Radiosynthesis of O-(1-[¹⁸F]fluoropropan-2-yl)-O-(4-nitrophenyl)methylphosphonate: A novel PET tracer surrogate of sarin

Abstract

O-(1-Fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate is a reactive organophosphate ester (OP) developed as a surrogate of the chemical warfare agent sarin that forms a similar covalent adduct at the active site serine of acetylcholinesterase. The radiolabeled O-(1-[¹⁸F]fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate ([¹⁸F] fluorosarin surrogate) has not been previously prepared. In this paper, we report the first radiosynthesis of this tracer from the reaction of bis-(4-nitrophenyl) methylphosphonate with 1-[¹⁸F]fluoro-2-propanol in the presence of DBU. The 1-[¹⁸F]fluoro-2-propanol was prepared by reaction of propylene sulfite with Kryptofix 2.2.2 and [¹⁸F] fluoride ion. The desired tracer O-(1-[¹⁸F]fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate was obtained in a >98% radiochemical purity with a 2.4% ± 0.6% yield (n = 5, 65 minutes from start of synthesis) based on starting [¹⁸F] fluoride ion and a molar activity of 49.9 GBq/μmol (1.349 ± 0.329 Ci/μmol, n = 3). This new facile radiosynthesis routinely affords sufficient quantities of [¹⁸F] fluorosarin surrogate in high radiochemical purity, which will further enable the tracer development as a novel radiolabeled OP acetylcholinesterase inhibitor for assessment of OP modes of action with PET imaging in vivo.

1 |. INTRODUCTION

Organophosphate esters (OP) are a class of compounds that include insecticides and chemical warfare agents (CWAs). CWAs such as the V-agents (eg, VX) and G-agents (eg, sarin, soman) are nondiscriminating poisons (Scheme 1) that pose a threat to both military and civilian populations.^1,2 Human toxicity resulting from OP exposures occurs via the initial inactivation of acetylcholinesterase (AChE), the enzyme responsible for the hydrolysis of the neurotransmitter acetylcholine (ACh). At high OP exposure levels, inhibition of available AChE pools in the central nervous system causes a rapid increase in synaptic ACh and triggers neurotoxic sequelae,^3–6 which may be ameliorated using antidote combinations such as 2-pyridine aldoxime methiodide (2-PAM) and atropine, if given in temporal proximity to the exposure.^7–10

SCHEME 1.

Inactivation of AChE by OP CWAs and postinhibitory pathways

The mechanism of inhibition of AChE by OP CWAs occurs through the displacement of a leaving group X (Scheme 1) by the active site serine hydroxyl of AChE.^4,11 The resultant OP-AChE adduct is relatively stable, but if the exposure is challenged early enough by certain antidotes such as 2-PAM, the OP can be nucleophilically displaced from AChE and the enzyme reactivated. Without antidote, the OP-AChE can remain intact or undergo “aging,” a process in which the second phosphonoester group is lost forming a methyl phosphonate anion attached to the serine that is fully refractory to reactivation. Some OP CWAs such as soman undergo aging within minutes, whereas sarin ages within hours although both afford the same OP adduct following loss of the ester moiety.

Previously, a novel radiolabeled OP surrogate based upon the CWA compound VX was prepared by replacing the β-aminothiol leaving group with a p-nitrophenyl (X = PNP) and the addition of a fluorine-18 atom to the ethyl ester side chain (R = CH₂CH₂F; Scheme 2) providing an OP-based PET imaging tracer.^12–15 The PNP group effectively mimics the leaving group in the CWA preserving the mechanism of inhibition forming the O-(2-fluoro-ethyl) methylphosphonate-AChE adduct thereby leaving the [¹⁸F]-OP attached to AChE. The rate of AChE inhibition by the VX and sarin PNP surrogates was slightly less than VX and sarin, but the presence of a fluorine on the alkoxy ester group reduced the rate of aging while favoring reactivation.^12,16 Thus, these surrogate compounds labeled with fluorine-18 furnish stable OP-AChE adducts that can aptly report on pools of active AChE in small animal models.¹⁴

SCHEME 2.

CWA fluoro-surrogates and their AChE adducts

Given our goal of developing tools for evaluation of exposure mechanism and reactivation therapies, we sought to prepare the [¹⁸F] fluorosarin surrogate PET tracer. The fluorine leaving group in the sarin molecule was replaced with a PNP group. An additional PNP leaving group (compound 1, Scheme 3) facilitates the incorporation of the fluorine-18-labeled isopropyl group. The fluorine-18 atom is located on the terminal methyl group of the isopropyl moiety to minimize interference with AChE reactivity. The resultant fluorosarin surrogate structure, O-(1-fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate (compound 2, Scheme 3) has been reported to possess similar inhibition, reactivation, and AChE adduct formation to sarin (without handling concerns)¹⁶ while demonstrating minimal aging. The bis PNP starting material allows for the incorporation of a fluorine-18 isopropyl group and formation of the radiolabeled AChE adduct with fluorine-18 in place of the methyl hydrogen on the sarin-AChE adduct (Scheme 2). In this study, we report the synthesis and purification of the PET tracer sarin surrogate O-(1-[¹⁸F] fluoropropan-2-yl)-O-(4-nitrophenyl)methylphosphonate.

SCHEME 3.

Synthesis of [¹⁸F]-2

2 |. EXPERIMENTAL

2.1 |. Materials

All reagents and solvents (purchased and used as anhydrous) were obtained from Sigma-Aldrich (St. Louis, MO) and used without further purification unless otherwise noted. Flash chromatography was conducted using silica gel (200–300 mesh), and thin-layer chromatography (TLC) was visualized by shortwave UV and/or staining by 2,5-dinitrobromoquinone-4-chlorimide (DBQ) and/or iodine. The nonradioactive molecular characterizations by ¹H-NMR, ¹⁹F-NMR, and ³¹P-NMR were recorded in CDCl₃ on a Varian Avance 400-MHz spectrometer, and chemical shifts (δ) are reported in ppm (relative to7.26 ppm) and coupling constants (J) in hertz (Hz). For radiochemical transformations, reversed-phase analytical HPLC was performed using a Waters 590 LC pump (Milford, MA) connected in series to a Shimadzu SPD-UV-visible detector (Columbia, MD) (λ = 254 nm) and a gamma counting in-line radiation flow detector (Model 105a, CRA; Berkeley, CA) with a Phenomenex Luna 5 μm C-18 (2) 100 Å 250 × 4.6 mm LC column. An isocratic solvent system of 40% acetonitrile/60% 5 mM pH 6.8 phosphate buffer at a flow rate of 1 mL/minute was used for analytical reversed-phase HPLC. Semipreparative reversed-phase HPLC purification of tracer was performed using a Waters 600 LC pump (Milford, MA) connected in series to a Shimadzu SPD-UV-visible detector (Columbia, MD) (λ = 254 nm) and a gamma counting in-line radiation flow detector (Model 105a, CRA; Berkeley, CA) with a Phenomenex Luna 10 μm C-18 (2) 100 Å 250 × 10 semipreparative HPLC column. HPLC chromatograms were acquired using SRI PeakSimple software (version 304—Torrance, CA). QMA light cartridges for concentrating [¹⁸F] fluoride ion and C-18 light Sep-Paks were purchased from Waters Scientific (Milford, MA).

2.2 |. Synthesis of O-(1-fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate, 2

Solid bis-(4-nitrophenyl) methylphosphonate,¹⁷ 1 (0.890 g; 2.63 mmol), was added to a conical flask (10 mL) containing 2.50 mL of anhydrous CH₂Cl₂ and stirred at 25°C. To this solution, 1,8-diazobicycloundec-7-ene (DBU; 430 μL, 2.9 mmol) was added giving a light yellow-gold solution that was stirred for 15 to 30 minutes whereupon 1-fluoro-2-propanol (220 μL, 2.9 mmol) was added via syringe, and the reaction was monitored by TLC. After 5 minutes, the solution was diluted with 50 mL CHCl₃ and extracted with 0.1 M HCl (2 × 25 mL), 1% NaOH (1 × 25 mL), water (1 × 25 mL), and brine (1 × 25 mL). The organic layer was then dried over anhydrous sodium sulfate, filtered, and concentrated to give 662 mg (91%) of 2. As an alternative to aqueous workup, the product can be isolated in 45% yield by silica gel flash chromatography (EtOAc/hexanes 1:1) of the crude oil following removal of solvent and reagents by rotary evaporation and high vacuum (<0.5 mmHg). ¹H-NMR (400.18 MHz, CDCl₃) δ 8.25 (dd, J = 17.8 Hz, 2H),7.42 (dd, J = 9.1 Hz, 2H), 4.95 to 4.82 (m, 1H), 1.76,1.74 (dd, J = 17.8 Hz), 1.38 (d, J = 6.5 Hz, 1.5H), 1.24 (d, J = 7.1 Hz, 1.5H). ¹³C-NMR (100.6 MHz, CDCl₃) δ 155.2 (d, J_C-P = 20.1 Hz), 144.6, 125.6, 121.1 (d, J _C-P = 11.0 Hz), 86.0 (d, J_C-P = 24.1 Hz), 84.3 (d, J_C-P = 25.0 Hz), 73.0 and 72.9 (d, J_C-P = 19.0, 26.2 Hz),17.0 (d, J_C-P = 29.0 Hz), 16.9 (d, J_C-P = 34.0 Hz), 13.0 (d, J_C-P = 35.0 Hz), 11.5 (d, J_C-P= 33.40 Hz). ³¹P-NMR (162 MHz, CDCl₃) δ 27.7, 29.2. ¹⁹F-NMR (376.3 MHz, CDCl₃) δ −223.9, −225.1. HRMS (ESI): Expected (M + H): 278.058815, Found: 278.059061.

2.3 |. Radiosynthesis of O-(1-[¹⁸F] fluoropropan-2-yl)-O-(4-nitrophenyl) methylphosphonate ([¹⁸F] fluorosarin surrogate), [¹⁸F]-2

[¹⁸F] Fluoride ion was produced in the UCSF GE PETtrace cyclotron by the ¹⁸O(p,n)¹⁸F reaction on enriched ¹⁸O-water. [¹⁸F] Fluoride ion was trapped on a QMA Light Sep-Pak preconditioned with 1 M Na₂CO₃ (10 mL) and H₂O (10 mL). A solution of Kryptofix[2.2.2] (18 mg) and 0.1 M K₂CO₃ (0.2 mL) dissolved in0.8 mL of CH₃CN was used to elute the [¹⁸F] fluoride ion from the QMA Sep-Pak. The solution was dried at 115°C under N₂ and repetitive azeotropic drying with CH₃CN (2 × 1 mL). A solution of 1,2-propyleneglycol sulfite (15 μL) in 300 μL CH₂Cl₂ was added, and the vial was sealed with a screw cap with a teflon septa and heated to 70°C for 20 minutes. The reaction was removed from heat and allowed to cool for 3 minutes, and a solution of 1 (8 mg, 23.6 μmol) and DBU (2 μL) in 300 μL CH₂Cl₂ was added. After 10 minutes, the reaction was dried at 80°C with a stream of N₂ under reduced pressure, dissolved in CH₃CN (300 μL), and diluted with 1.7 mL of0.1 M NaH₂PO₄. The solution was purified by semipreparative HPLC using 50% CH₃CN/50% aqueous0.1% (w/v) ascorbic acid. The fraction containing [¹⁸F]-2was diluted to 30 mL with pH 6.8 10 mM phosphate buffer, passed through a C-18 Light Sep-Pak preconditioned with EtOH (10 mL) and H₂O (10 mL). The Sep-Pak was washed with 5 mL of pH 6.8 10 mM phosphate buffer. The activity was then eluted with CH₃CN and formulated by dilution with pH 6.8 phosphate-buffered saline to 10% CH₃CN/PBS final concentration. The desired tracer identity, [¹⁸F]-2, was confirmed using the identical HPLC conditions and coelution with cold standard 2. Radiochemical purity of the final product was >98% with a yield of 2.4% ± 0.6% (n = 5, 65 minutes from start of synthesis (SOS)) based on starting [¹⁸F] fluoride ion and a molar activity of49.9 GBq/μmol (1.349 ± 0.329 Ci/μmol) at the end of synthesis (n = 3). Starting from 11.1 to 18.5 GBq (300–500 mCi) of [¹⁸F] fluoride ion, the range of final isolated yields of [¹⁸F]-2 was 0.11 to 0.37 GBq (3–10 mCi). Identity of [¹⁸F]-2 was confirmed by HPLC coinjection with 2 (data not shown).

3 |. RESULTS AND DISCUSSION

The inhibition of AChE occurs through the displacement of a leaving group on the OP by nucleophilic reaction with an active-site AChE serine hydroxyl (Scheme 1).^4,5,11

In order to follow the fate of a radiolabeled sarin, the radioisotope must be on either the methyl or the isopropoxy moieties. As previously described for the [¹⁸F]-VX surrogate,^12–15 the fluorine-18 label was placed on the alkoxy group, where the electronegativity would have minimal effect at phosphorus both in terms of the enzyme reaction and also subsequent hydrolysis of the alkoxy substituent.

Two approaches were applied to prepare the [¹⁸F]-VX surrogate. The first approach involved alkylation of a cesium phosphonate salt with 2-[¹⁸F] fluoroethyltosylate ([¹⁸F]FCH₂CH₂OTs) using microwave acceleration.¹⁴ Alternatively, transesterification of bis-(4-nitrophenyl) methylphosphonate 1 with 2-[¹⁸F] fluoroethanol in the presence of DBU provided more efficient and reproducible preparation of the [¹⁸F]-VX surrogate.¹⁵ While both of these radiosyntheses were successful, improved yields, overall synthesis time, and the efficiency of the transesterification process prompted application of this approach to prepare the sarin surrogate (Scheme 3).

Direct application of the reaction conditions used to produce the [¹⁸F]-VX surrogate^12,18 including the acetonitrile solvent resulted in low yields of [¹⁸F]-2. The ring opening of 1,2-propyleneglycol sulfite with [¹⁸F] fluoride ion in CH₃CN gave the desired 1-[¹⁸F]fluoro-2-propanol in high yield although absolute quantification of the conversion measured by radio-TLC may be underestimated due to product volatility. In spite of the suitable yield of 1-[¹⁸F]fluoro-2-propanol, the transesterification radio-chemical yields were less than 20% by analytical reversed-phase HPLC. The lower transesterification yields may be due to the greater steric demands of the 1-fluoro-2-propyloxy group as compared with the 1-fluoroethoxy group. Additionally, losses of the final product during HPLC isolation contributed to the low yield of [¹⁸F]-2 with CH₃CN. Overall, the CH₃CN route to form the [¹⁸F]-2 gave inadequate product yield for further evaluation.

As originally reported, the DBU-mediated transesterification used CH₂Cl₂ as solvent.¹⁷ In order to avoid a solvent change, CH₃CN was replaced by CH₂Cl₂ as the solvent for both labeling and transesterification. The initial [¹⁸F] fluorination reaction to the intermediate 1-[¹⁸F]fluoro-2-propanol gave radiochemical conversion comparable with the CH₃CN reaction by radio-TLC. The subsequent CH₂Cl₂ transesterification reaction gave improved yields of [¹⁸F]-2. Analytical HPLC of the crude reaction showed sufficient production of [¹⁸F]-2 (Figure 1B). In order to purify [¹⁸F]-2 by reversed-phase HPLC, it was necessary to remove the CH₂Cl₂ by evaporation and reconstitute the residue in CH₃CN and phosphate buffer. This solvent exchange process takes time and with the extra handling usually leads to loss of radioactive product. In this case, the increased transesterification yield in CH₂Cl₂ allowed isolation of [¹⁸F]-2 post HPLC in spite of the extra time and handling involved. Additionally, the necessary pre-HPLC solvent exchange enabled facile removal of unreacted 1-[¹⁸F] fluoro-2-propanol by evaporation. The semipreparative HPLC was carried out in 50/50 CH₃CN/0.1% ascorbic acid. The ascorbic acid was added to the HPLC eluent, as previously noted in Neumann et al,¹⁵ to buffer the pH <7 in order to reduce tracer decomposition during the purification step. The final yield of [¹⁸F]-2 was 2.4% ± 0.6% (SOS, 65 min) with a molar activity of49.9 GBq/μmol (1349 ± 329 Ci/mmol) and >98% radiochemical purity (Figure 1A). Thereafter, tracer was formulated in 10% CH₃CN/pH 6.8 PBS as developed for the [¹⁸F]-VX (James et al¹⁴) post HPLC to enable meaningful in vivo assessments.

FIGURE 1.

A, Analytical HPLC radiochromatogram of [¹⁸F]-2 post purification. B, Analytical HPLC radiochromatogram of crude transesterification reaction mixture

Tracer product stability was an issue. The analytical HPLC of the CH₂Cl₂ transesterification reaction (Figure 1B) shows unreacted starting material and the fluorine-18-labeled phosphonic acid analog, resulting from the loss of PNP, indicating that [¹⁸F]-2 may be hydrolyzing even in the mild room temperature transesterification reaction. It was noted that upon the addition of DBU that the reaction solution turns yellow presumably from the liberation of PNP. Additionally, exchanging the solvent from CH₂Cl₂ to CH₃CN followed by the addition 0.1 M aqueous NaH₂PO₄, to facilitate the reversed-phase semipreparative HPLC purification, may have contributed to further loss of [¹⁸F]-2. Fortu nately, the higher yields afforded by the CH₂Cl₂ transesterification reaction provided an opportunity to isolate adequate quantities of the pure [¹⁸F]-2 for subsequent evaluation.

4 |. CONCLUSIONS

A new fluorine-18-labeled surrogate of the organophosphate nerve agent sarin was prepared. Transesterification of bis-(4-nitrophenyl) methylphosphonate with DBU and 1-[¹⁸F]fluoro-2-propanol in CH₂Cl₂ gave the desired product. This approach uses mild conditions with modest radiochemical yields and high specific activity. Facile production of [¹⁸F] sarin surrogate in sufficient quantities will enable its use as a tracer for assessing OP modes of action with PET imaging in vivo.