Radiosynthesis and Analysis of (S)-4-(3-[18F]Fluoropropyl)-L-Glutamic Acid

Purpose (S)-4-(3-[18F]Fluoropropyl)-L-glutamic acid ([18F]FSPG) is an L-glutamate derivative used as a PET biomarker to assess intracellular redox status in vivo through targeting of the cystine/glutamate antiporter protein, xc− transporter. In this report, we describe a radiosynthesis of [18F]FSPG for use in PET studies that address specific challenges in relation to the radiotracer purity, molar activity, and quality control testing methods. Procedures The radiosynthesis of [18F]FSPG was performed using a customised RNPlus Research automated radiosynthesis system (Synthra GmbH, Hamburg, Germany). [18F]FSPG was labelled in the 3-fluoropropylmoiety at the 4-position of the glutamic acid backbone with fluorine-18 via substitution of nucleophilic [18F]fluoride with a protected naphthylsulfonyloxy-propyl-L-glutamate derivative. Radiochemical purity of the final product was determined by radio HPLC using a new method of direct analysis using a Hypercarb C18 column. Results The average radioactivity yield of [18F]FSPG was 4.2 GBq (range, 3.4–4.8 GBq) at the end of synthesis, starting from 16 GBq of [18F]fluoride at the end of bombardment (n = 10) in a synthesis time of 50 min. The average molar activity and radioactivity volumetric concentration at the end of synthesis were 66 GBq µmol−1 (range, 48–73 GBq µmol−1) and 343–400 MBq mL−1, respectively. Conclusion Stability tests using a 4.6 GBq dose with a radioactivity volumetric concentration of 369 MBq mL−1 at the end of synthesis showed no observable radiolysis 3 h after production. The formulated product is of high radiochemical purity (> 95%) and higher molar activity compared to previous methods and is safe to inject into mice up to 3 h after production.


3
The radiosynthesis of [ 18 F]FSPG ( Fig. 1) is typically carried out as a one-pot, two-step synthesis followed by purification and formulation for i.v. injection. In the first step, fluorine-18 incorporation into a protected precursor is achieved via nucleophilic substitution. This is followed by deprotection via acid hydrolysis to yield [ 18 F]FSPG.

3
The first reported methods (Table 1) describe the use of a precursor with a 4-nitrophenylsulphonate, leaving the group as the starting material for [ 18 F]nucleophilic displacement [21,22]. Acid hydrolysis of O-tert-butyl and N-tert-butoxycarbonyl groups then furnished [ 18 F]FSPG, with radiochemical yields (RCY) from [ 18 F]fluoride varying from 30 to 63% (decay corrected). A more recent method has employed the 2-naphthylsulphonate substituent as the leaving group and O-tert-butyl and N-triphenylmethyl substituents as the protecting groups [17]. For this method, the reported yield of [ 18 F]FSPG from [ 18 F] fluoride was 35-49% (decay corrected). While the radiolabelling and hydrolysis steps are optimised, purification of the final product has proved challenging as several methods have been applied with varied success. One of the earliest methods used a combination of reversephase HPLC and solid phase extraction to isolate [ 18 F] FSPG with 93% radiochemical purity (RCP) [15]. Later methods have switched to cartridge purification using a combination of reverse-phase and cation-exchange stationary phases [23,24], sometimes as part of chemical reagent cassette-based methods [17,25]. Radiochemical purities using this approach have been as low as 90% [17,23], although more recently values above 95% have been achieved [25,26] with molar activities (MA) in the range of 4-48 GBq µmol −1 at the end of synthesis (EOS). Here, we present a non-cassette based method for the radiosynthesis of [ 18 F]FSPG followed by HPLC purification and formulation via solid phase extraction that delivers a high purity product in high molar activity that is suitable for preclinical and clinical PET research.

Radio-HPLC Analytical Chromatography
Radio-HPLC analysis was carried out on a DionexUltimate 3000 HPLC system, consisting of an IC-3000 solvent delivery system, a WPS-3000SL-Analyt autosampler, a VWD3100 absorbance detector, a TCC3000SD column compartment, and a BioscanFC-1000 Flow countrate counter coupled to a BioscanFC-3500 PIN diode detector. The system was operated using Chromeleon software (version 6.8).

Radiosynthesis of [ 18 F]FSPG
The fully automated synthesis of [ 18 F]FSPG was carried out on a customised RNPlusSynthra module (Fig. 2)

Chemical and Radiochemical Purity (Direct Analysis)
[ 18 F]FSPG formulated in PBS (10-µL aliquot) was analysed by HPLC using a Thermo Fisher Hypercarb C 18 column (150 × 4.6 mm) eluted at 0.9 mL min −1 with a mixture of 0.1% trifluoroacetic acid and acetonitrile (94:6). The eluate was monitored continuously for radioactivity and absorbance at 218 nm. The retention time for FSPG was 4.6 min.
The formulated product gave one radioactive peak with the same retention time (4.6 min) as authentic FSPG. The radiochemical purity was 95-97%. For confirmation, [ 18 F]FSPG formulated in PBS (50 µL) was mixed with a 50-µL sample of a 100-µM reference solution of authentic FSPG. A 10-µL aliquot of the resultant mixture was then analysed by HPLC as described above.

Molar Activity Determination (Analysis Using Pre-Column Derivatisation)
[ 18 F]FSPG formulated in PBS (a 10-μL aliquot) was mixed with OPA reagent solution (10 μL) and stirred for 1 min at room temperature. The resultant sample was analysed by HPLC using a Waters Spherisorb C 18 , 3-µm column (250 × 4.6 mm) eluted at 1.0 mL min −1 with a mixture of 10-mM NaH 2 PO 4 (pH adjusted to 7.0 using 5.0-M NaOH) and methanol (60:40). The eluate was monitored continuously for radioactivity and absorbance at 338 nm. A sample of authentic FSPG (100 µM, 50 µL) when incubated at room temperature for 1 min with 50 µL of OPA reagent gave a single peak with a retention time of 4.6 min.
The derivatised 18 F-fluorinated product gave one radioactive peak and one stable peak both with the same retention time (4.6 min) as the OPA derivative of authentic FSPG. Typically, the radiochemical purity was 95-97%, and the average molar activity was 66 GBq µmol −1 (range, 44-73 GBq µmol −1 ) at the end of synthesis (EOS).

Results and Discussion
This work was undertaken to support a preclinical study of the PET imaging characteristics of the transport system x c − using [ 18 F]FSPG in a mouse tumour model. The study requires that samples of [ 18 F]FSPG be prepared on demand and transported off-site to an animal PET scanning facility prior to use. It was, therefore, necessary to have a reliable radiosynthesis method that was able to produce [ 18 F]FSPG in high chemical and radiochemical stability at room temperature. Furthermore, the method had to yield [ 18 F]FSPG in high radioactivity volumetric concentration and molar activity to provide the small volume doses needed for preclinical PET.
The substitution reaction of nucleophilic [ 18 F]fluoride with a naphthylsulfonyloxy-propyl-L-glutamate derivative, containing tert-butyl and trityl-protecting groups at the carboxylic O-hydroxy and N-amino groups, respectively, was used to synthesise [ 18 F]FSPG labelled in the fluoropropylmoiety with fluorine-18 (Fig. 1b). The radiolabelling step was performed in acetonitrile under sealed, thermal conditions. The resultant radiolabelled product was then heated under open-vessel conditions for cleavage of tert-butyl and trityl-protecting groups. While still under open-vessel conditions, the reaction mixture was heated at 100 °C with concentrated sodium hydroxide. This was to degrade or remove any potential volatile by-products of the radiolabelling and deprotection steps, such as tert-butyl alcohol (tert-butanol: boiling point = 83 °C) [27], which might hinder purification of the product. For purification, semi-preparative HPLC was used to isolate [ 18 F]FSPG, while the lipophilic3-(2-naphthylsulfonyl)oxy)propyl precursor was strongly retained on the HPLC column and was later eluted using 50% ethanol (Fig. 3). The [ 18 F]FSPG was collected in a dilute acid solution then trapped by solid phase extraction on a cation exchange cartridge, containing a sorbent with strong anionic functional groups [28]. This ion exchange trapping process requires that [ 18 F]FSPG behaves in aqueous solution as a zwitterion in the same manner as its glutamic acid parent compound [29]. Thus, at low pH, under acidic and aqueous conditions, it exists in a predominantly protonated form with a net charge of + 1 due to a positive charge at the amino group. This allows it to be trapped on the cartridge due to ionic interaction with anionic groups on the cartridge sorbent. On elution of the cartridge with PBS (pH = 7.4), the [ 18 F]FSPG is converted to a deprotonated form with a net charge of − 1 due to deprotonation at both carboxylic acid groups, while the amino group retains a positive charge. In this deprotonated form, it is released from the cartridge into the PBS solution (pH = 7.4) then filtered for use in clinical and preclinical PET imaging studies. Using this method, [ 18 F]FSPG was produced typically in 39-43% radiochemical yield (decay corrected) from [ 18 F]fluoride. After dispensing, the average radioactivity yield of [ 18 F] FSPG as formulated for preclinical use was 4.2 GBq (range, 3.4-4.8 GBq), starting from 16 GBq of [ 18 F]fluoride at the end of bombardment (EOB) (n = 10). The average molar activity was 66 GBq µmol −1 (range, 44-73 GBq µmol −1 ) at EOS, and the radioactivity volumetric concentration was 343-400 MBq mL −1 at EOS. No chemical impurities were observed in the final product at 218-nm absorbance (Fig. 4).
Values of [ 18 F]FSPG molar activity at the end of synthesis for previously reported methods (Table 1) are typically between 4 and 48 GBq µmol −1 [16,17,[24][25][26]. These methods rely on solid phase extraction (SPE) using multiple cation exchange cartridges in series for product purification and isolation. However, SPE in this way can be problematic in terms of poor recovery, reproducibility issues, and sample extracts being insufficiently clean. There is a potential for structurally similar side products to form and co-elute as part of the final formulated product. The authors of these reported methods discuss the challenges of SPE in relation to [ 18 F]FSPG purification [22,25]. Product purification is challenging because SPE does not involve real-time analysis of the recovered eluate. In the case of the synthesis of [ 18 F] FSPG, which uses harsh acidic and basic reagents, the HPLC purification method employed here offers a more accurate and reproducible way to recover products in high purity and high molar activity. This is important from a PET imaging perspective since [ 18 F]FSPG is used as a transporter radioligand for the transport system x c − [15,16]. Glutamine and cystine are the endogenous AA targets for this transport system and so [ 18 F]FSPG molar activity must be sufficiently high to avoid saturation of the transporter. Tumour cell uptake studies using L- [5-11 C]glutamine with molar activity of 1.85 GBq µmol −1 at EOS have a maximum uptake of 22% in SF188 tumour cells and in dynamic small-animal PET studies using rats, bearing 9L tumor xenografts; it shows barely measurable tumour uptake and retention [30,31]. By comparison, in the same xenograft model, [ 18 F](2S,4S)-4-(3-Fluoropropyl)glutamine shows good tumour uptake compared to the muscle (background) regions [8]. PET studies using [ 18 F]FSPG at the molar activity reported here could potentially allow improved tumour uptake characteristics.
Amino acids are typically difficult to detect by UV absorbance due to a lack of a suitable UV chromophore. Pre-column derivatisation using OPA Reagent is, therefore, used as a method for indirect HPLC analysis, whereby the amino acid analyte is converted to a UV detectable isoindoline derivative. The method is used for amino acid analysis in biological, industrial, and environmental applications [32][33][34] where a high detection sensitivity is necessary. However, the reagent is expensive, requires careful handling, has a limited shelf life, and, crucially, it may not react with product sample impurities to produce compounds identifiable by UV detection. Despite these drawbacks, the reagent is highly sensitive and reacts with trace amounts of a wide variety of compounds containing primary amino substituents. Using this indirect method, the radiochemical purity (RCP) and molar activity (MA) of the formulated radioactive product were measured at 95-97% and 44-73 GBq µmol −1 , respectively, with [ 18 F]fluoride as the only radioactive impurity (Fig. 5). Radiostability over time was tested using an 18 Fproduct dose of 4.6 GBq with a radioactivity volumetric concentration of 369 MBq mL −1 at EOS. Radioanalyses of the final product at EOS and at 1.5 and 3.0 h after EOS were performed, and, at all 3 time points, the RCP was measured at 95% with [ 18 F]fluoride as the only radioactive impurity. In between radioanalyses, the product was stored under sealed conditions at room temperature. For confirmation, the radiostability analysis was repeated using an alternative method, which allowed direct analysis of [ 18 F]FSPG. The method used a Thermo Fisher Hypercarb C 18 column, which has a sufficiently high carbon loading to retain hydrophilic FSPG and allow it to be eluted with a reasonable retention time (Fig. 4). Using this method, the identity of the final product was first confirmed by co-elution with an authentic FSPG reference prior to radiostability testing. The results of the radiostability analysis were the same as those of the indirect method, confirming that the product is suitable for use up to 3-h post-radiosynthesis. The direct analysis method is simpler but has a lower limit of detection for FSPG of ca. 50 µM and, therefore, is not sufficiently sensitive for FSPG analysis at the no-carrier-added level. This simplified procedure could be used for routine radiochemical purity and chemical purity measurements as well as radiostability testing, whereas the indirect method involving analyte derivatisation is useful for determining molar activity.

Conclusions
FSPG was labelled in the fluoropropylmoiety at the 4-position of the glutamic acid backbone with 18 F via the substitution reaction between nucleophilic [ 18 F]fluoride and di-tert-butyl-(2S,4S)-2-(3-((naphthalen-2-ylsulfonyl)oxy) propyl)-4-(tritylamino)pentanedioate. Radiolabelling was achieved by a method utilising semi-preparative reversephase HPLC purification and solid-phase extraction. This is in contrast to previously reported methods, employing cartridge purification, which do not allow real-time analysis of the purification process and are, therefore, prone to issues

3
with reproducibility and product separation efficiency. The method developed here delivers a product with improved molar activity at the end of synthesis compared to previously reported methods, which could potentially have better PET imaging characteristics with respect to imaging the amino acid transport system x c − . To simplify the product analysis, we have employed a direct method of radio-HPLC analysis, which avoids the need for pre-column derivatisation of the radiolabelled product using OPA Reagent, an expensive derivatising agent with limited chemical reactivity and a short shelf life. Stability tests using both direct and indirect methods of radioanalysis show that the formulated product is of high radiochemical purity (RCP > 95%), undergoes no observable radiolysis, has an acceptable range of molar activity (MA = 44-73 GBq µmol −1 EOS) and is suitable to inject into mice up to 3 h after production.