Development of an automated, GMP compliant FASTlab radiosynthesis of [18F]GE-179 for the clinical study of activated NMDA receptors

Khan Imtiaz (Orcid ID: 0000-0001-8070-7900)

Development of an automated, GMP compliant FASTlab^TMradiosynthesis of [¹⁸F]GE-

179 for the clinical study of activated NMDA receptors

Imtiaz Khan,^1*Tom Christian Berg,²Jane Brown,¹Rajiv Bhalla,¹Anthony Wilson,¹Andrew

Black,¹Graeme McRobbie,¹James Nairne,¹Andreas Olsson,²and William Trigg¹

1. Core Imaging R&D, Life Sciences, GE Healthcare, Pollards Wood, Nightingales Lane,

Chalfont St Giles, Bucks, HP8 4SP, UK

2. Core Imaging R&D, Life Sciences, GE Healthcare, Nycoveien 1, PO Box 4220 Nydalen,

0401 Oslo, Norway.

Abstract

N-(2-chloro-5-(S-2-[¹⁸F]fluoroethyl)thiophenyl)-N’-(3-thiomethylphenyl)-N’-

methylguanidine, ([¹⁸F]GE-179), has been identified as a promising positron emission

tomography (PET) ligand for the intra-channel phencyclidine (PCP) binding site of the N-

methyl-D-aspartate (NMDA) receptor. The radiosynthesis of [¹⁸F]GE-179 at low

radioactivity levels has been previously reported.¹However, the manufacture of a GMP

compliant product at high radioactivity levels was required for clinical studies. We describe

the development of a process using the GE FASTlab^TMradiosynthesis platform coupled with

HPLC purification. The radiosynthesis is a two-step process, involving the nucleophilic

fluorination of ethylene ditosylate, 11, followed by alkylation to the deprotonated thiol

precursor, N-(2-chloro-5-thiophenol)-N’-(3-thiomethylphenyl)-N’-methyl guanidine, 8.

The crude product was purified by semi-preparative HPLC to give the formulated product in

an activity yield (AY) of 7 ± 2% (n = 15) with a total synthesis time of 120 minutes. The

radioactive concentration (RAC) and radiochemical purity (RCP) were 328 ± 77 MBq/mL

and 96.5 ± 1% respectively and the total chemical content was 2 ± 1 µg. The final

formulation volume was 14 mL.

This article has been accepted for publication and undergone full peer review but has not

been through the copyediting, typesetting, pagination and proofreading process which may

lead to differences between this version and the Version of Record. Please cite this article as

doi: 10.1002/jlcr.3831

The previously described radiosynthesis of [¹⁸F]GE-179 was successfully modified to deliver

an process on the FASTlab^TMthat allows the manufacture of a GMP quality product from

high starting radioactivitity (up to 80 GBq) and delivers a product suitable for clinical use.

Background

N-Methyl-D-aspartate (NMDA) receptors are ligand-gated heteromeric ion channels that

regulate fast neurotransmission by allowing prolonged influx of Ca²⁺, Na⁺and K⁺ions, into

the synapse. The NMDA receptor is part of the glutaminergic excitatory pathway involved in

many neurological processes including synaptic transmission and plasticity.^{2, 3}Abnormalities

in the level of activation of NMDA receptors are known to be involved in the

pathophysiology of many diseases, (e.g. epilepsy, neuropathic pain). It is believed that an

increased level of activation is important to these disease states rather than the up-regulation

of the receptor population.^4-9A tracer that can image activated channels has potential utility

in epilepsy, TBI and other central nervous system (CNS) disorders.

Our group have previously reported the design and evaluation of [¹⁸F]GE-179,¹which has

been identified as a lead candidate to progress to clinical studies to assess its potential as a

PET tracer for imaging activated NMDA receptors.^10-14The reported precursor synthesis and

radiosynthesis were appropriate to support a biological screening program for the

identification of the lead compound, however, the manufacture of a GMP quality product was

required for clinical studies.¹⁵The original radiosynthesis¹was suitable for the preparation of

[¹⁸F]GE-179 for pre-clinical studies but was not suitable for delivering a GMP quality

product for clinical studies with high starting radioactivity (up to 80 GBq). In order to

deliver an automated process capable of delivering a GMP quality product, a number of steps

needed to be introduced or updated. These included the development of an [¹⁸F]fluoride

cartridge trapping and elution regime, and the replacement of the insoluble base with a

soluble base to deprotonate the thiol precursor, 8. Furthermore, the original radiosynthesis¹

was a 2-pot process with two semi-preparative HPLC purifications to purify both [¹⁸F]9 and

[¹⁸F]GE-179. As the FASTlab^TMonly has one reaction vessel and capacity for one semi-

preparative HPLC purification step, the original radiosynthesis was not compatible and

required modifications in order to automate the radiosynthesis of [¹⁸F]GE-179.

In this work, we describe the development of an improved, scalable precursor synthesis and

the development of a GMP compliant automated radiosynthesis of [¹⁸F]GE-179 on the

FASTlab^TMsynthesiser using a single use, disposable cassette coupled with semi-preparative

HPLC purification.

Methods

Materials and analysis

All reagents and solvents (unless stated otherwise) were purchased from Alfa Aesar

(Massachusetts, U.S), TCI Europe (Zwijndrecht, Belgium), VWR International (Lutterworth,

UK) or Sigma-Aldrich (Gillingham, UK) and used without further purification. FASTlab^TM

cassette and consumables were supplied by Clinical Supplies, GE Healthcare (Oslo, Norway).

Saline for injection (0.9% w/v) and dehydrated ethanol BP were purchased from Hameln

Pharmaceuticals (Gloucester, UK) and Martindale Pharmaceuticals (High Wycombe, UK)

respectively. QMA-carbonate-Sep Pak (p/n 186004051) and tC18 SPE (p/n WAT036810)

cartridges were purchased from Waters (Elstree, Hertfordshire, UK). Millex GS, 0.22 µm

filter unit (P/N SLG5V255F) was purchased from Sigma-Aldrich (Gillingham, UK).

HPLC conditions for radiochemistry

Semi-preparative radio-HPLC was carried out on a Knauer Smartline 1000 system equipped

with a Bioscan Flowcount FC3200 sodium iodide/PMT gamma detector (Lablogic, Sheffield,

UK) and a Knauer Smartline UV detector.

Analytical HPLC analysis was carried out on an Agilent 1100 series system equipped with a

Bioscan Flowcount FC3200 Iodide/PMT gamma detector and Laura 3 software (Lablogic,

Sheffield, UK).

Analytical conditions for synthetic chemistry

Nuclear magnetic resonance (¹H NMR and ¹³C NMR) spectra were obtained on either a

Bruker Avance DPX-400 MHz or a Varian 500 MHz NMR spectrometer. Chemical shifts are

reported in parts per million (ppm) with reference to an internal solvent. HPLC was obtained

on an Agilent 1100 series HPLC system with UV detection. The analytical HPLC column

used was an Xbridge shield Reverse Phase 18, particle size 2.5 µM with 4.6 mm id and a

length of 50 mm. Analytical thin layer chromatography (TLC) was carried out using Merck

TLC Silica gel 60 F₂₅₄aluminum plates. Flash column chromatography was carried out using

a Biotage Flash system using pre-packed silica columns or using a glass column packed with

Merck Silica gel 60 (0.040-0.063 mm).

Synthesis of thiol precursor (8) and GE-179 reference standard (10)

Synthesis of N-(2-chloro-5-thiophenol)-N’-(3-thiomethylphenyl)-N’-methyl guanidine (8)

The synthesis of 8 and unlabelled standard GE-179 described in our original publication was

somewhat challenging and not as scalable or robust as is ideal for a GMP manufacture.¹In

particular, the final debenzylation step was difficult to reproduce at multigram scales.

Therefore, an alternative strategy was developed wherein the dimeric molecule 7 was

prepared, negating the need for a protecting group that is difficult to remove. The final step

of the synthesis was a reduction to cleave the disulfide to give the free thiol. The improved

synthetic route is described in Scheme 1 and has been employed to synthesise multigram

quantities of 8 for GMP radiosynthesis.

N-(3-(Methylthio)phenyl)cyanamide (2)

3-(Methylthio)aniline (99.8 g, 0.72 mol) was dissolved in diethyl ether (600 mL) in a flask

equipped with a gas scrubber containing a sodium hypochlorite solution. The mixture was

cooled to 0-5 °C. A solution of cyanogen bromide (38.1 g, 0.36 mol) dissolved in diethyl

ether (500 mL) was added drop-wise to the reaction mixture, while maintaining the reaction

temperature below 5 °C. Upon complete addition, the reaction mixture was slowly allowed to

come to ambient temperature and stirred for a total of 18 h. A solution of hydrochloric acid

(550 mL of a 0.028% v/v solution) was added to quench the reaction. The organic phase was

washed with an additional portion of HCl (550 mL of a 0.028% v/v solution) and dried over

MgSO₄, filtered and concentrated in vacuo at 20-25 ^oC to afford the product as a brown oil

which solidified upon standing (53.0 g, 90%). ¹H NMR (400 MHz, CDCl₃) δ 7.23 (d, 1H),

6.96 (dm, 1H), 6.87 (t, 1H), 6.75 (dt, 1H), 6.01 (broad, s, 1H), and 2.48 (s, 3H). ¹³C NMR

(100 MHz, CDCl₃) δ 141.2, 137.7, 130.1 (1C), 121.7, 112.7, 111.9, 110.5, and 15.4. Purity by

HPLC (Agilent 1100 series, X-Bridge column) 98.6 % area.

N-Methyl-N’-(3-(methylthio)phenyl)cyanamide (3)

The cyanamide 2 (53 g, 0.32 mol) was dissolved in THF (400 mL) and cooled to 0 ^oC.

Sodium hydride, 14.9 g of a 60% dispersion in mineral oil, 0.37 mol) was washed with

heptane (2 x 195 mL) and suspended in THF (145 mL). The NaH suspension was added to

o

the reaction mixture portion wise while maintaining the reaction temperature below 20 C.

The resulting mixture was cooled to 0-5 °C and methyl iodide (92.3 g, 0.65 mol, 40.2 mL)

was added dropwise while maintaining the reaction temperature below 5 °C. The reaction

mixture was slowly allowed to come to ambient temperature and stirred for a total of 18.5 h.

THF was removed in vacuo at 20-25 °C to form a thick paste. The thick paste was suspended

in diethyl ether (645 mL), stirred for 45 minutes and filtered. The filtrate was evaporated in

vacuo at 20-25 °C. The crude material was purified using a Biotage flash system with a pre-

packed silica column (400 g silica), using DCM as eluent. The product fractions were

combined and concentrated in vacuo at 20-25 °C to afford the product as a pale-yellow oil

(54.1 g, 94%). ¹H NMR (400 MHz, CDCl₃) δ 7.26 (t, 1H), 6,96 (m, 1H), 6.95 (m, 1H), 6.82,

(m, 1H), 3.31 (s, 3H), and 2.48 (s, 3H). ¹³C NMR (100 MHz, CDCl₃) δ 140.9 (1C), 129.8

(1C), 121.0 (1C), 115.6 (1C), 113.8 (1C), 112.6 (1C) 111.4 (1C), 36.9 (1C), and 15.5 (1C).

Purity by HPLC (Agilent 1100 series, X-Bridge column) 98.7 % area.

5,5’-disulfanediylbis(2-chloroaniline) HCl salt (6)

Tin(II) chloride (205 g, 1.08 mol) was dissolved in concentratedhydrochloric acid (495 mL)

and added 4-chloro-3-nitrobenzenesulfonyl chloride (25.1 g, 0.01 mol) The resulting mixture

was refluxed for 4 h. The reaction mixture was then allowed to come to ambient temperature

and stirred overnight. The product was isolated by filtration and dried using suction to afford

the aminothiophenol 5 as a pale-yellow solid that was used without further purification. The

aminothiophenol 5 was suspended in water (125 mL) and agitated at ambient temperature. An

aqueous solution of potassium iodide (59.8 g, 0.36 mol) and iodine (61 g, 0.24 mol) in water

(900 mL) was added to the mixture portionwise. The reaction was monitored by HPLC

analysis. Upon complete conversion, the pH in the reaction mixture was adjusted to pH 12 by

addition of sodium hydroxide (6 M). The mixture was extracted using diethyl ether (3 x 1000

mL) and the combined organic phases were dried over MgSO₄, filtered and concentrated in

vacuo to approximately 1000 mL. HCl in diethyl ether (65 mL of a 2 M solution) was added

dropwise. The precipitated solid was collected by filtration and further washed with diethyl

ether (100 mL). The solid was dried in vacuo overnight to afford the product as an off-white

solid (16.3 g, 86%). ¹H NMR (400 MHz, d₆-DMSO) δ 7.35 (broad s, 4H), 7.23 (d, 2H), 7.05

(d, 2H), and 6.73 (dd, 2H). ¹³C NMR (100 MHz, d₆-DMSO) δ143.8, 134.9, 129.8, 117.3,

116.2, and 113.9. Purity by HPLC (Agilent 1100 series, X-Bridge column) 98.5 % area.

1,1'-(Disulfanediylbis(6-chloro-1,3-phenylene))bis(3-methyl-3-(3-

(methylthio)phenyl)guanidine) (7)

The cyanamide 3 (21.1 g, 0.12 mol) was heated to 140-145 °C in a reaction flask equipped

with a mechanical stirrer. The aniline hydrochloride 6 (11.6 g, 0.03 mol) was added in

portions to the melt.. Upon complete addition, the resulting melt was stirred for 5 h, in

process HPLC analysis confirmed that the reaction had reached completion. The reaction

melt was cooled to ambient temperature and the residue was re-dissolved in chloroform

(1055 mL) and washed with hydrochloric acid (4 x 1055 mL of a 0.1% v/v solution). The pH

of the combined aqueous phases was adjusted to pH 9 using sodium hydroxide (2M). The

aqueous phase was then back-extracted with diethyl ether (4 x 950 mL), and the combined

organic phases were dried over MgSO₄, filtered and evaporated to afford a yellow oil. The oil

was purified by flash chromatographic purification, using silica gel eluting with chloroform

added methanol (0-15%) afford the product as a foamy yellow solid (15.27 g, 78%). ¹H NMR

(400 MHz, CDCl₃) δ 7.30 (d, 2H), 7.29 (d, 2H), 7.17 (t, 2H), 7.13 (m, 2H), 7.11 (d, 2H), 7.06

(m, 2H), 7.04 (m, 2H), 3.89 (broad s, 4H), 3.34 (s, 6H), and 2.49 (s, 6H). ¹³C NMR (100

MHz, CDCl₃) δ150.7, 147.7, 144.9, 140.6, 135.9, 130.5, 129.9, 127.3, 124.5, 124.3, 124.1,

123.3, 122.6, 38.8, and 15.6. Purity by HPLC (Agilent 1100 series, X-Bridge column) 95.2 %

area.

N-(2-Chloro-5-mercaptophenyl)-N’- N’-methyl -N’-(3-thiomethylphenyl) guanidine (8)

All solvents were degassed with argon prior to use. The disulfide 7 (5.1 g, 7.6 mmol)

dissolved in ethanol (100 mL) was treated with sodium borohydride (954.7 mg, 25.2 mmol)

by portion wise addition. The reaction progress was monitored by in-process HPLC analysis.

Upon complete reaction, the reaction mixture was concentrated in vacuo at 20-25 °C to afford

a foamy pale yellow solid. The solid was purified by filtration through celite using

chloroform as eluent. The collected product fractions were acidified to pH 1 by addition of

hydrogen chloride in diethyl ether (2 M) and concentrated in vacuo at 20-25 °C to afford the

product as a foamy pale-brown solid (4.8 g, 95%). ¹H NMR (500 MHz, CDCl₃) δ 9.68 (broad

s, 1H), 8.53 (broad s, 1H), 7.13 (t, 1H), 7.03 (d, 1H), 7.02 (m, 2H), 6.92 (s, 1H), 6.89 (d, 1H),

6.86 (d, 1H), 3.74 (s, 1H), 3.66 (s, 3H), and 2.46 (s, 3H). ¹³C NMR (100 MHz, CDCl₃) δ

150.7, 147.7, 144.9 , 140.6, 135.9, 130.5, 129.9, 127.3, 124.5,124.3, 124.1, 123.3, 122.6,

38.8, and 15.6. Purity by HPLC (Agilent 1100 series, X-Bridge column) 93.3 % area.

3-(2-Chloro-5-(2-fluoroethylthio)phenyl)-N’-methyl-N’-(3-

(methylthio)phenyl)guanidine, GE-179, (10)

The route described in Scheme 2 was employed to synthesise GE-179 for analytical reference

standard or cold competition studies.

The disulfide 7 (3.0 g, 4.5 mmol) was dissolved in ethanol (120 mL) and 2-fluoroethyl

tosylate, 9, (2.1 g, 9.8 mmol) was added. The reaction mixture was heated to 60°C under an

inert atmosphere and sodium borohydride (0.8 g, 22.3 mmol) was added portionwise. The

resulting mixture was stirred for 45 minutes before being evaporated to dryness in vacuo. The

crude product was purified by flash column chromatographic purification over silica gel,

eluting with chloroform added methanol (2%) to afford the product as a clear oil that

crystalised on standing (2.2g, 5.6 mmol, 62.9%).

¹H NMR (300MHz, CDCl₃): δ 7.32 (t, 1H), 7.30 (d, 1H), 7.21 (t, 1H), 7.14 (ddd, 1H), 7.09

(ddd, 1H), 7.07 (d, 1H), 6.95 (dd, 1H), 4.53 (dt, 2H), 3.90 (broad s, 2H), 3.41 (s, 3H), 3.18

(dt, 2H), and 2.51 (s, 3H). 13C NMR (75 MHz, CDCl3) δ_C150.7, 147.7, 145.0, 140.6, 133.9,

130.4, 130.0, 126.6, 126.2, 124.7, 124.6, 124.4, 123.4, 81.6 38.8, 33.8, and 15.5. Purity by

HPLC (Agilent 1100 series, X-Bridge column) 96.1 %.

Radiosynthesis of [¹⁸F]GE-179

Preparation of reagents and FASTlab^TMcassette for the radiosynthesis of [¹⁸F]GE-179

The reagents and consumables for the radiosynthesis of [¹⁸F]GE-179 were prepared as

described in Table 1 and were contained in sealed vials or bottles in the cassette positions

according to Figure 1. The tC18 Sep-Pak cartridge was pre-conditioned with ethanol (2 mL)

followed by water (10 mL) before being placed on the FASTlab^TMcassette.

Production of [¹⁸F]Fluoride

No carrier added [¹⁸F]fluoride (1.5 mL, 40-80 GBq) was produced via an ¹⁸O(p,n)¹⁸F nuclear

reaction on a GE PETtrace cyclotron by irradiation of an enriched [¹⁸O]H₂O target using 16.4

MeV protons. The aqueous [¹⁸F]fluoride solution was transferred through a Teflon line by

helium overpressure of the target from the GE PETtrace cyclotron directly into the collection

funnel at position 6 on the FASTlab^TMcassette (Figure 1).

Quality Control

Quality control of [¹⁸F]GE-179 was carried out in accordance with the European

Pharmacopeia.¹⁶

Visual Inspection and pH

Doses were visually examined to be clear, colourless and free of particulate matter.

ColourpHast 2.0-9.0 pH-indicator strips (Fisher Scientific, Loughborough, UK) were used to

confirm the pH by comparison with the scale provided.

Radionuclidic Identity

Radionuclidic identity was confirmed by taking three activity measurements at 5 minute

intervals using the Capintec CRC-15R radioisotope dose calibrator. The half-life was

calculated from the generated calibration and compared with the known half-life of

[¹⁸F]fluorine (109.8 mins).

Bioburden, Sterility and Endotoxin Testing

Bioburden and sterility was performed by international Laboratory Services (Shardlow, UK).

Samples prior to aseptic dispensing were analysed for total yeast, mould and total aerobic

microbial count. Post aseptic sample were incubated for 14 days and tested by EP/USP

membrane filtration, absence of culture growth after 14 days was indicative of sterility.

Endotoxin content was analysed using a Charles River Laboratories Endosafe Portable

Testing System according to European Pharmacopeia guidelines.¹⁶

Aseptic Dispensing and Filter Integrity Testing

Aseptic dispensing was performed via a single-use arrangement of tubing, filter and needles

provided as a sterile (gamma-irradiated) pre-assembled kit from Pall (Portsmouth, UK).

Integrity testing was achieved by bubble point analysis performed in situ by Pall Palltronic

Flowstar II integrity test instrument.

Residual solvent and Kryptofix^®222 analysis

Residual solvent analysis was determined by using an Agilent gas chromatography (GC)

system, equipped with headspace autoinjector, split/splitless inlet, flame ionisation detector

(FID) and a Restek column (Stabilwax 30 m x 0.25 mm, 0.25 µm, G16 stationary phase).

Kryptofix^®222 levels were analysed using the established spot test method.^{17, 18}

Specification for the release of formulated [¹⁸F]GE-179

[¹⁸F]GE-179 for clinical use was analysed in accordance with European Pharmacopeia

guidelines for radiopharmaceuticals.¹⁶Table 3 summarises the specification for the release of

[¹⁸F]GE-179 for clinical use with a maximum injected dose of 10 mL.

Results and Discussion

Radiosynthesis of [¹⁸F]GE-179

The radiosynthesis of [¹⁸F]GE-179 (Scheme 3) is a two-step process involving the

nucleophilic fluorination of 11 to produce 2-[¹⁸F]fluoroethyl tosylate, [¹⁸F]9, which is a

common intermediate for the addition of a [¹⁸F]fluoroethyl group.^19-21The second step is the

alkylation of the thiolate, 12. The crude product is purified by semi-preparative HPLC

followed by reformulation on a tC18 Sep-Pak cartridge. In our original synthesis, a 2-pot-2-

HPLC purification process was employed which was both not compatible with the

FASTlab^TMand low yielding. In this work, we have developed a one-pot-one-HPLC

purification process which results in a higher yielding radiosynthesis.

Development of the automated FASTlab^TMradiosynthesis of [¹⁸F]GE-179

The radiosynthesis of [¹⁸F]GE-179 for pre-clinical studies has previously been reported.¹In

order to deliver [¹⁸F]GE-179 for clinical studies, a GMP compliant automated radiosynthesis

was required. However, the original radiosynthetic route was not compatible with

automation on FASTlab^TMfor three reasons:

1) [¹⁸F]Fluoride needs to be introduced directly from the cyclotron. Therefore, an

automated and reproducible [¹⁸F]fluoride trapping and elution process at high

radioactive levels was required

2) Cesium carbonate is not soluble in either anhydrous acetonitrile or ethanol. Therefore,

an alternative soluble base to deprotonate 8 was required

3) A 2-pot-2-semi-preparative HPLC process is not compatible with the FASTlab^TM,

due to the restricted number of cassette positions. Therefore, a one-pot-one-HPLC

purification process was required

The original radiosynthesis involved drying a mixture of [¹⁸F]fluoride in target [¹⁸O]H₂O

(200-300 µL), Kryptofix^®222 (5 mg, 13 µmol) and potassium carbonate (1 mg, 7 µmol) in

anhydrous acetonitrile (800 µL) and water (200 µL).¹Automated processes require

[¹⁸F]fluoride to be trapped on a QMA-carbonate Sep-Pak cartridge followed by elution using

a fixed volume of eluent (e.g. Kryptofix^®222 and potassium carbonate solution). This allows

for a fixed [¹⁸F]fluoride drying process. Secondly, we previously used cesium carbonate as

the base to deprotonate 8 to create the thiolate, 12. This base is not suitable for use on the

FASTlab^TMas it is not soluble in either anhydrous acetonitrile or ethanol. Furthermore, the

introduction of a second base, in addition to the QMA eluent, increased the number of

reagent vials on the cassette and limited the number of vial positions for other reagents on the

FASTlab^TMcassette. Therefore, the Kryptofix^®222 and potassium carbonate solution was

investigated as the base for the deprotonation of 8.

The QMA eluent vial for the commercial synthesis of [¹⁸F]FDG on FASTlab^TMcontains

Kryptofix^®222 (44 mg, 117 µmol), and potassium carbonate (7.8 mg, 57 µmol) in acetonitrile

(660 µL) and water (165 µL). To enable the use of this higher concentration FASTlab^TMvial,

we had to use a smaller volume (70 µL) which contained the same amount of Kryptofix^®222

and base as the original radiosynthesis. This was diluted with water (500 µL) and used to

elute the trapped [¹⁸F]fluoride from the QMA-carbonate Sep-Pak cartridge into the reaction

vessel. After the [¹⁸F]fluoride solution was evaporated to dryness, ethylene ditosylate in

anhydrous acetonitrile was added and the reaction mixture was heated at 100 °C for 10

minutes. These conditions delivered crude [¹⁸F]9 with a radiochemical conversion of > 80%

(Figure 2).

The Kryptofix^®222 and potassium carbonate solution was found to be a suitable base for the

deprotonation of 8 (Experiment 1, Table 2). Therfore, a series of experiments were performed

with different amounts of eluent from the commercially available FASTlab^TMeluent vial

(Experiments 2-6, Table 4). Analysis of the crude reaction mixtures showed that a

conversion of [¹⁸F]9 to [¹⁸F]GE-179 was >90% for the experiment using 350 µL of the

FASTlab^TMeluent and confirmed the FASTlab eluent to be a suitable base required to

deprotonate the thiol precursor 8, prior to the alkylation reaction.

Due to the position of the eluent vial (P2) and the QMA-carbonate Sep-Pak cartridge (P4) on

the FASTlab^TMcassette, the volume of eluent in the reagent vial (825 µL) was sufficient to

deliver 70 µL of eluent to elute the [¹⁸F]fluoride from the QMA-carbonate Sep-Pak cartridge,

however, due tothe position of the thiol precursor 8 vial (P16) , this volume was not sufficient

to deliver 350 µL of base to the thiol precursor vial at P16. The volume within the

FASTlab^TMmanifold between P2 and P16 is ca. 500 µL , therefore, to deliver the required

amount of base, the volume of the QMA eluent solution was increased to 1.5 mL while

maintaining the same concentration of reagents [i.e. Kryptofix^®222 (79 mg, 210 µmol),

potassium carbonate (14.2 mg, 103 µmol) in anhydrous acetonitrile (1300 µL) and water (200

µL)].

Semi-preparative HPLC purification of crude [¹⁸F]GE-179 proved to be a challenge due to

the presence of a close eluting impurity (Figure 3). It was postulated that the close eluting

peak was the acetate, 13. Formation of the acetate was thought to derive from the reaction of

acetonitrile with the hydroxyethylsulfide (Scheme 4), followed by partial hydrolysis of the

imidate.(Scheme 4). Based on this hypothesis, the formation of impurity 14 could be avoided

by replacing acetonitrile with another solvent (or removal of 11 by HPLC purification).

Therefore, a series of experiments were performed to test this hypothesis.

In the first experiment, both the synthon labelling and alkylation reactions were performed in

acetonitrile. In the second experiment, acetonitrile from the labelling reaction was removed

by evaporation prior to the alkylation and ethanol was used as the solvent for the alkylation

reaction and in the third experiment, HPLC purified [¹⁸F]9 in acetonitrile was added to 8,

which was also dissolved in acetonitrile.

The crude reaction material from each of these experiments were analysed by analytical

HPLC and the chromatograms are shown in Figures 3, 4 and 5 respectively. Figure 3 shows

the close eluting peak when using acetonitrile for the alkylation reaction, whereas Figure 4

shows the result of removing acetonitrile by evaporation and using ethanol as the alkylation

solvent. Thus, performing the alkylation reaction in ethanol prevented the formation of the

impurity, 14. Figure 5 shows that by removing unreacted 11 by HPLC purification of [¹⁸F]9

and performing the alkylation reaction in acetonitrile also prevented the formation of 14.

Therefore, the data presented in Figures 3-5 support the hypothesis that the close eluting

impurity is 14. Figure 6 shows the synthesised reference standard 14, co-elutes with the

close eluting peak 14, present in the crude reaction.

As a consequence of these results, a change to the composition of the eluent vial and removal

of acetonitrile from the [¹⁸F]9 labelling reaction was required. Therefore, the eluent vial was

prepared with ethanol instead of acetonitrile. Over 95% of the [¹⁸F]fluoride was eluted from

the QMA cartridge using the modified eluent vial and there was no impact on the

radiochemical yield of [¹⁸F]9. After formation of crude [¹⁸F]9 in acetonitrile, the reaction

was evaporated to dryness by heating at 100 °C under a nitrogen flow and also by applying a

vacuum to the reaction vessel. The subsequent alkylation reaction was performed in ethanol.

The FASTlab^TMonly has one reaction vessel and the capacity for one HPLC purification

step. Therefore, the automated process was modified so that both the labelling and alkylation

reactions were performed in the same reaction vessel. Furthermore, as a single HPLC

purification step was required for the purification of [¹⁸F]GE-179, it was necessary to remove

the the HPLC purification of [¹⁸F]9.

A consequence of using crude [¹⁸F]9 was the presence of ethylene ditosylate, 11, in the

alkylation reaction which competes with the reaction of [¹⁸F]9 with 8. When using HPLC

purified [¹⁸F]9, ethylene ditosylate, 11, would not be present in the alkylation reaction and as

a result this competing reation would not be present. This would allow the use of lower

amounts of 8, however, in the case of using crude [¹⁸F]9, it was necessary to increase the

amount 8. It was observed that when using 3 mg of 11 and 4 mg of 8, the conversion of

[¹⁸F]9 to [¹⁸F]GE-179 was ca. 40%. However, increasing the amount 8 to 10 mg increased

the conversion of [¹⁸F]9 to [¹⁸F]GE-179 to ca. 80% (Figure 7).

Automated FASTlab^TMradiosynthesis of [¹⁸F]GE-179

The incoming [¹⁸F]fluoride solution was trapped on a Waters QMA cartridge and eluted into

the reaction vessel. The eluent was prepared in syringe 1 (P3) by diluting the QMA eluent

solution with water (500 µL). The [¹⁸F]fluoride solution was then evaporated to dryness with

a combination of nitrogen flow and vacuum at 100 °C for 20 minutes.

Following the [¹⁸F]fluoride evaporation, 11 (2.5 mg, 6.7 µmol) in anhydrous acetonitrile (1

mL) was added. The radiolabelling reaction was conducted at 86 °C for 15 minutes to

produce [¹⁸F]9. Subsequently, the crude reaction mixture was evaporated to dryness with a

combination of nitrogen flow and vacuum at 80 °C for 15 minutes.

In parallel to the [¹⁸F]9 labelling reaction, the QMA eluent (350 µL) was added to the vial

containing 8 (15 mg, 44 µmol) dissolved in ethanol (1.8 mL). The resulting ethanolic thiolate

solution (1.5 mL, 10 mg, 30 µmol) was added to the reaction vessel containing dried crude

[¹⁸F]9. The alkylation reaction was carried out at 80 °C for 2 minutes followed by 100 °C for

10 minutes. Crude [¹⁸F]GE-179 was then diluted with water (2 mL) and 0.1 M hydrochloric

acid (1 mL) and loaded onto a 10 mL HPLC loop. [¹⁸F]GE-179 was purified using reverse

phase isocratic semi-preparative HPLC process on a Hichrom ACE C18 (100 x 10 mm, 5

µm) column with a flow rate of 3 mL/min with mobile phase A (0.8% Triethylamine (pH

adjusted to ca. pH 7-7.3 with 85% H₃PO₄) and mobile phase B (100% acetontrile). The

isocratic method was 60% (B) for 60 mins.

Purified [¹⁸F]GE-179 was collected into an external vial containing water (50 mL) and then

trapped on to a tC18 cartridge on the FASTlab™ cassette. After trapping, the tC18 cartridge

was washed with water (20 mL) and dried under a stream of nitrogen. Finally, [¹⁸F]GE-179

was eluted into an external formulation vial with ethanol (1 mL). The formulated product

consisted of saline (0.9% w/v, 11.5 mL) and ascorbic acid (1 mg/mL in water, 1.5 mL) to

give a final formulation in a volume of 14 mL. Quality control was performed using a

reverse phase isocratic analytical HPLC process on a Primesep B2, (250 x 4.6 mm, 5 µm)

HPLC column at a flow rate of 1.6 mL/min with mobile phase A (200 mM Ammonium

acetate) and mobile phase B (100% Acetonitrile). The isocratic method was 44% (B) for 35

mins.

Clinical production of [¹⁸F]GE-179 using the FASTlab^TMsynthesiser

The FASTlab^TMsynthesiser was developed by GE Healthcare for the automated production

of [¹⁸F]fluorine radiotracers. The synthesiser was designed to aid PET radiopharmaceutical

manufacturers to comply with the GMP requirements for PET radiopharmaceuticals.^22-33The

disposable, single-use cassette, removes the need to validate a cleaning protocol as the

disposable fluid path removes any potential for cross contamination between productions.

The radiosynthesis of [¹⁸F]GE-179 described above resulted in a process that delivered an

AY of 7 ± 2% (n = 15) with a radioactive concentration (RAC) of 289 ± 48 MBq/mL. The

total synthesis time, including purification, was 120 minutes. Quality control was carried out

in accordance with the European Pharmacopeia.¹⁶The formulated product was clear,

colourless, pH neutral and free of particulate matter. The product successfully passed the

yeast, mould count and total aerobic microbial count from the bioburden placed on the

aseptic dispensing filter. Absence of culture growth following 14 days of incubation

demonstrated the sample sterility post filtration. All samples were pyrogen free.

Analytical radio-HPLC established the radiochemical purity was 96 ± 1% and remained

greater than 95% up to at least 6 hours. The molar activity and total chemical content was

calculated to be 627 ± 416 GBq/µmol and 2 ± 1 µg respectively. The final formulation

volume was 14 mL.

Residual solvent analysis showed the levels of acetonitrile to be within the permitted daily

exposure (PDE) of < 410 ppm.³⁴Formulation of the product in 0.9% saline (11.5 mL),

ascorbic acid solution (1 mg/mL, 1.5 mL) and ethanol (1 mL) gave a formulation volume of

14 mL with an ethanol content of 7%. Ethanol content is acceptable up to a maximum of 10

% when added as an excipient.³⁴The formulated product contained less than 50 µg/mL of

16

Kryptofix^®222 by visual comparison with positive and negative controls.^15,

The

specifications for the release of [¹⁸F]GE-179 for clinical use are shown in Table 3.

Conclusion

An improved, robust and scalable precursor synthesis has been developed which enabled the

preparation of multigram quantities of the thiol precursor to GMP standard. A successful

transfer of the previously reported 2-pot-2-HPLC radiosynthesis of [¹⁸F]GE-179 to an

automated process using FASTlab^TMhas been achieved. The new process has been

implemented to GMP standard at up to 80 GBq starting activity and has enabled the delivery

of [¹⁸F]GE-179 for clinical use. The radiosynthesis delivered an AY of 7 ± 2%, with an RAC

of 289 ± 48 MBq/mL and molar activity of 627 ± 416 GBq/µmol. The RCP was 96 ± 1% and

remained over 95% up to at least 6 hours. The total chemical content was 2 ± 1 µg in a

formulation volume of 14 mL.

Abbreviations

Activity Yield; AY, COC: Cyclic Olefin Copolymer; FDG: Fluorodeoxyglucose; GMP:

Good manufacturing practice;; HPLC: High Performance Liquid Chromatography; NMDA:

N-Methyl-D-aspartate; NMR: Nuclear Magnetic Resonance; ODS: Octadecyl-silica; PET:

Position Emission Tomography; QC: Quality control; QMA: Quaternary Methylammonium;

RAC: Radioactive Concentration; RCP: Radiochemical Purity; RV: Reaction vessel; TLC:

Thin Layer Chromatography ; TEA: Triethylamine; UV: Ultraviolet

Acknowledgements

The authors would like to thank, and for their contribution and also Maria Antalffy and John

Woodcraft at Hammersmith Imanet for the delivery of the clinical preparations.

References

1.

2.

E. G. Robins, Y. Zhao, I. Khan, A. Wilson, S. K. Luthra and E. Årstad, Bioorg. Med.

Chem. Lett., 2010, 20, 1749.

R. Dingledine, K. Borges, D. Bowie and S. F. Traynelis, Pharmacological Reviews,

1999, 51, 7-61.

3.

4.

5.

6.

C. G. Lau and R. S. Zukin, Nature Reviews Neuroscience, 2007, 8, 413-426.

N. A. Cavara and M. Hollmann, Molecular Neurobiology, 2008, 38, 16-26.

E. Gamzu, Eur. Neuropsychopharm., 1995, 5, 171.

K. Muir, D. Grosset, E. Gamzu and K. Lees, British Journal of Clinical

Pharmacology, 1994, 38, 33-38.

7.

8.

P. Paoletti and J. Neyton, Current Opinion in Pharmacology, 2007, 7, 39-47.

M. R. Walters, A. P. J. Bradford, J. Fischer and K. R. Lees, British Journal of Clinical

Pharmacology, 2002, 53, 305-311.

9.

R. N. Waterhouse, Nucl. Med. Biol., 2003, 30, 869-878.

10.

F. López-Picón, A. Snellman, O. Shatillo, P. Lehtiniemi, T. J. Grönroos, P.

Marjamäki, W. Trigg, P. A. Jones, O. Solin, A. Pitkänen and M. Haaparanta-Solin, J.

Nucl. Med., 2016, 57, 1442-1447.

11.

12.

13.

C. J. McGinnity, A. Hammers, D. A. R. Barros, S. K. Luthra, P. A. Jones, W. Trigg,

C. Micallef, M. R. Symms, D. J. Brooks, M. J. Koepp and J. S. Duncan, J. Nucl.

Med., 2014, 55, 423-430.

C. J. McGinnity, M. J. Koepp, A. Hammers, D. A. Riaño Barros, R. M. Pressler, S.

Luthra, P. A. Jones, W. Trigg, C. Micallef, M. R. Symms, D. J. Brooks and J. S.

Duncan, Journal of Neurology, Neurosurgery & Psychiatry, 2015, 86, 1150-1157.

Schoenberger M, Schroeder F.A, Placzek M.S, Carter R.L, Rosen B.R, Hooker J.M,

Sander C.Y, In Vivo [¹⁸F]GE-179 Brain Signal Does Not Show NMDA-Specific

Modulation with Drug Challenges in Rodents and Nonhuman Primates ACS Chem

Neuroscience, 2018, 9(2):298-305

14.

McGinnity C.J, Årstad E, Beck K, Brooks D.J, Coles J.P, Duncan J.S, Galovic M,

Hinz R, Hirani E, Howes O.D, Jones P.A, Koepp M.J, Luo F, Riaño Barros D.A,

Singh N, Trigg W, Hammers A, Comment on " In Vivo [¹⁸F]GE-179 Brain Signal

Does Not Show NMDA-Specific Modulation with Drug Challenges in Rodents and

Nonhuman Primates, , ACS Chem Neuroscience, 2019, 10(1):768-772.

European Association of Nuclear Medicine Radiopharmacy Committee: Guidelines

on current good radiopharmacy practice (cGRPP) in the preparation of

radiopharmaceuticals, version 2., 2007.

15.

16.

European Pharmacopoeia, Council of Europe, EDQM (European Directorate for the

Quality of Medicines), 2008.

17.

18.

19.

B. H. Mock, W. Winkle and M. T. Vavrek, Nucl. Med. Biol., 24, 193-195.

P. J. H. Scott and M. R. Kilbourn, Appl. Radiat. Isot., 2007, 65, 1359-1362.

A. Bauman, M. Piel, R. Schirrmacher and F. Rösch, Tetrahedron Letters, 2003, 44,

97-99.

20.

21.

T. Kniess, M. Laube, P. Brust and J. Steinbach, MedChemComm, 2015, 6, 1714-1754.

T. R. Neal, S. Apana and M. S. Berridge, Journal of Labelled Compounds and

Radiopharmaceuticals, 2005, 48, 557-568.

22.

S. R. Dubash, N. Keat, P. Mapelli, F. Twyman, L. Carroll, K. Kozlowski, A. Al-

Nahhas, A. Saleem, M. Huiban, R. Janisch, A. Frilling, R. Sharma and E. O.

Aboagye, J. Nucl. Med., 2016, 57, 1207-1213.

23.

24.

25.

26.

R. Fortt, G. Smith, R. O. Awais, S. K. Luthra and E. O. Aboagye, Nucl. Med. Biol.,

2012, 39, 1000-1005.

J. Henrottin, C. Lemaire, D. Egrise, A. Zervosen, B. van Den Eynde, A. Plenevaux,

X. Franci, S. Goldman and A. Luxen, Nucl. Med. Biol., 2016, 43, 379-389.

Y.-Y. Huang, S. Taylor, J. Koziorowski, Y.-N. Chang, W.-H. Kao, K.-Y. Tzen and

C.-Y. Shiue, Ann. Nucl. Med., 2016, DOI: 10.1007/s12149-016-1097-4, 1-9.

A. Jackson, M. R. Battle, D. M. O'Shea, W. F. Chau, A. Gaeta, S. L. Brown, A. L.

Ewan, C. L. Jones, P. A. Jones, J. L. Woodcraft, D. R. Bouvet, B. B. Guilbert and W.

Trigg, Nucl. Med. Biol., 2014, 41, 196-202.

27.

A. Jackson, B. B. Guilbert, S. D. Plant, J. Goggi, M. R. Battle, J. L. Woodcraft, A.

Gaeta, C. L. Jones, D. R. Bouvet, P. A. Jones, D. M. O'Shea, P. H. Zheng, S. L.

Brown, A. L. Ewan and W. Trigg, Bioorganic and Medicinal Chemistry Letters,

2013, 23, 821-826.

28.

29.

R. Krasikova, Curr. Org. Chem., 2013, 17, 2097-2107.

C. Lemaire, L. Libert, X. Franci, J. L. Genon, S. Kuci, F. Giacomelli and A. Luxen,

Journal of Labelled Compounds and Radiopharmaceuticals, 2015, DOI:

10.1002/jlcr.3291, n/a-n/a.

30.

C. Lemaire, L. Libert, X. Franci, J. L. Genon, S. Kuci, F. Giacomelli and A. Luxen,

Journal of Labelled Compounds and Radiopharmaceuticals, 2015, 58, 281-290.

J. Z. Long, M. S. Jacobson and J. C. Hung, J. Nucl. Med. Technol., 2013, 41, 32-34.

D. Thonon, D. Goblet, E. Goukens, G. Kaisin, J. Paris, J. Aerts, S. Lignon, X. Franci,

R. Hustinx and A. Luxen, Mol. Imaging Biol., 2011, 13, 1088-1095.

T. Wickstrøm, A. Clarke, I. Gausemel, E. Horn, K. Jørgensen, I. Khan, D. Mantzilas,

T. Rajanayagam, D. J. In 'T Veld and W. Trigg, Journal of Labelled Compounds and

Radiopharmaceuticals, 2014, 57, 42-48.

31.

32.

33.

34.

PMP/ICH/283/95 ICH guidelines: ICH topic Q3C impurities: residual solvents

European Medicines Agency, 1998.

Table 1: FASTlab^TMreagents, components and vial composition

CASSETTE POSITION

SILICONE LINE COMPONENT

LENGTH

P1

P2

14 cm

O-18 water Recovery vial

Kryptofix^®222 (79 mg, 210 µmol) and

potassium carbonate (14.2 mg, 103 µmol)

Solvent: [water (200 µL) and ethanol (1.3

mL)]

(11 mm glass vial)

Syringe 1

P3

P4-P5

P6

14 cm

QMA-carbonate Sep-Pak cartridge

O-18 water Collection Funnel

COC Reaction vessel

COC Reaction vessel centre

Empty

P7

14 cm

P8

P9

P10

P11

P12

Empty

Syringe 2

Ethylene ditosylate, 11, (4.0 mg, 11 µmol)

in anhydrous acetonitrile (1.6 mL)

(11 mm glass vial)

P13

P14

Ethanol (4.0 mL)

(13 mm glass vial)

0.1 M HCl (4.0 mL)

(13 mm glass vial)

P15

P16

Water bag spike and water bag (100 mL)

Thiol precursor 8 (15 mg, 40 µmol) in

ethanol (1.8 mL)

(13 mm glass vial)

Line to HPLC

P17

42 cm

Table 2: Effect of the amount of FASTlab^TMeluent on the conversion of [¹⁸F]9 to [¹⁸F]GE-

179

Volume of eluent Kryptofix^®222

(µL)

K₂CO₃

8

K₂CO₃: 8

Conversion of [¹⁸F]9

to [ F]GE-179 (%)

18

1

2

3

4

5

6

50

2.7 mg, 7 µmol

8 mg, 21.2 µmol

0.5 mg, 3.6 µmol

10 mg, 30 µmol

1:8

12

70

69

89

91

90

150

350

700

1.4 mg, 10.3 µmol

10 mg, 30 µmol

1:3

18.7

49.7µmol

mg, 3.3 mg, 24.0 µmol

1:1

18.7 mg, 49.7 3.3 mg, 24.0 µmol

µmol

1:1

37.3 mg, 99 6.6 mg, 48.0 µmol

µmol

1.6:1

Table 3: Specification for the release of formulated [¹⁸F]GE-179*

Parameter

Acceptance criteria

Clear, colourless and particulate free

4-7

Appearance

pH

< 17.5 EU/mL

< 16 µg/dose

< 1.5 µg/dose

< 5 µg/dose

≤ 50 µg/mL

≤ 410 ppm

Bacterial endotoxin

Amount of [¹⁹F]GE179

Amount of any individual unknown impurity

Amount of total chemical impurities

Kryptofix® 222

Residual solvent: Acetonitrile

Residual solvent: Ethanol

RCP

≤ 10 %

≥ 95%

Pass

Sterility

≥ 3.45 bar

Filter Integrity

* Based on a maximum patient injection of 10 mL

Table 4: Summary of results from the clinical preparations of [¹⁸F]GE-179 at Hammersmith

Imanet

Starting

Activity Yield

(AY) (%)

Formulated

RAC

GE-179 only

(µg)

Total Chemical

Content (µg)

RCP (%)

Molar Activity

(GBq/µmol)

Activity (MBq)

product (MBq)

(MBq/mL)

59108

27956

57123

47862

65514

69650

61663

70765

78492

47823

63197

66495

53319

52180

66479

7.4

13.7

7.0

10.3

6.2

5.6

5.0

5.8

5.4

6.7

8.2

7.4

7.2

5.4

4385.1

3816.6

4018.6

4931.1

4038.7

4328.7

3426.3

3569.3

4573.0

2568.1

4223.6

5455.5

3966.6

3735.5

3582.4

313.2

272.6

287.0

352.2

288.5

309.2

244.7

255.0

326.6

183.4

301.7

389.7

283.3

266.8

255.9

0.9

0.6

2.0

0.7

0.6

0.5

0.7

0.2

0.5

1.1

0.6

2.7

2.2

1.7

0.6

0.9

2.1

4.6

1.8

0.6

3.2

2.3

1.0

1.8

1.4

3.1

3.5

3.4

0.6

95.3

97.8

96.1

95.6

96.5

96.7

97.2

97.1

95.7

97.9

95.3

96.1

95.8

95.9

96.1

1177.9

498.6

194.7

601.3

1409.0

308.0

293.8

788.3

1043.5

318.9

689.4

358.5

230.2

234.5

1265.1

Scheme 1: Synthesis of N-(2-chloro-5-thiophenol)-N’-(3-thiomethylphenyl)-N’-methyl

guanidine (8)

Scheme 2: Synthesis of N-(2-chloro-5-(2-fluoroethylthio)phenyl)-N’-methyl-N’-(3-

(methylthio)phenyl)guanidine, GE-179.

Scheme 3: Radiosynthesis of [¹⁸F]GE-179

Scheme 4: Postulated mechanism of the formation of 13 and 14

Figure 1: FASTlab^TMcassette for [¹⁸F]GE-179 radiosynthesis

Figure 2: Analytical HPLC chromatogram of crude [¹⁸F]9 (radiochemical in red, UV in blue)

Figure 3: Analytical HPLC chromatogram of crude [¹⁸F]GE-179 (radiochemical in red, UV

in blue) showing the close eluting impurity, 13, when the alkylation reaction was performed

using unpurified [¹⁸F]9 in acetonitrile.

Figure 4: Analytical HPLC chromatogram of crude [¹⁸F]GE-179 (radiochemical in red, UV

in blue) showing that the close eluting impurity was not present when the alkylation reaction

was performed using unpurified [¹⁸

F]9 in ethanol.

Figure 5: Analytical HPLC chromatogram of crude [¹⁸F]GE-179 (radiochemical in red, UV

in blue) when the alkylation reaction was performed using HPLC purified [¹⁸F]9 in

acetonitrile.

Article Doi

Development of an automated, GMP compliant FASTlab radiosynthesis of [18F]GE-179 for the clinical study of activated NMDA receptors

DOI: 10.1002/jlcr.3831

Source and publish data:

Authors:

Article abstract of DOI:10.1002/jlcr.3831

Full text of DOI:10.1002/jlcr.3831

Products guided by the article

R&D Labs maybe for 70502-87-3

Relevant to this article

Hot Product