Triazole-Based Radioligands for PET of P2X₇R: Syntheses, Conformational Studies, and Preliminary Autoradiographic Evaluation of [¹⁸F]AM-10

Abstract

The P2X₇ receptor is an emerging target for molecular imaging of inflammation in the brain and peripheral tissues. In this work, we focus on five triazole-based ligands with high affinity and selectivity for P2X₇ receptors (JNJ-64413739, JNJ-55308942, AM-10, AM-12, and AM-15), which are amenable to autoradiography and positron emission tomography (PET) imaging. We studied the phenomenon of conformational and rotational changes of these molecules by NMR and ab initio calculations. The reaction of ligands AM-10 and AM-12 with [¹⁸F]­fluoride resulted in an isotopic exchange on the pyrimidine ring, leaving the halogen atoms on the acyl moieties intact. The reaction yielded [ ¹⁸ F]­AM-10 with a radiochemical yield as high as 27% and a molar activity as high as 152 GBq/μmol. Quantitative autoradiography with [ ¹⁸ F]­AM-10 in sagittal mouse brain cryostat sections indicated a maximum specific binding (B _max) of 15.8 ± 2.8 pmol/g of wet weight and a dissociation constant (K _D) of 16.6 ± 5.1 nM. Thus, we present the first synthesis of [ ¹⁸ F]­AM-10 by isotopic exchange and confirm its specific binding at mouse brain P2X₇ receptors, which should warrant its use in animal and human PET investigations.

Introduction

P2X₇ is a ligand-gated ion channel, which has widespread expression in the brain, immunocompetent cells of the central and peripheral nervous system, and in peripheral

tissues. , Activation of P2X₇ receptors by the endogenous ligand ATP can mediate host immune responses against exogenous pathogens or endogenous factors, , with important involvement in tumor biology. , Due to its roles in diverse aspects of human physiology and pathology, the P2X₇ receptor is a rapidly emerging target for the development of pharmaceuticals , and radioligands for molecular imaging by positron emission tomography (PET) − (Figure A).

1.

Importance of the P2X₇ receptor in pathology and disease (A) and molecular structures of triazole-based P2X₇ ligands (B). Created with Biorender.com.

Tricyclic triazole derivatives show high affinity in vitro for the P2X₇ receptor. The main pharmacophore is considered to be a rigid triazolo-piperidine arrangement. One of the most affine and selective ligands for P2X₇ yet described is the antagonist ligand JNJ-55308942 (Figure B), which binds with high affinity at recombinant P2X₇ receptors from human (IC₅₀ 4.8 ± 1.1 nM) and rat (IC₅₀ 5.9 ± 1.0 nM), while showing little affinity for cytochrome P450 enzymes (IC₅₀ > 10 μM). Compound 29 (Figure B) is a structurally similar tricyclic ligand possessing higher lipophilicity, which might be a beneficial property for brain PET imaging studies. Furthermore, Chrovian et al. showed good affinity for 29 at human (IC₅₀ 10.0 ± 3.7 nM) and rat P2X₇ receptors (IC₅₀ 15.0 ± 7.4 nM) in vitro. Given these results, we hypothesized that 29 and [ ¹⁸ F]­JNJ-55308942 might both serve admirably in brain PET studies. Although there has been no report on the radiosynthesis and binding properties of [ ¹⁸ F]­JNJ-55308942, its structural congener [ ¹⁸ F]­JNJ-64413739 (Figure B) showed high-affinity binding at human (IC₅₀ 1.0 ± 0.2 nM) and rat (IC₅₀ 2.0 ± 0.6 nM) P2X₇ receptors. Furthermore, PET investigations in nonhuman primates with [ ¹⁸ F]­JNJ-64413739 showed occupancy in vivo by the P2X₇ antagonist JNJ-54175446 (Figure B), and reversible binding in the brains of healthy volunteers, with quantitation relative to an image-derived arterial input function. Other work established the dosimetry and test-retest reproducibility [ ¹⁸ F]­JNJ-64413739 for PET quantification of P2X₇ receptors in the human brain. Whereas the original SAR publication stated that the fluorine substituent in the pyrazine fragment ensured higher specificity for the P2X₇ receptor, it has been uncertain why fluorine was replaced with hydrogen during the development of the PET radiopharmaceutical [ ¹⁸ F]­JNJ-64413739. While earlier radiosynthesis employed aromatic nucleophilic substitution of the chlorine atom with [¹⁸F]­fluorine, we hypothesized that the fluorine-19 substituent in the pyrazine fragment of compounds AM-10 and AM-12 would also be amenable to ¹⁸F-labeling by an isotopic exchange mechanism.

In this work, we describe the preparation of three ¹⁸F-radiolabeled ligands of P2X₇ receptors, namely, the structurally similar PET radiopharmaceutical candidates AM-10 and AM-12 (Scheme B), which differ with respect to their lipophilicity, and the known PET radiopharmaceutical [ ¹⁸ F]­JNJ-64413739 (Figure B, Scheme A), which may represent the current gold standard for P2X₇ receptor PET investigations. For the precursors AM-10 and AM-12 (Scheme C), we tested the hypothesis that [¹⁸F]fluorine isotopic exchange in the pyrazine fragment of the precursor would lead to the labeled products [ ¹⁸ F]­AM-10 and [ ¹⁸ F]­AM-12, rather than ¹⁸F-labeling of the acyl residues by the nucleophilic substitution of the respective halogen atoms (Scheme C). We also undertook a preliminary quantitative autoradiographic study of [ ¹⁸ F]­AM-10 binding in mouse brain cryostat sections. Since there is poor understanding of the rotational and conformational behavior of triazole-based ligands, we undertook a detailed NMR study of these properties of JNJ-55308942, AM-10, AM-12, and AM-15, with confirmation of experimental observations by ab initio calculations.

1. Synthesis of Triazole-Based P2X₇ Receptor Ligand Precursors and Radioligands .

^a A – Reproduction of the synthesis of [¹⁸F]­JNJ-64413739 according to Kolb et al., 2019.; B – Synthesis of analogues of JNJ-55308942; C – Radiolabelling of AM-10 and AM-12. Reagents and conditions: (a) (COCl)₂, DCM, DMF (cat.), 17 h, RT; (b) TEA, DCM, 3 h, RT; (c) i. toluene, pyrrolidine, t = 4 h for JNJ-64410047 and t = 20 min for s3 and s4, 110 °C; ii. 0 °C, DCM, NaHCO₃, m-CPBA, NaOH (aq.), 30 min, RT; (d) K₂₂₂/K oxalate and carbonate/[¹⁸F]­F^–, DMSO, 0 min, 135 °C; (e) i. HCl in 1,4-dioxane, DCM, RT, 4 h; ii. MeOH, RT, 3 days; iii. NaHCO₃, DCM, RT, 1 h; (f) i. (hetero)­aromatic acyl chlorides were prepared by procedure (a); ii. TEA, DCM, RT, 1 h; (g) Bu₄NOMs/K⁺/[¹⁸F]­F^–, DMSO, 135 °C, 5 min or K₂₂₂/K⁺/[¹⁸F]­F^–, DMSO, 135 °C, 5 min; (h) Bu₄NOMs/K⁺/[¹⁸F]­F^–, DMSO, 135 °C, 5 min.

Results and Discussion

Chemistry and Radiochemistry

For the synthesis of [ ¹⁸ F]­JNJ-64413739 (Scheme A), we preformed the synthesis of the precursor JNJ-64410047 according to the procedure described by Chrovian et al. Kolb et al. lacked sufficient detail to enable reproduction of the synthesis. However, there is a brief analytical description and report of chemical yields in the Supplemental data of Kolb et al. In the reaction, the acylation of (S)-2-methyl-4-oxopiperidine with 3-chloro-2-(trifluoromethyl)­isonicotinoyl chloride provided a ketone (s1), which was then transformed by a 1,3-dipolar cycloaddition/Cope elimination to give the desired chromatographic standard JNJ-55308942 in 59% yield (ref 40%). The cycloaddition yielded two regioisomers but always favored the C-7 methylation (see Figure B for numbering). NMR analysis of our JNJ-64410047 product clearly showed the presence of rotamers, although HPLC analysis did not confirm the presence of the second expected regioisomer. Thus, we suppose that the second regioisomer was present in a negligible proportion. The radiotracer [ ¹⁸ F]­JNJ-64413739 was obtained after due optimization of the reaction conditions (Table ), with 2–6% decay-corrected radiochemical yield (RCY; Table ), which is (in the best-case scenario) superior to the published report of 3.1 ± 2.0% RCY. The radiochemical purity (RCP) was >99%, and the total synthesis time was 70 min.

1. Synthesis of [¹⁸F]­F-Labeled JNJ-64413739 .

				labeling				radiochemical yield
entry	initial activity 18F [GBq]	QMA for trapping 18F (form, weight)	eluent/catalyst	solvent	precursor (mg)	temperature (°C)	time (min)	(%, d.c.)
1	24	K2CO3,130 mg	8.3 mg Bu4NHCO3	0.8 mL DMSO	1.8	130	20	none
2	31			0.9 mL DMF	0.9	100	10	none
3	15		15 mg K222 + 4.8 mg KHCO3		0.9	100	10	none
4	5		8.3 mg Bu4NHCO3	0.9 mL DMSO	2.0	100	10	none
5	4	C2K2O4,130 mg	15 mg K222 + 5.8 mg C2K2O4 + 0.04 mg K2CO3		2.8	135	10	5.6
6	11	C2K2O4,46 mg	10.5 mg 18-Crown-6 + 6.4 mg C2K2O4 + 0.04 mg K2CO3	0.6 mL DMSO	0.6	135	10	none
7	11		15 mg K222 + 5.8 mg C2K2O4 + 0.04 mg K2CO3	0.6 mL DMA	1.8	135	10	0.1
8	11	C2K2O4,60 mg		0.6 mL DMSO	1.0	135	10	6.1
9	44				1.4	135	10	2.1
10	15			0.6 mL DMI	2.1	135	10	none
11	33	C2K2O4,130 mg		0.6 mL 2,3-dimethyl-2-butanol	1.3	135	10	none
12	14		15 mg K222 + 9.6 mg KH2PO4 + 0.04 mg K2CO3	0.6 mL DMSO	1.6	135	10	0.0

The approach of Chrovian et al. inspired our syntheses of the ligands JNJ-55308942, AM-10, AM-12, and AM-15 (Scheme B). However, our approach differed from that for the preparation of JNJ-64410047 with respect to our use of a Boc-protected ketone and 2-azido-5-fluoropyrimidine (s2) for the cycloaddition/Cope elimination reaction. The yield of mixed triazole products s3 and s4 was 47%, compared to the 65% reported by Chrovian et al. Kolb et al. introduced supercritical fluid chromatography (SFC) purification for regioisomer separation, whereas we used silica gel column chromatography (CHCl₃-MeOH 100:1, v/v) for partial separation of the regioisomers by fractionation of the eluent (see Supporting Information 2. Synthetic Procedures). After the first separation, we obtained pure s4 in 24% yield and another fraction containing both isomers. Chromatography of the impure fraction gave an additional 8% yield of s4 and traces of pure s3 (0.05%). NMR spectra of the separated regioisomers are shown in Supporting Information Figure S1. We performed Boc cleavage by treating s4 with HCl to give the secondary amine s5 with 49% yield (ref 53%). The final reactions were N-acylations of s5 by appropriate substituted (hetero)­aromatic carboxylic acid chlorides, closely following the procedure described above for JNJ-64410047 production. The yields of acylation giving AM-10, AM-12, and AM-15 were generally lower than the 87% for JNJ-55308942 (for structure, see Figure B).

Surprisingly, during radiolabeling of AM-10 nor AM-12 with fluorine-18, we observed no aromatic nucleophilic substitution of chlorine or bromine at the carbon-12 with [¹⁸F]fluorine (Scheme C, Table ). Therefore, we hypothesized that the reaction proceeded by fluorine-18 isotopic exchange in the pyrazine fragment of the precursor at C-20 (for numbering used, see Figure B). This model may explain the Szardening group’s decision to develop [ ¹⁸ F]­JNJ-64413739, which omits the fluorine-19 atom in the pyrazine fragment that is present in JNJ-55308942. In their reports, , the authors have not explicitly expressed any rationale for their decision to prepare an analogue without the fluorine atom in the pyrazine fragment. In the case of [ ¹⁸ F]­AM-10, the intact chlorine atom secured higher affinity toward P2X₇ receptors compared to the analogue where the chlorine is substituted with fluorine (JNJ-55308942).

2. Radio-TLC Yields [ ¹⁸ F]­AM-10 and [ ¹⁸ F]­AM-12 According to the Temperature and Time of the Reaction .

entry	compd. AM-10/12 [mg]	precondition of QMA	elution of QMA	T [°C]	time [min]	RCC [%]	RCY [%]
1	10 2.8 mg	C2K2O4	K222 (40 μmol), C2K2O4 (35 μmol), K2CO3 (0.3 μmol), 50% MeCN	135	5	61	N/A
2	10 2.8 mg	C2K2O4	K222 (40 μmol), C2K2O4 (35 μmol), K2CO3 (0.3 μmol), 50% MeCN	135	10	52	N/A
3	10 2.8 mg	C2K2O4	K222 (40 μmol), C2K2O4 (35 μmol), K2CO3 (0.3 μmol), 50% MeCN	135	10	N/A	6
4	12 2.8 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	155	5	71	N/A
5	12 2.8 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	155	10	70	N/A
6	10 2.8 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	155	5	59	N/A
7	10 2.8 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	155	10	55	N/A
8	10 2.0 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	73	N/A
9	10 2.0 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	10	65	N/A
10	10 1.0 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	56	N/A
11	10 1.0 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	10	45	N/A
12	10 0.5 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	36	N/A
13	10 0.5 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	10	35	N/A
14	10 0.1 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	8	N/A
15	10 0.1 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	10	6	N/A
16	10 0.2 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	N/A	2.5
17	10 0.2 mg	K2CO3	Bu4NOMs (20 μmol), 50% MeCN	135	5	N/A	4.7

^aThe QMAs were preconditioned with C₂K₂O₄ or K₂CO₃ (0.5 M, 10 mL).

^bAnion-exchange cartridge.

^cRadiochemical conversion.

^dRadiochemical yield.

The literature procedure for the compound JNJ-64410047 with chlorine substitution by fluorine-18 (ref ) inspired our first experiments toward labeling of AM-10. We thus prepared [ ¹⁸ F]­AM-10 by reaction of the corresponding chloro-precursor AM-10 with [¹⁸F] fluoride in the presence of potassium carbonate and Kryptofix 222 (K₂₂₂). The radiochemical conversion (RCC) was 61% after five min, and repetition under the same conditions gave a 6% RCY, which is superior to the previously published value of 3.1 ± 2.0 for ¹⁸F-JNJ-64413739 (ref ). To increase the molar activity, we applied a modified procedure, which secured higher RCYs, while using a lower amount of the precursor. The corresponding chloro-precursor AM-10 or bromo-precursor AM-12 reacted with [¹⁸F]­fluoride in the presence of Bu₄NOMs (Table ). Furthermore, in comparing the yields of the reaction at 135 and 155 °C, we found a better RCC at the lower temperature. When starting with 8.28 GBq of ¹⁸F-fluoride, the lowest tested amount of precursor AM-10 (0.2 mg) gave the highest molar activity (9.3 GBq/μmol), which is sufficient for autoradiography in vitro and small animal PET experiments and also meets the general requirements for human PET applications (Table ).

After due optimization, our most successful radiosynthesis with this precursor gave a molar activity of 152 GBq/μmol when starting with 378 GBq of ¹⁸F-fluoride (Table ). In the next step of the radiofluorination optimization, we discovered that Bu₄NOMs is prone to decomposition during prolonged storage, and that using partially decomposed Bu₄NOMs in the reaction reduces the RCY. Our original procedure was to prepare a stock solution of Bu₄NOMs in methanol and store the solution in the dark at 4 °C for up to six months. The original bottle with the solid chemical had been similarly stored for up to six years. Entries 1–12 in Table correspond to the usage of the stock solution of Bu₄NOMs. When using a new bottle of Bu₄NOMs for preparation in a fresh solution in methanol (entries 13–15 in the Table ), the optimal reaction temperature proved to be 120 °C rather than 135 °C, and the reaction time 5 min instead of 8 min. Using the fresh solution of Bu₄NOMs in methanol at 135 °C led to only 2.1% RCY. Here, we used a smaller 32 mg QMA cartridge equilibrated with potassium phosphate solution instead of the 64 mg QMA cartridge. The milder reaction conditions led to greater survival of AM-10 during the fluorine isotopic exchange reaction, leading to lower molar activity of the product (Table , entries 13 and 14). At the same time, the decay-corrected RCYs were much higher, at 24.2 and 26.9%. To obtain higher molar activity, we halved the amount of the precursor (0.1 mg instead of the previously used 0.2 mg; Table , entry 15). Indeed, this procedure gave higher molar activity (18 GBq/μmol compared to 14 GBq/μmol for the entry 14), but RCY dropped to 7.5%. Further optimization of the reaction conditions is underway.

3. Synthesis of [¹⁸F]­F-Labeled AM-10 .

				labeling
entry	initial activity 18F[GBq]	QMA for trapping 18F (form, weight)	eluent/catalyst	DMSO [mL]	AM-10 [μg]	temperature [°C]	time [min]	separation of AM-10	total time of synthesis [min]	radiochemical yield [%, d.c.]	specific activity [GBq/μmol]
1	24	K2CO3, 130 mg	1 mL 20 mM-Bu4NOMs	0.9	200	135	5	HPLC	63	1.1	24.4
2	9	C2K2O4, 130 mg	15 mg K222 + 5.8 mg C2K 2 O4 + 0.04 mg K2CO3 in 0.7 mL MeCN/H2O 1:1	0.9	200	135	5	HPLC	67	1.6	n. d.
3	12	K2CO3, 130 mg	1 mL 20 mM-Bu4NOMs	0.9	200	135	5	SPE	47	7.9	16.2
4	83	K2CO3, 130 mg	1 mL 20 mM-Bu4NOMs	0.9	200	135	5	SPE	38	3.1	n. d.
5	63	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs	0.6	200	135	10	HPLC	73	7.3	43.7
6	35	K2HPO4, 60 mg	1 mL 20 mM-Bu4NOMs	0.6	280	135	10	HPLC	79	1.6	53.7
7	63	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs	0.6	200	135	10	HPLC	79	1.8	43.7
8	143	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs+0.1 mL tert-amyl alcohol	0.6	200	135	5	HPLC	65	1.7	12.4
9	378	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs+0.1 mL tert-amyl alcohol	0.6	200	135	8	HPLC	72	1.8	152.1
10	395	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs+0.1 mL tert-amyl alcohol	0.6	200	135	8	HPLC	71	2.0	81.8
11	350	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs	0.6	200	135	8	HPLC	68	2.5	121.7
12	307	K3PO4, 60 mg	1 mL 20 mM-Bu4NOMs	0.6	200	135	8	HPLC	65	2.3	116.7
13	1.3	K3PO4, 32 mg	1 mL 20 mM-Bu4NOMs	0.6	200	120	5	HPLC	65	24.2	1.7
14	10	K3PO4, 32 mg	1 mL 20 mM-Bu4NOMs	0.6	200	120	5	HPLC	66	26.9	13.5
15	7	K3PO4, 32 mg	1 mL 20 mM-Bu4NOMs	0.6	100	120	5	HPLC	62	7.5	17.9

In Vitro Autoradiography

We tested our new tracer [ ¹⁸ F]­AM-10 in the setting of autoradiography in vitro. Displacement studies in 20-μm cryostat sections prepared from the frozen mouse brain indicated widespread specific binding (Figure ). In saturation binding studies (n = 4), we obtained maximal specific binding (B _max) for entire coronal sections of 15.8 ± 2.8 pmol/g and an apparent dissociation constant (K _D) of 16.6 ± 5.1 nM. We present a comparison of these results with literature reports for other P2X₇ ligands in the literature (Table S1). This binding affinity of [ ¹⁸ F]­AM-10 measured in mouse brain cryostat sections falls within the range of reports for other P2X₇ radioligands in various preparations, which extended from 1 nM [¹¹C]­GSK1482160 (ref ) to 25 nM [¹⁸F]­FTTM (ref ). The B _max in cryostat sections was at least five-fold lower than earlier reports for P2X₇ ligands in brain membrane preparations (see Table S1), perhaps reflecting binding site compartmentation or the effects of a chaperone lost during membrane processing.

2.

(A) In vitro autoradiography of [ ¹⁸ F]­AM-10 in mouse brain sagittal cryostat sections (20 μm-thick). (B) Representative binding curves for the P2X₇ ligand [ ¹⁸ F]­AM-10. Total binding (TB) and nonspecific binding (NSB) are shown for three different radioligand concentrations (1, 3, and 10 nM) after 60 min incubation at room temperature. Nonspecific binding was defined with the addition of 10 μM GSK1482160. Specific binding (SB) was calculated by subtracting nonspecific from total binding. (C) Scatchard plot of the saturation binding data.

NMR and Computational Study

The ¹H and ¹⁹F NMR spectra of the compounds JNJ-55308942 and AM-10, AM-12, and AM-15 at variable temperatures revealed that these compounds exist as a mixture of four rather stable conformers, which interconvert at higher temperature (Figure S2-6 and Table S1). Since a molecule’s conformation affects its pharmacological properties, we explored that phenomenon more deeply by QM calculation studies of JNJ-55308942. In accord with the spectral observations, we identified four energetically meaningful conformers (Figure ), with populations as shown in Table . The conformers have almost identical geometry on the piperidine ring but differ regarding the configuration of their N-substituent. The most stable conformers Aa and Ab carry the carbonyl group close to proton H9b (dihedral angle C9–N8–C10–C11, ExTor2 around 0°), while conformers Ba and Bb have the carbonyl close to proton H7 (ExTor2 around 180°). Proton H16 lies above the piperazine ring (i.e., on the same side as the methyl group) for conformers Aa and Ba, whereas H16 lies under the ring for conformers Ab and Bb.

3.

Populated conformers of JNJ-55308942 found by QM calculation.

4. Calculated Populations and Torsion Angles for JNJ-55308942 (See Figure ).

conformer	population [%]	N2–N1–C17–N18 [°]	C7–N8–C10–C11 [°]	N8–C10–C11–C16 [°]
JNJ-55308942-Aa	47	–19.1	13.4	–124.2
JNJ-55308942-Bb	35	–22.8	–176.6	–129.1
JNJ-55308942-Ab	9	–21.2	–1.7	113.2
JNJ-55308942-Ba	9	160.1	166.6	118.4

Overall, the DFT computational results agreed very well with the NMR data for JNJ-55308942. The additive computed populations of conformers Aa and Ab (56%) and Ba and Bb (44%) were close in magnitude to the experimentally observed values (see Table S2). Conformers Aa and Ab have torsion angles of 13.7° at C7–N8 and – 1.7° at C10–C11 (i.e., close to 0°), while conformers Ba and Bb have C10–C11 torsion angles close to 180° (166.6 and – 176.6°, respectively).

NMR gave good identification of each of the four conformers for the compounds AM-10 and AM-12, in which the energy barrier of the rotation around the C10–C11 bond was significantly higher due to the bulkiness of chlorine or bromine in the ortho position. In the cases of compounds JNJ-55308942 and AM-15 (fluorine in the ortho position), the rotation barrier was significantly lower such that most of the NMR signals of conformers Aa and Ab (as well as Ba and Bb) were around coalescence, i.e., two broad signals for Aa and Ab, or one averaged signal for both.

Due to its magnetic anisotropy, there is a strong NMR signal from the carbonyl position. The major conformers (Aa and Ab, or their average A) have a significantly higher chemical shift of H9b (5.40–5.55 vs 4.20–4.66 ppm) and a lower shift of C9 (34.6–35.1 vs 38.7–39.8 ppm). Conversely, the minor conformers (Ba and Bb, or their average B) have a higher chemical shift of H7 (5.39–5.44 versus 3.98–4.26 ppm) and a lower shift of C7 (42.1–42.7 ppm vs 48.6–48.8 ppm).

The orientation of the aryl group can be detected by observing the NOE of proton H16 in the NOESY spectrum. Based on the optimized molecular geometry, the Aa conformer would show the NOE on the methyl and proton H7, versus the protons H7 and H6b for the Ab. This specification enabled identification of the major conformers, i.e., AM-10-Aa, AM-10-Ab, AM-12-Aa, and AM-12-Ab. An analogous estimation for the minor conformers Ba and Bb failed due to low and overlapping signals.

Computation predicted four populated conformers from the computational study of compounds AM-10 and AM-15, but spectroscopy detected only two conformers in the case of compound AM-12. Following the same designation of conformers as above, we present the corresponding populations and torsion angles in Table .

5. Calculated Populations and Torsion Angles for AM-10, AM-15, and AM-12 .

conformer	population [%]	N2–N1–C17–N18 [°]	C7–N8–C10–C11 [°]	N8–C10–C11–C16 [°]
AM-10-Ab	40	160.0	5.4	92.0
AM-10-Aa	35	158.8	1.5	–97.1
AM-10-Bb	13	159.0	179.2	–107.4
AM-10-Ba	12	159.3	171.5	98.8
AM-15-Aa	36	159.4	9.2	–118.0
AM-15-Ab	32	158.6	–4.2	116.6
AM-15-Bb1	23	159.9	–176.3	–128.9
AM-15-Bb2	9	159.3	–178.5	–132.5
AM-12-Ab	60	159.4	2.6	96.1
AM-12-Aa	40	159.3	5.7	–101.9

As can be seen from the torsion angles, conformer Ba is absent for compound AM-15, and the two Bb conformers (designated as 1 and 2) are very similar, with RMS for the overlay of only 0.14 Å. For compound AM-12, we found only A conformers; based on Gibbs free energies, the B conformers were unpopulated. For the geometries of populated conformers of compounds AM-10, AM-15, and AM-12, see Supporting Information Figures S33–S35, respectively. Cartesian coordinates of the preferred conformers are also available in the Supporting Information (Table S6).

In order to evaluate rotational flexibility around the C9–N8–C10–C11 (ExTor2) and N8–C10–C11–C16 (ExTor3) torsion angles, we performed free energy relaxed scans at the DFT level. Furthermore, to better assess the effect of substituents in the close vicinity to these torsions, two additional compounds were studied computationally, namely, ethyl (Et-JNJ-55308942) and isopropyl ( ⁱ Pr-JNJ-55308942) derivatives of JNJ-55308942 at C7. Comparison of free energy maps of compounds JNJ-55308942 and AM-15 (see Figures and S36a, respectively) shows that both maps are very similar, both in the positions of local minima and with respect to the barriers at low-energy transitions between them.

4.

Relaxed scan free energy map of JNJ-55308942 (energy is given in kcal/mol).

That clearly shows that the atom in position 14, either N or C, has a negligible effect on the flexibility of both torsions. On the contrary, the halogen atom in position 12 has a very significant effect on this flexibility. Energy maps for compounds AM-10 and AM-12 (see Figure S36b,c, respectively), with Cl or Br atoms in position 12, respectively, revealed much higher energy barriers on low-energy transitions between various rotamers. Calculations indicated energies of about 13 kcal/mol for compound AM-10 compared to JNJ-55308942, and about 5 kcal/mol for compound AM-12. It is evident that a halogen atom larger than fluorine substantially hinders free rotation of vicinal torsions. Substitution of a methyl group in position 7 with an ethyl group in compound Et-JNJ-55308942 did not lead to a significant change in the energy map (see Figure S36d) compared to JNJ-55308942. We conclude that an ethyl group can rotate in a manner not to hinder free rotation around adjacent exocyclic torsions. However, similar rotation of an isopropyl group is impossible, as can be seen from the energy map of compound ⁱ Pr-JNJ-55308942 (see Figure S36e), where the energy barriers for low-energy transitions are about 5 kcal/mol higher compared to Et-JNJ-55308942. Free energy relaxed scans thus show that the halogen atom in position 12 is the structural feature having the largest effects on free rotation around the C9–N8–C10–C11 (ExTor2) and N8–C10–C11–C16 (ExTor3) torsion angles.

These considerations may call for an evaluation of the biological activity of the studied compounds (JNJ-55308942, AM-15, AM-10, and AM-12, and their derivatives) in light of the in vivo availability of their bioactive conformers. Assuming that only one conformation can access the binding site of the target receptor, it follows that the bioactivity of a given drug should depend directly on the content of the proper conformer and on the thermodynamics of its formation by isomerization.

Conclusions

We demonstrate an efficient fluorine ¹⁹F/¹⁸F isotopic exchange without aromatic nucleophilic substitution of chlorine or bromine atoms by fluorine-18 in precursor candidates of tricyclic triazole-based ligands for PET examination of the P2X₇ receptor. After due optimization, the ¹⁹F/¹⁸F isotopic exchange yielded sufficiently high molar activity for [ ¹⁸ F]­AM-10 (152 GBq/μmol, starting with 350 GBq of [¹⁸F]­fluoride) for quantitative autoradiography in vitro and, likewise, potentially for PET experiments in experimental animals and humans, without exceeding the conditions for tracer studies (usually <5% occupancy in vivo). The decay-corrected RCY was as high as 27% when lower molar activity was acceptable. Autoradiographic evaluation of [ ¹⁸ F]­AM-10 in mouse brain cryostat sections demonstrated high specific binding to brain P2X₇ receptors, with maximal binding (B _max) of 15.8 ± 2.8 pmol/g for the entire sagittal sections, and dissociation constant (K _D) 16.6 ± 5.1 nM, which is in a range of other radioligands for P2X₇ receptors. The P2X₇ ligand [¹¹C]­GSK1482160, which had low nM affinity binding in HEK cells, showed displaceable binding in the brains of living mice pretreated with lipopolysaccharide to evoke microglial activation; it remains to be established if [ ¹⁸ F]­AM-10 can likewise reveal specific binding in the living organism.

Experimental Section

Materials and Methods

For thin-layer chromatography (TLC), we used aluminum silica gel sheets for detection in UV light (TLC silica gel 60 F254, Merck). For column chromatography, we used 30–60 μm silica gel (ICN Biomedicals, Costa Mesa, USA). A JEOL 400 MHz spectrometer (Peabody, USA) served for basic NMR spectra, and a JEOL 500 MHz instrument (Peabody, USA) provided spectra as a function of temperature. The signals are represented by chemical shift (δ) in ppm, followed by multiplicity and corresponding coupling constants (J) in Hz, and by signal assignment, which is based on an analysis of ordinary ¹H–¹H COSY, ¹H–¹H NOESY, H–¹³C and ¹H–¹⁵N HSQC, and on ¹H–¹³C and ¹H–¹⁵N HMBC correlation spectra. The ¹H and ¹³C chemical shifts are referenced to TMS (using the solvent signals from CHD₂SOCD₃ (2.50 ppm) and CD₃SOCD₃ (39.52 ppm). The ¹⁵N and ¹⁹F chemical shifts are referenced (using the frequency of solvent deuterium) to liquid NH₃ and CFCl₃, respectively.

Quadrupole LC/MS (ESI ionization) with an Infinity III LC system (Agilent Technologies, Santa Clara, USA) was used for LR-MS and HPLC analyses (C18 column: 100 mm; UV detection). We used HPLC analyses (C18 column, UV and RAD detection) for the characterization of radioactive products and semipreparative HPLC (C18, UV and RAD detections) for purification. We performed radiosynthesis and purification procedures manually or on the GE TRACERlab FX FN synthesizer. For autoradiography, we used FAD mouse brains, Tissue-Tek gel (Sakura, Torrance, CA, USA), Superfrost Plus microscope slides (Hampton, USA), and a Microtome cryostat (Leica Biosystems, USA).

The following chemicals were purchased from commercial sources and used as delivered: Abcr GmbH (Karlsruhe, Germany): 3-fluoro-2-(trifluoromethyl)­isonicotinic acid (95%), 1-N-boc-(S)-2-methylpiperidine-4-one (95%), and 2-bromo-3-(trifluoromethyl)­benzoic acid (98%); Fluorochem Ltd. (Hadfield, UK): 3-chloro-2-(trifluoromethyl)­isonicotinic acid (97%); Sigma-Aldrich (Missouri, USA): ethylene glycol-bis­(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid – EGTA (≥97.0%), magnesium chloride (≥98%), m-chloroperoxybenzoic acid – m-CPBA (≥77%), oxalyl chloride – (COCl)₂ (≥99%), pyrrolidine (99%), sodium bicarbonate (99.5%), sodium sulfite (≥98%), tetrabutylammonium methanesulfonate – Bu₄NOMs (≥97%), triethylamine – TEA (99.5%), and tris­(hydroxymethyl)­aminomethane hydrochloride (≥99%); and TCI Europe (Zwijndrecht, Belgium): 2-fluoro-3-(trifluoromethyl)­benzoic acid (≥98%).

[¹⁸F]-(S)-(3-(Fluoro)-2-(trifluoromethyl)­pyridin-4-yl)­(6-methyl-1-(pyrimidin-2-yl)-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone ([¹⁸F]­JNJ-64413739)

Manual Synthesis

N.c.a. [¹⁸F]­fluoride was produced via ¹⁸O­(p,n)¹⁸F nuclear reaction by proton irradiation (18 MeV 80 μA) of enriched [¹⁸O]­H₂O. Irradiated [¹⁸O]­H₂O from the cyclotron target (TR-24 cyclotron, Nb-target, ACSI, Canada) was passed through a preconditioned QMA cartridge (10 mL, 0.5 M K₂CO₃) to trap [¹⁸F]­fluoride. Subsequently, the [¹⁸F]­fluoride was eluted with 1 mL 20 mM NBu₄OMs in MeOH (Merck, Germany), and then dried by azeotropic distillation using N₂ at 120 °C and the addition of dry MeCN (2 × 0.5 mL). For experimental details and RCYs, see Table .

Radiosynthesis Using the TRACERlab FX FN Synthesizer

Irradiated [¹⁸O]­H₂O from the cyclotron target was passed through a preconditioned QMA cartridge (10 mL 0.5 M K₂CO₃ or 0.25 M K₂C₂O₄, 20 mL water (WFI) to trap the [¹⁸F]­fluoride. Subsequently, [¹⁸F]­fluoride was eluted with 1 mL eluent (15 mg Kryptofix 222, 5.8 mg K₂C₂O₄, 0.04 mg K₂CO₃) in 50% MeCN. After elution, the [¹⁸F]­fluoride was dried by azeotropic distillation using N₂ at 120 °C and the addition of dry MeCN (2 × 0.5 mL). To the dried [¹⁸F]­fluoride was added 1–3 mg of precursor JNJ-66410047 in 0.6 mL of DMSO. The mixture was heated (135 °C, 10 min) and then diluted with 2 mL of mobile phase MeCN: 20 mM-phosphate buffer pH 5 30/70 (v/v). The reaction mixture was injected into the semipreparative HPLC, and the reactor was washed with 2 mL of the mobile phase.

The reaction mixture was analyzed on a semipreparative HPLC (Atlantis Prep T3 column, 250 × 10 mm², Waters; MeCN: 20 mM-phosphate buffer pH 5 30/70 (v/v; 8 mL/min). The radioactive peak fraction at 15–16 min was diluted with 100 mL of distilled water, and the mixture was passed through a preconditioned tC18 Plus cartridge (Waters, USA; 5 mL of EtOH, 20 mL of H₂O), eluted with 1 mL of EtOH, and diluted with 10 mL of saline. [ ¹⁸ F]­JNJ-64413739 was isolated in 2.1–6.1% radiochemical yield (RCY), > 99% radiochemical purity (RCP), and a total synthesis time of 70 min. The molar activity was not determined due to the absence of the chromatographic standard. The radiolabeled product was identified by HPLC-MS.

General Procedure (GP) for N-Acylation

To a solution of (hetero)­aromatic benzoic acid (1 equiv) in DCM (2 mL) was added (COCl)₂ (2 equiv). One drop of DMF was added, and the mixture was stirred for 2 h at RT. The solvents were removed under reduced pressure, and the residue was coevaporated with toluene (3 × 5 mL). Crude chloride salt was dissolved in DCM (2–4 mL) and added to amine s5 (0.5 equiv) in DCM (2 mL). TEA (2 equiv) was added via a syringe, and the mixture was stirred overnight at RT. The solvents were then evaporated under reduced pressure, and after aqueous workup, the residue was chromatographed (2–3% MeOH-DCM).

(S)-(3-Fluoro-2-(trifluoromethyl)­pyridin-4-yl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone (JNJ-55308942)

JNJ-55308942 (34 mg, 0.08 mmol) was prepared by the GP from s5 (47 mg, 0.2 mmol) in 40% yield. R _F = 0.3 in DCM-MeOH 20:1 (v/v). MS (ESI): for C₁₇H₁₂F₅N₇O calcd 425.10 Da, found m/z 426.1 [M + H]⁺. ¹H, ¹³C, and ¹⁹F NMR characteristics are depicted in Supporting Information Table S2 and NMR spectra are in Figures S2–S7.

(S)-(3-Chloro-2-(trifluoromethyl)­pyridin-4-yl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone (AM-10)

AM-10 (60 mg, 0.14 mmol) was prepared by GP from s5 (0.2 mmol) in 68% yield. R _F = 0.7 in AcOEt-MeOH 20:1 (v/v). MS (ESI): for C₁₇H₁₂ClF₄N₇O calcd 441.07 Da; found m/z 442.0 [M + H]⁺. ¹H, ¹³C, and ¹⁹F NMR characteristics are depicted in Supporting Information Table S3 and studied spectra are in Figures S8–S17.

(S)-(2-Bromo-3-(trifluoromethyl)­phenyl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone (AM-12)

AM-12 (40 mg, 0.08 mmol) was prepared by the GP from s5 (0.2 mmol) in 41% yield. R _F = 0.7 in AcOEt-MeOH 20:1 (v/v). MS (ESI): for C₁₈H₁₃BrF₄N₆O calcd 484.03 Da; found m/z 485.0 [M + H]⁺. ¹H, ¹³C, and ¹⁹F NMR characteristics are depicted in Supporting Information Table S4 and studied spectra are in Figures S18–S25.

(S)-(2-Fluoro-3-(trifluoromethyl)­phenyl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone (AM-15)

AM-15 (38 mg, 0.09 mmol) was prepared by GP from s5 (0.2 mmol) in 45% yield. R _F = 0.3 in DCM-MeOH 20:1 (v/v). MS (ESI): for C₁₈H₁₃F₅N₆O calcd 424.11 Da; found m/z 425.0 [M + H]⁺. ¹H, ¹³C, and ¹⁹F NMR characteristics are depicted in Supporting Information Table S5 and studied spectra are in Figures S26–S32.

[¹⁸F]-(S)-(3-Chloro)-2-(trifluoromethyl)­pyridin-4-yl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone ([¹⁸F]­AM-10)

Manual Synthesis

For the production of n.c.a. [¹⁸F]­fluoride, see the synthesis of [ ¹⁸ F]­JNJ-64413739 above. To the dried [¹⁸F]­fluoride was added AM-10 (0.2 mg, 0.45 μmol) in 0.9 mL of DMSO. The mixture was heated to 135 °C for five min, whereupon the reaction mixture was quenched with 2 mL of distilled water. The reaction mixture was analyzed on a semipreparative HPLC (Luna C18 column, 250 × 10 mm; 45% MeCN, 0.1% TFA, 4 mL/min). The fraction at 600 s was reinjected to the HPLC along with a cold reference for compound identification (Luna 5 μm C18 column, 150 × 4.6 mm; 25–95% MeCN in 7 min, 0.1% TFA, 1.5 mL/min, R _T = 5.5 min). Subsequently, the fraction was diluted with 50 mL of distilled water, passed through a preconditioned C18 cartridge (5 mL; 50% EtOH), and eluted with 3 mL of EtOH. [ ¹⁸ F]­AM-10 was isolated in 4.7% radiochemical yield (RCY) and 99.9% radiochemical purity (RCP) with a 9.26 MBq/μmol specific molar activity (SA) in a total synthesis time of 50 min.

Radiosyntheses Using the TRACERlab FX FN Synthesizer

N.c.a. [¹⁸F]­fluoride was produced via the ¹⁸O­(p,n)¹⁸F nuclear reaction, as described above. The radionuclide was eluted into the reactor. After elution with 1 mL of 20 mM Bu₄NOMs (Merck, Germany) in MeOH, [¹⁸F]­fluoride was dried as described above. To the dried [¹⁸F]­fluoride was added AM-10 (200 μg, 0.45 μmol) in 0.6 mL of DMSO. The mixture was heated to 135 °C for 8 min and then diluted with 2 mL of the mobile phase. The reaction mixture was injected into the semipreprative HPLC, and the reactor was washed with 2 mL of the mobile phase.

HPLC analysis was performed on a Kinetex EVO C18 column (100 × 21.2 mm², Phenomenex) with the eluent MeOH/20 mM-phosphate buffer pH 7.0 57/43 (v/v) delivered at 5 mL/min. The radioactive peak fraction at 13–16 min was diluted with 100 mL of distilled water, and the mixture was passed through a preconditioned tC18 Plus cartridge (Waters, USA; 5 mL of EtOH, 20 mL of H₂O) and eluted with 1 mL of EtOH. The diluted saline was eluted just before use.

[¹⁸F]-(S)-(2-Bromo-3-(trifluoromethyl)­phenyl)­(1-(5-fluoropyrimidin-2-yl)-6-methyl-1,4,6,7-tetrahydro-5H-[1,2,3]­triazolo­[4,5-c]­pyridin-5-yl)­methanone ([¹⁸F]­AM-12)

Manual synthesis. For the production of n.c.a. [¹⁸F]­fluoride, see the synthesis of [ ¹⁸ F]­JNJ-64413739 above. To the dried [¹⁸F]­fluoride was added AM-12 (2.8 mg, 5.4 μmol) in 0.9 mL of DMSO, and the mixture was heated to 135 °C. After heating for 2, 5, and 10 min, 30 μL of samples was taken and diluted with 70 μL of water, and after 18 min at RT, the reaction mixture was quenched with 2 mL of water. Along with cold reference standards for compound identification, the samples and the reaction mixture were injected into the HPLC (Kinetex 2.6 μm C18 column, 150 × 4.6 mm; 35% MeCN, 0.1% TFA, 1.5 mL/min, R _T = 4.1 min). We did not isolate [ ¹⁸ F]­AM-12; its quantitation by TLC indicated 73% radiochemical conversion (RCC) at 2 min.

Ab Initio Calculations

For computational studies, low-energy conformers for all compounds (JNJ-55308942 and AM-10, AM-12, and AM-15) were searched using the Conformational Search routine in the MacroModel module at the molecular mechanics level using the OPLS4 force field. Found structures were optimized at the density functional theory (DFT) level with the B3LYP-D3 functional and the 6–31G** basis set in implicit water using the CPCM solvent model. All of the identified local minima were verified by a frequency calculation at the same theoretical level, and Gibbs free energies were used for the estimation of Boltzmann populations of each conformer p _i according to

$$p_i=\frac{\exp (-\frac{{\epsilon }_i}{kT})}{\sum _i\exp (-\frac{{\epsilon }_i}{kT})}$$

where ε _i is the Gibbs free energy of the i-th conformer, k is the Boltzmann constant, and T is the absolute temperature. All DFT calculations were performed in the Jaguar package.

In Vitro Autoradiography

Brains harvested from male mice of C57BL/6N background were placed on ice prior to immersion in isopentane (−40 °C), followed by storage at −80 °C. A cerebral hemisphere was placed on the cryotome stand and frozen in place with tissue gel Tissue-Tek O.C.T. Compound (Sakura) for cutting at −20 °C. Serial 20 μm-thick sagittal brain sections were mounted on Superfrost Plus adhesive glass slides (Fischer) and stored at −80 °C until use. After thawing and preincubation for 10 min (50 mM Tris-HCl, pH 7.4, 5 mM MgCl₂, 2 mM EGTA, 0.1% BSA), the sections were incubated for 1 h with 0.3; 1, 3; and 10 nM [^{1 8}F]­AM-10 solutions. Nonspecific binding was assessed on consecutive sections in the additional presence of 10 μM unlabeled GSK1482160 as a blocking agent. Following the washing steps (3 × 1 min in ice-cold incubation buffer and a 30 s dip in distilled water) and drying at room temperature, the brain sections and slides containing spots of known radiochemical concentration were exposed overnight to a phosphor storage screen (Cyclone Plus (PerkinElmer) to obtain autoradiograms. After quantitation of the total nonspecific binding, Scatchard analysis of the specific binding component gave estimates of the saturation binding parameters B _max and K _D.

Ethics Statement

The animal study protocol was approved by the Ethics Committee of Samo Biomedical Centre, Pardubice, Czech Republic, for studies involving animals.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c04531.

• Reported P2X₇ ligand values of dissociation constant (K _D) and specific binding (B _max); synthetic procedures; NMR spectra, conformational behavior; and DFT data (PDF)

○.

Department of Clinical Physiology, Nuclear Medicine and PET, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen, Denmark

◆.

RadioMedic, Husinec - Řež 289, 250 68 Řež, Czech Republic

¶.

Czech Institute of Informatics, Robotics and Cybernetics, Czech Technical University in Prague, Jugoslávských partyzánů 1580/3, 160 00 Praha, Czech Republic

).

Samo Biomedical Centre, Na Klínku 1082, 530 06 Pardubice, Czech Republic

A.M.: synthesis, radiosynthesis, manuscript writing, autoradiography, data evaluation. M.J.: basic chemical design, chemical synthesis, manuscript writing. I.R.: DFT calculations. P. Drašar, P. Džubák, M.M.H., and O.L.: supervision, corrections in draft, funding. B.D.: NMR spectroscopy, data evaluation. V.S., L.P., H.V., and D.S.: radiosynthesis, radioanalysis, data evaluation, manuscript writing. P.C.: discussion, manuscript writing, experimental design (autoradiography), language corrections. A.P.: project management, supervision, manuscript writing, consultations.

The animal study protocol was approved by the Ethics Committee of Samo Biomedical Centre, Pardubice, Czech Republic, for studies involving animals.

The authors declare no competing financial interest.