1H), 1.27 to 1.4 (m, 2H).

13C NMR (101 MHz, CD3Cl): δ (ppm) 171.3, 154.9, 136.5, 126.7, 126.6, 125.62, 122.0, 122.0, 121.4, 121.3, 119.8, 119.2, 119.1, 119.1, 111.4, 79.6, 69.7, 69.6, 60.5, 58.6, 58.5, 55.8, 55.8, 55.1, 54.9, 46.5, 45.2, 44.9, 36.2, 35.3, 34.7, 34.7, 34.1, 34.1, 34.0, 34.0, 32.1, 31.7, 30.4, 29.0, 28.9, 28.6, 21.2, 14.3.

(S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (13)

To tert‐butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidine‐1‐carboxylate (12) (35 mg, 0.08 mmol) in DCM (15 mL) was added hydrogen chloride in diethyl ether (2.0 M, 409 μL, 0.82 mmol) at RT under air. A sudden colour change from yellow to orange was observed, and red precipitate was formed. The mixture was stirred at RT overnight. The solvent was removed on the rotavapor; and the orange solid, (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (13) (27 mg, 0.07 mmol, 91%), was used in the next step without further purification.

LC/MS (M + 1): 328.4 (100%), 329.4 (21%).

(E)‐3‐(3,4,5‐Trifluorophenyl)acryloyl chloride (9)

(E)‐3‐(3,4,5‐Trifluorophenyl)acrylic acid (500 mg, 2.47 mmol) was dissolved in DCM (15 mL), and two drops of dimethylformamide (DMF) were added. Oxalyl dichloride (419 μL, 4.95 mmol) was slowly added at RT. The clear solution was stirred at RT under air for 2 hours. The resulting mixture was concentrated under reduced pressure to afford (E)‐3‐(3,4,5‐trifluorophenyl)acryloyl chloride (9) (540 mg, 2.45 mmol, 99%) as a yellow oil in quantitative yield, which was used directly in the next step.

1H NMR (400 MHz, CD3Cl): δ (ppm) 7.65 (d, J = 15.6 Hz, 1H), 7.21 (dd, J = 7.6, 6.6 Hz, 2H), 6.56 (d, J = 15.6 Hz, 1H).

13C NMR (101 MHz, CD3Cl): δ (ppm) 166.3, 166.2, 165.6, 164.0, 147.2, 146.8, 124.9, 124.8, 113.3, 113.2, 113.0, 105.1.

(S,E)‐1‐(4‐(1‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (4)

2‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (13) (26 mg, 0.08 mmol) and (E)‐3‐(3,4,5‐trifluorophenyl)acryloyl chloride (9) (17.51 mg, 0.08 mmol) were combined in dry THF (6 mL) followed by a few drops of triethylamine (11.07 μL, 0.08 mmol) under air. The mixture was stirred at RT for 5 hours. The solid residue was purified via HPLC (30‐90% MeCN–0.2% NH4OH over 25 min, XbridgeTM Prep C‐18 10 μm OBDTM 19 × 250 mm column, 20 mL/min). The solvent was removed under reduced pressure to afford a white solid, (S,E)‐1‐(4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (4) (22.0 mg, 54.2%).

LC/MS (M + 1): 512.3 (100%), 513.3 (31%), 514.3 (3%).

1H NMR (400 MHz, CD3Cl): δ (ppm) 8.05 (s, 1H), 7.62 (d, J = 7.0 Hz, 1H), 7.49 (d, J = 15.3 Hz, 1H), 7.37 (d, J = 7.9 Hz, 1H), 7.19 (t, J = 7.3 Hz, 1H), 7.09 to 7.16 (m, 3H), 6.97 (s, 1H), 6.82 (d, J = 15.1 Hz, 1H), 4.75 (s, 1H), 4.10 (s, 1H), 3.59 (s, 1H), 3.49 (s, 1H), 3.25 to 3.39 (m, 1H), 3.03 to 3.22 (m, 2H), 2.85 to 3.01 (m, 3H), 2.59 to 2.74 (m, 2H), 2.50 (s, 1H), 2.06 to 2.15 (m, 3H), 1.94 (d, J = 8.3 Hz, 1H), 1.63 to 1.86 (m, 3H), 1.35 to 1.49 (m, 1H), 1.19 to 1.28 (m, 1H).

13C NMR (101 MHz, CD3Cl): δ (ppm) 164.4, 152.8, 152.7, 150.3, 139.6, 136.5, 131.7, 126.7, 122.2, 121.3, 119.8, 119.8, 119.3, 119.1, 111.8, 111.8, 111.6, 111.4, 69.5, 58.6, 54.6, 54.6, 51.0, 46.2, 45.8, 45.4, 42.6, 36.4, 36.2, 34.6, 34.1, 33.9, 32.7, 31.5, 29.9, 28.8.

Iodination

(S)‐2‐(4‐(5‐Iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (14) (failed)

Under air, tert‐butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidine‐1‐carboxylate (13) (127 mg, 0.30 mmol) and NIS (80 mg, 0.36 mmol) were combined in DCM/TFA (1:1 v/v, 3 mL). The solution quickly turned dark red/purple and was stirred at RT overnight. The dark solution was dried under nitrogen flow. The solid residue was further purified via HPLC (20‐95% MeCN–0.1% TFA over 25 min, XbridgeTM Prep C‐18 10 μm OBDTM 19 × 250 mm column, 20 mL/min), which afforded a mixture of regio‐isomers that could unfortunately not be successfully separated.

Note: Attempts to separate these were to no avail, since the compound was not stable on the HPLC column at pH 10, coelution occurred at pH 3, and it appeared to degrade upon storage in DMSO.

LC/MS (M + 1) of isomeric mixture: 454.2 (100%).

Ethyl (S)‐2‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (6)

Under air, tert‐butyl (S)‐4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐ethoxy‐2‐oxoethyl)piperidine‐1‐carboxylate (5) (84 mg, 0.18 mmol) and NIS (48.3 mg, 0.21 mmol) were combined in DCM/TFA (1:1 v/v, 3 mL). The solution quickly turned dark red/purple and was stirred at RT overnight. The dark solution was dried under nitrogen flow. The solid residue was further purified via preparative HPLC (20‐80% MeCN–0.1% TFA over 25 min, XbridgeTM Prep C‐18 10 μm OBDTM 19 × 250 mm column, 20 mL/min). The solvent was removed on the rotavapor, and the isomers were separated via SFC (MeOH/H2O/NH3 97/3/50 mM, Waters Prep 100 SFC MS with a Waters BEH 5 μm 30 × 250 mm column) ethyl (S)‐2‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (6) (12.0 mg, 13.54%) was obtained as an off‐white solid.

LC/MS (M + 1): 496.2 (100%), 497.2 (24%).

1H NMR (400 MHz, CD3Cl): δ (ppm) 8.02 (s, 1H), 7.94 (d, J = 1.4 Hz, 1H), 7.42 (dd, J = 8.5, 1.6 Hz, 1H), 7.13 (d, J = 8.5 Hz, 1H), 6.92 (s, 1H), 4.17 to 4.27 (m, 2H), 3.11 to 3.25 (m, 2H), 2.96 (d, J = 10.4 Hz, 2H), 2.87 (d, J = 11.5 Hz, 1H), 2.52 to 2.76 (m, 4H), 2.25 to 2.37 (m, 1H), 1.99 (d, J = 13.2 Hz, 5H), 1.76 (td, J = 12.4, 12.3, 3.8 Hz, 1H), 1.55 to 1.64 (m, 2H), 1.21 to 1.37 (m, 5H).

13C NMR (101 MHz, CD3Cl): δ (ppm) 171.1, 171.0, 135.5, 130.3, 129.5, 128.2, 121.3, 120.5, 113.3, 82.7, 73.0, 60.1, 54.1, 47.0, 46.1, 45.8, 34.5, 33.9, 33.8, 33.6, 29.9, 29.3, 14.9.

(S,E)‐1‐(4‐(2‐Hydroxy‐1‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)ethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (10)

Under air, (S,E)‐1‐(4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (4) (20 mg, 0.04 mmol) and NIS (11.43 mg, 0.05 mmol) were combined in DCM/TFA (3:1 v/v, 2 mL) and stirred overnight at RT. The solvent was removed on the rotavapor, and the light brown residue was purified via preparative HPLC (20‐80% MeCN–0.1% TFA over 30 min, XbridgeTM Prep C‐18 10 μm OBDTM 19 × 250 mm column, 20 mL/min) to afford a beige solid, (S,E)‐1‐(4‐(2‐hydroxy‐1‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)ethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (10) (17.0 mg, 68.2%).

LC/MS (M + 1): 638.2 (100%).

Deuterium trial reaction

Ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate‐d 1 (15)

Ethyl 2‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (6) (2 mg, 4.04 μmol), palladium on activated charcoal (2 mg, 0.02 mmol), and triethylamine (40 μL, 4.04 μmol) were combined in ethanol (0.5 mL). The mixture was frozen in liquid nitrogen, and the flask was evacuated before closing the vacuum tap and being allowed to warm to RT. This was performed twice. Then the mixture was frozen once more, and deuterium gas was pulled into the flask (which was still under vacuum; ie, the tap was closed). The mixture was allowed to warm to RT and stirred for 4 hours. The suspension was diluted with ethanol (5 mL) and filtered through a phase separator cartridge, and the solvent removed on the rotovap. The product was afforded as a yellow solid that was used directly in the next step.

LC/MS (M + 1): 371.3 (100%).

(S)‐2‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol‐d 1 (16)

An oven‐dried, round‐bottom flask was evacuated and backfilled with nitrogen three times. To it was added lithium aluminium hydride (0.22 mL, 0.22 mmol) in THF (1M), and it was then cooled to 0°C. Another flask containing ethyl 2‐(4‐(1H‐indol‐3‐yl‐5‐d)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate‐d 1 (2 mg, 5.40 μmol) was evacuated and backfilled with nitrogen three times before adding anhydrous THF (3 mL). The yellow solution was added slowly to the LiAlH4 under N2. The mixture was stirred at 0°C for 1 hour before being allowed to warm to RT. It was stirred for another 1.5 hours at RT. Water (0.5 mL) was slowly added to quench, followed by 4M of NaOH (1 mL) and water (0.5 mL). The mixture was filtered over a syringe filter and dried over a phase separator cartridge, and the aqueous phase was washed twice with ethyl acetate (still on the phase separator). The solvent was removed on the rotary evaporator to yield a solid.

LC/MS (M + 1): 329.2 (100%).

(S,E)‐1‐(4‐(1‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (4‐d 1)

To 2‐(4‐(1H‐indol‐3‐yl‐5‐d)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (16) (1.5 mg, 4.57 μmol) was added (E)‐3‐(3,4,5‐trifluorophenyl)acryloyl chloride (9) (4 mg, 0.02 mmol), and the flask was evacuated and backfilled with nitrogen three times. Then dry THF (3 mL) followed by a few drops of triethylamine (0.637 μL, 4.57 μmol) was added under N2. The mixture was stirred at RT for 3 hours; then the solvent was removed on the rotovap. LCMS indicated formation of (S,E)‐1‐(4‐(1‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl) piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl) prop‐2‐en‐1‐one (4‐d 1).

LC/MS (M + 1): 513.3 (100%).

Radiochemistry—synthesis

[3H] ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (7)

Approach 1

Ethyl (S)‐2‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (6) (2.4 mg, 4.84 μmol), palladium on activated charcoal (2.2 mg, 0.02 mmol), and triethylamine (40 μL, 4.84 μmol) were combined in ethanol (0.5 mL). The mixture was frozen in liquid nitrogen, and the flask evacuated before closing the vacuum tap and being allowed to warm to RT. This was performed twice. The mixture was frozen once more and the flask evacuated; then tritium was released from the uranium bed by heating it with a blowtorch. The reaction was run with a partial pressure of tritium gas (337.3 GBq) and was warmed to RT with stirring for 3 hours. The tritium gas in the flask was blown of together with the solvent under nitrogen flow. Then 0.5 mL of methanol was added and the solvent evaporated as before. This was performed twice. It was taken up in methanol once more and then passed through a syringe filter, and the solvent was evaporated under reduced pressure to yield a solid (2.5 GBq), which was purified via preparative HPLC (10‐80% MeCN–0.2% NH3OH over 30 min, Xbridge Prep C‐18 5 μm OBD 19 × 100 mm column, 10 mL/min).

[3H] ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(1‐ethylpiperidin‐4‐yl)acetate (17):

LC/MS (M + 1): 398.3 (27%), 400.3 (100%), 402.3 (98%), 404.3 (44%), 406.3 (28%), 408.3 (11%), 410.3 (1%).

3H NMR (533 MHz, CD3OD): δ (ppm) 7.04 (t, J = 7.8 Hz), 2.58 (b), 1.14 (b).

[3H] ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(1‐methylpiperidin‐4‐yl)acetate (18).

LC/MS (M + 1): 384.2 (26%), 386.2 (100%), 488.3 (74%), 490.3 (3%).

3H NMR (533 MHz, CD3OD): δ (ppm) 7.04 (t, J = 7.7 Hz), 2.33 (b).

Approach 2

Ethyl 2‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (6) (2.4 mg, 4.84 μmol), palladium on activated charcoal (2.2 mg, 0.02 mmol), and triethylamine (40 μL, 4.84 μmol) were combined in acetonitrile (0.7 mL). The mixture was frozen in liquid nitrogen, and the flask evacuated before closing the vacuum tap and being allowed to warm to RT. This was performed twice. The mixture was frozen once more and the flask evacuated; then tritium was released from the uranium bed by heating it with a blowtorch. The reaction was run with a partial pressure of tritium gas (336.4 GBq) and was warmed to RT with stirring for 3 hours. The tritium gas in the flask was blown of together with the solvent under nitrogen flow. Then 0.5 mL of acetonitrile was added and the solvent evaporated as before. This was performed twice. The mixture was then passed through a syringe filter, and the solvent was evaporated under a stream of nitrogen to yield a yellow solid, (approximately 700 MBq). This was purified by preparative HPLC (25‐75% MeCN–0.2% NH3 over 35 min, Xbridge Prep C‐18 5 μm OBD 19 × 100 mm column, 10 mL/min).

LC/MS (M + 1): 372.2 (100%), 373.2 (30%).

[3H] (S)‐2‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)ethan‐1‐ol (8) (failed)

An oven‐dried, round‐bottom flask was evacuated and backfilled with nitrogen three times. It was charged with lithium aluminium hydride in THF (1M, 0.2 mL, 0.20 mmol) and cooled to 0°C. Another flask containing [3H] ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate (7) (200 MBq) was evacuated and backfilled with nitrogen before adding anhydrous THF (3 mL). The yellow solution containing [3H] ethyl (S)‐2‐(4‐(1H‐indol‐3‐yl)piperidin‐1‐yl)‐2‐(piperidin‐4‐yl)acetate was slowly added to the LiAlH4 solution under N2. The mixture was stirred at 0°C for 1 hour before being allowed to warm to RT. It was stirred for another 2 hours at RT. The mixture was quenched with TFA (50 μL), the solvent evaporated under a stream of nitrogen, and the crude material purified via preparative HPLC (10‐60% MeCN–0.1% TFA over 30 min, Xbridge Prep C‐18 5 μm OBD 19 × 100 mm column, 10 mL/min). Unfortunately, the product could not be detected in any of the fractions or the reaction flask. Most active fraction contained approximately 40 MBq. It is possible that the product stuck to the aluminium salts and remained on the pre‐column filter.

Crude LC/MS (M + 1): 330.2 (100%), 331.2 (11%).

[3H] (S,E)‐1‐(4‐(1‐(4‐(1H‐Indol‐3‐yl)piperidin‐1‐yl)‐2‐hydroxyethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one ([3H]4)

(S,E)‐1‐(4‐(2‐Hydroxy‐1‐(4‐(5‐iodo‐1H‐indol‐3‐yl)piperidin‐1‐yl)ethyl)piperidin‐1‐yl)‐3‐(3,4,5‐trifluorophenyl)prop‐2‐en‐1‐one (10) (0.6 mg, 0.94 μmol), palladium on calcium carbonate (1.944 mg, 0.94 μmol), and triethylamine (20 μL, 0.94 μmol) were combined in ethanol (0.6 mL). The mixture was frozen in liquid nitrogen, and the flask evacuated before closing the vacuum tap and being allowed to warm to RT. This was performed twice. The mixture was frozen once more and the flask evacuated; then tritium (50 mbar) was released from the uranium bed by heating it with a blowtorch. The mixture was warmed to RT and stirred for 1 hour. The tritium gas in the flask was blown off together with the solvent under nitrogen flow. Then 0.5 mL of methanol was added and the solvent evaporated as before. This was performed twice. The crude mixture was taken up in methanol once more and then passed through a syringe filter, and the solvent was evaporated under reduced pressure to yield a solid (approximately 500 MBq) that was purified via preparative HPLC (10‐70% MeCN–0.1% TFA over 30 min, Xbridge Prep C‐18 5 μm OBD 19 × 100 mm column, 10 mL/min). The radioactive purity of the residue (370 MBq, 0.9% radiochemical, and 37% chemical yield) was determined to be 96% by HPLC.

The isotopic incorporation was determined by mass spectrometry to be 16% unlabelled, 82% monolabelled, and 2% dilabelled to give a molar activity of 900 GBq/mmol.

Chemical purity: greater than 90%. Chiral purity: 99 ee.

HPLC analysis method: Waters Xbridge C18 3.5 μm, 4.6 × 100 mm column, −60°C. Eluent A: Water + 10 mM ammonium formate pH 3. Eluent B: MeCN, flow rate 0.6 mL/min, UV detection 254 nm, retention time 14.5 min. Gradient: 0 minute—5% B; 3 minutes—5% B; 25 minutes—95% B; 30 minutes—95% B.

LC/MS (M + 1): 514.3 (100%), 515.2 (37%), 516.3 (14%).

3H NMR (533 MHz, CD3OD): δ (ppm) 7.36 (d, J = 8.6 Hz), 7.09 to 7.14 (m), 7.02 (t, J = 8.0) (integral ratio 0.2:0.4:1.0).

Post‐mortem brain tissues of different pathologies were obtained from the Netherlands Brain Bank and used for the in vitro ARG. The frozen sections were allowed to reach room temperature, pre‐incubated for 15 minutes with binding buffer (50 mM Tris HCl 120 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, at pH 7.4), and then incubated for 1 hour at room temperature with the tritiated compound using different concentrations (3 and 1 nM). To determine the specific binding, adjacent sections were coincubated with the cold compound (unlabelled) at 10 μM. After incubation, the slides were washed three times in washing buffer (50 mM Tris HCl, pH 7.4) followed by a brief wash in distilled water. The slides were then dried and exposed to phosphor imaging plates (Fujifilm Plate BAS‐TR2025, Fujifilm, Tokyo, Japan). Tritium microscale standards (American Radiolabeled Chemicals Inc) were placed in cassettes together with the sections for calibration and quantification of the binding density. The phosphor imaging plates, exposed for approximately 90 hours, were scanned, and the resulting images were processed in a Fujifilm BAS‐5000 phosphor imager (Fujifilm, Tokyo, Japan). Analysis was performed using Multi Gauge 3.2 phosphor imager software (Fujifilm, Tokyo, Japan). Specific binding was calculated by subtracting the level of nonspecific binding from the total binding for each section only considering the grey matter as the region of interest (ROI) for the quantification. ROIs were drawn manually on the autoradiogram using multigauge software and were used for the semiquantitative analyses. Graphs were created with GraphPad Prism 5.0 (GraphPad Software, Inc, San Diego, CA).

CONFLICT OF INTEREST

None declared.

FUNDING INFORMATION

H2020 Marie Sklodowska‐Curie Actions, Grant/Award Number: 675417