An efficient preparation of labelling precursor of [11C]L-deprenyl-D2 and automated radiosynthesis

Background The synthesis of [11C]L-deprenyl-D2 for imaging of astrocytosis with positron emission tomography (PET) in neurodegenerative diseases has been previously reported. [11C]L-deprenyl-D2 radiosynthesis requires a precursor, L-nordeprenyl-D2, which has been previously synthesized from L-amphetamine as starting material with low overall yields. Here, we present an efficient synthesis of L-nordeprenyl-D2 organic precursor as free base and automated radiosynthesis of [11C]L-deprenyl-D2 for PET imaging of astrocytosis. The L-nordeprenyl-D2 precursor was synthesized from the easily commercial available and cheap reagent L-phenylalanine in five steps. Next, N-alkylation of L-nordeprenyl-D2 free base with [11C]MeOTf was optimized using the automated commercial platform GE TRACERlab® FX C Pro. Results A simple and efficient synthesis of L-nordeprenyl-D2 precursor of [11C]L-deprenyl-D2 as free base has been developed in five synthetic steps with an overall yield of 33%. The precursor as free base has been stable for 9 months stored at low temperature (−20 °C). The labelled product was obtained with 44 ± 13% (n = 12) (end of synthesis, decay corrected) radiochemical yield from [11C]MeI after 35 min synthesis time. The radiochemical purity was over 99% in all cases and specific activity was (170 ± 116) GBq/μmol. Conclusions A high-yield synthesis of [11C]L-deprenyl-D2 has been achieved with high purity and specific activity. L-nordeprenyl-D2 precursor as free amine was applicable for automated production in a commercial synthesis module for preclinical and clinical application.

pathologies has been studied with PET (Fowler et al. 2015). The PET tracer [ 11 C]L-deprenyl-D 2 binds selectively and irreversibly to the MAO-B (Fowler et al. 1987;Fowler et al. 2005).This compound acts as a suicide inhibitor of the MAO-B through a covalent linkage during normal catalytic stage, which involves cleavage of the C-D bond in the methylene carbon of the propargyl group (Fowler et al. 2015;Fowler et al. 2002). In the last years, this PET radiotracer has been applied to investigate astrocytosis in neurodegenerative diseases including Alzheimer's disease, Creutzfeldt Jakob disease and Amiotrophic Lateral Sclerosis (Engler et al. 2003;Engler et al. 2012;Choo et al. 2014;Carter et al. 2012;Santillo et al. 2011;Johansson et al. 2007). These studies indicated that [ 11 C]L-deprenyl-D 2 can be used as in vivo marker for reactive astrocytosis, providing information concerning processes leading to neuronal loss.
To facilitate the studies of [ 11 C]L-deprenyl-D 2 in humans and small animals, we have developed an efficient synthesis of the precursor for [ 11 C]L-deprenyl-D 2 as well as its radiosynthesis. Previous work describes the preparation of the labelled precursor [ 11 C]L-deprenyl-D 2 from the activated d 2 -propargyl group and L-amphetamine as starting material. L-amphetamine is extremely hard to access to, especially because only few companies market it, as well as the import requirements given by the competent national authorities take long time and are difficult to succeed. Because of this, it is convenient to develop a new synthetic strategy.
In addition, the automated syntheses provide advantages over manual or semiautomated methods. Automated syntheses generally are more reproducible than manual and semi-remote syntheses minimizing the possibility of human errors. Therefore, an efficient alternative to the synthesis of L-nordeprenyl-D 2 precursor of [ 11 C]L-deprenyl-D 2 as free base and an improved automated synthetic method have been developed. This paper describes both aspects of the improved synthesis of [ 11 C]L-deprenyl-D 2 .

Organic synthesis
All chemicals and reagents were purchased from Aldrich, Merck and Dorwil. Analytical TLC were performed on silica gel 60F-254 plates and visualized with UV light (254 nm) and p-anisaldehyde in acidic ethanolic solution or iodine vapours. Column chromatography was performed using silica gel (SAI,. NMR spectra were recorded on a Bruker DPX-400 spectrometer. The assignment of chemical shifts was based on standard NMR experiments ( 1 H, 1 H-COSY, HETCOR and 13 C-NMR). The chemical shifts values were expressed in ppm relative to tetramethylsilane as internal standard. Mass spectra were determined on a Shimadzu DI-2010 (EI-MS) or Applied Biosystem API 2000 (ESI-MS). IR were obtained using a Shimadzu IR equipment Affinity-1 (Fourier Transform Infrared Spectrophotometer). Materials, instruments, protocols and documents used for precursor synthesis were in agreement with GMP recommendations.
The ice/water bath was removed and the mixture stirred at room temperature for 20 min. Then, the mixture was again cooled to 0°C and L-phenylalanine (1 g, 6 mmol) was added. The ice/water bath was removed, and the reaction mixture was stirred at room temperature for 12 h. The reaction mixture was cooled to 0°C, and methanol (9 mL) was added dropwise, followed by aqueous sodium hydroxide (5 mL, 2.5 M). Finally, the mixture was evaporated in vacuum, and the residue extracted with chloroform (5 × 5 mL). The combined extracts were dried with Na 2 SO 4 , filtered, and evaporated in vacuum. The white solid obtained was dried under vacuum for 24 h to yield 1 (0.84 g, 92% yield).
i-b) To a solution of L-Phenylalanine methyl ester hydrochloride (300 mg, 1.39 mmol) in a 1:1 (v/v) mixture of water and ethanol (3.5 mL) was added slowly with stirring a solution of lithium borohydride (103 mg, 4.73 mmol) in the same solvent (3.5 mL) cooled externally in an ice/water bath. When the addition of borohydride was complete the mixture was stirred for 1 h at room temperature. Next, the solution was evaporated under reduced pressure and the residual aqueous solution treated first with sodium hydroxide and then with sodium chloride to saturate the solution before extraction with ethyl acetate (5 × 5 mL). The extract was washed with brine, dried over anhydrous Na 2 SO 4 , and evaporated under reduced pressure to yield 1 as white solid (0.172 g, 82% yield). 1  (S)-tert-Butyl (1-hydroxymethyl-2-phenylethyl)-carbamate (2): To a magnetically stirred suspension of 1 (1.0 g, 6.6 mmol) in water (6.5 mL) was added di-tert-butyl dicarbonate ((Boc) 2 O, 1.5 g, 9.9 mmol) at room temperature. After stirring for 25 min the reaction mixture, the white solid formed was filtered, washed with water and dried under vacuum for 48 h to yield 2 (1.31 g, 79% yield). (S)-tert-Butyl (1-iodomethyl-2-phenylethyl)-carbamate (3): A mixture of iodine (1.59 g, 6.28 mmol), imidazole (0.47 g, 6.9 mmol) and triphenylphosphine (1.65 g, 6.26 mmol) in dry dichloromethane (50 mL) was cooled at 0°C with stirring for 15 min. Next, the mixture was stirred at room temperature for another 15 min, and a solution of 2 (1.44 g, 5.71 mmol) in dry dichloromethane (18 mL) was added dropwise. The mixture was stirred for 15 min at room temperature; the solid formed was filtered and the organic layer washed with diluted aqueous Na 2 S 2 O 3 and water, dried with Na 2 SO 4 and evaporated in vacuo. After the workup, the crude was purified by column chromatography (SiO 2 , Hexane/EtOAc (9:1)), yielding derivative 3 as a white solid (1.6 g, 80%). (S)-tert-Butyl (1-methyl-2-phenylethyl)-carbamate (4): A mixture of 3 (1.53 g, 4.24 mmol) in anhydrous tetrahydrofuran (32 mL) was cooled to −10°C under nitrogen atmosphere. Next, a solution of sodium tri-sec-butylborohydride (N-Selectride) 1 M in tetrahydrofuran (6.36 mL, 6.36 mmol) was added dropwise and the resulting mixture was stirred at 0-5°C for about 2 h. The reaction was quenched by the slow addition of water (3.0 mL) followed by the dropwise addition of a solution made by combining 45 mL of H 2 O, 3.0 g of K 2 CO 3 , and 23 mL of 10% H 2 O 2 . The reaction mixture was stirred at room temperature for 1 h. The THF was evaporated under reduced pressure, and the product was extracted with dichloromethane (4 × 15 mL). The organic layers were dried with Na 2 SO 4 and the solvent evaporated in vacuo. After the workup the crude was purified by column chromatography (SiO2, Hexane/EtOAc (9:1)), yielding derivative 4 as a white solid (0.92 g, 93%). (1,1-d 2 )-2-propyn-1-ol (5): A 1 M solution of LiAlD 4 (29.0 mL, 29.0 mmol) in ether was cooled to −55°C under nitrogen atmosphere in a two neck round-bottomed flask. Next, a solution of methyl propiolate (2.7 mL, 30 mmol) in anhydrous ether (10 mL) was added dropwise, over a period of about 60 min. The reaction mixture was stirred for another 90 min at −30°C and was then allowed to warm to room temperature over a period of about 3 h and stirred overnight. Finally, the mixture was cooled to about 0°C and quenched by the slow addition of water (1.5 mL) followed by the dropwise addition of a solution of NaOH (0.11 g in 0.75 mL) and 1 mL of H 2 O. The solid was allowed to settle and decanted. The solid formed was filtered, washed with ether (2 × 25 mL), the organic layers dried with Na 2 SO 4 and the ether was evaporated under vacuum. d 2 -Propargyl alcohol was obtained as an oil (∼50% by 1H NMR signals) and was used in the next reaction without further purification. 1 H-NMR (CDCl 3 ) δ (ppm): 3.4 (s, 1H, OH), 2.4 (s, 1H, CH). 13 C-NMR (CDCl 3 ) δ (ppm): 60.4, 73.7, 81.0.
(1,1-d 2 )Propargyl p-toluenesulphonate (6): A mixture of 5 (crude mixture of the reduction process) and p-toluenesulfonyl chloride (5.8 g, 30 mmol) in anhydrous ether (70 mL) was cooled a − 10°C under nitrogen atmosphere. Next, KOH (8.50 g, 152 mmol) was added and the mixture was allowed to warm to room temperature, over a period of about 1 h, and then stirred for 2 h. The solid decanted was filtered, washed with ether (20 mL) and the organic layer washed with brine, dried with Na 2 SO 4 and evaporated in vacuo. After the workup the crude was purified by column chromatography (SiO 2 , Hexane/EtOAc (9:1)), yielding derivative 6 as a yellow oil (2.45 g, 40% two steps). L-nordeprenyl-D 2 : To a solution of 4 (117 mg; 0.50 mmol) in dichloromethane (1.0 mL) was added trifluoroacetic acid (0.25 mL) and stirred at room temperature for 2 h. The volatile components were removed under reduced pressure. Then, anhydrous DMF (5 mL), potassium carbonate (138 mg, 1.0 mmol) and d 2 -propargyl tosylate 6 (110 mg, 0.5 mmol) were added at room temperature. The resulting mixture was stirred at ambient temperature for about 24 h. The mixture was then diluted with water (20 mL) and extracted with diethyl ether (3 × 10 mL). The organic layers were combined, washed with brine, dried, and concentrated in vacuo. The resulting residue was then purified by flash column chromatography (hexane/ ethyl acetate (7:3)) to give the desired product (53 mg; 61% . Once the radioactivity in the reactor reached a plateau, solution was heated to 80°C for 1 min. Crude [ 11 C]L-deprenyl-D 2 was separated from its precursor, the solvent and other minor radiochemical impurities using semipreparative reverse-phase HPLC (Nucleosil C18ec, 250 × 10, Macherey-Nagel; CH 3 COONH 4 0.1 M:MeCN 40:60, flow rate 6 mL/min, UV and gamma detection). The fraction containing the [ 11 C]L-deprenyl-D 2 was diluted in water (50 mL) for injection, passed through a SPE cartridge (Sep-pak C18 light), and eluted with EtOH (1 mL). [ 11 C]L-deprenyl-D 2 was formulated with saline (9 mL) and subjected to sterilizing filtration (0.22 μ). Chemical and radiochemical impurities were detected and quantified using radio-HPLC: a mixture of TFA 0.1% and acetonitrile (75:25; v/v) was used as the mobile phase at a flow rate of 1.5 mL/min on a Nucleodur C18-ec 100-5 250 × 4.6 column (Macherey-Nagel). The whole HPLC analysis was completed within 10 min. The retention times of the L-nordeprenyl-D 2 and L-deprenyl-D 2 4.4 ± 0.3 min and 5.4 ± 0.3 min, respectively. The chemical identity of [ 11 C]L-deprenyl-D 2 was determined by comparing the retention time of the unlabelled reference compound. The radiochemical purity was calculated considering the portion of [ 11 C]L-deprenyl-D 2 in relation to total radioactivity. The specific activity was determined considering total radiopharmaceutical activity and the amount of the unlabelled product.
The residual solvents (such as acetone, MEK and acetonitrile) and ethanol were analysed by gas chromatography (GC) in accordance with USP general chapter <467>. The appearance of the solution was checked by visual inspection, and pH was determined using a calibrated pH-meter. Radionuclidic purity was assessed by recording the corresponding gamma spectrum and radionuclidic identity by measuring the physical half-life.
Sterility and concentration of bacterial endotoxins were tested in accordance with USP general chapters <71>and <85>, respectively.

Results and discussion
Organic synthesis of L-nordeprenyl-D 2 The synthesis of L-nordeprenyl-D 2 was initially reported through direct N-alkylation reaction between L-amphetamine and propargyl bromide-α-α-D 2 (Scheme 1) Fowler et al. 1988). L-amphetamine was purchased commercially, while deuterated propargyl bromide was prepared by the reduction of methyl propiolate with LiAID 4 followed by bromination with PBr 3 . Under this condition the deuterated key compound was obtained in low yield as a mixture difficult to purify, containing 15% of allyl bromide-1,1,3-D 3 . Another drawback of applying this methodology is the difficult access to L-amphetamine for use in research, as mentioned above.
To avoid these problems, we have developed a new route to synthesizing Lnordeprenyl-D 2 from L-phenylalanine in five steps (Scheme 2).
Using this methodology, the key precursors to obtain are the derivative of Lamphetamine protected with Boc group, compound 4 (Scheme 2) and propargyl tosylate deuterated 6 (Scheme 3). First, to synthesize the derivative Boc-L-amphetamine 4, a previously described synthetic sequence was adopted with some improvements in certain reaction steps (Quagliato et al. 2000;Gant and Sarshar 2010). At the beginning, Lphenylalanine as starting material was reduced in the presence of TMS-Cl and LiBH 4 , activating and reducing agent, respectively, yielding L-phenylalanilol 1 in excellent yield (92%, condition i-a, Scheme 2). In this context, when L-phenylalanine methyl ester was used as starting material and LiBH 4 as a reducing agent, L-amino alcohol 1 was obtained in 82% yield and short reaction time (condition i-b, Scheme 2) (Hvidt et al. 1988). Subsequently the amino group of compound 1 was converted to N-t-Boc derivative by reaction with (Boc) 2 O in aqueous medium under mild conditions. The procedure was carried out using in short reaction times, and the L-Boc-phenylalanilol 2 was isolated by simple filtration in high yield (Scheme 2). Then, alcohol 2 was transformed into the iodomethyl 3 in presence of about 1 equivalent of triphenylphosphine-iodine-imidazole system under mild reaction conditions (15 min at room temperature). Subsequent reduction of iodomethyl derivative 3 using N-Selectride as reducing agent leads to the formation of L-Bocamphetamine 4 in excellent yield. This last key intermediate was obtained with 54% Scheme 1 Conventional organic synthesis of L-nordeprenyl-D 2 using L-amphetamine as starting material Scheme 2 i-a) LiBH 4 , TMS-Cl, THF, r.t., 12 h, 92%; i-b) LiBH 4 , EtOH-H 2 O, r.t., 1 h, 82%; ii) (Boc) 2 O, H 2 O, 79%; iii) Triphenylphosphine, iodine, imidazole, CH 2 Cl 2 , 80%; iv) N-selectride, THF, 93%; v) a) TFA, CH 2 Cl 2 ; b) 6, K 2 CO 3 , DMF, 61% two steps overall yield following the synthetic methodology developed in this work. Considering that the next step requires the use of a propargyl deuterated derivative activated for N-alkylation reaction, we aimed obtaining the tosylate 6, since this derivative could be easily isolated and purified by column chromatography. Thus, through a first reduction step of methyl propiolate with LiAlD 4 the corresponding d2-propargyl alcohol 5 (Scheme 3) was obtained. The d2-propargyl tosylate 6 was efficiently obtained (40% in two reaction steps) from alcohol 5 by reaction with tosyl chloride in basic medium at room temperature. Finally, the precursor L-nordeprenyl-D 2 was synthesized by a first step of deprotecting the derivative L-Boc-amphetamine 4 in the presence of TFA, followed by reaction of N-alkylation with d2-propargyl tosylate 6 using K 2 CO 3 and DMF as solvent. The precursor as free base was stored in freezer at −20°C, where its purity was 99.1% controlled by HPLC for 9 months (data not shown).
Through the development of this methodology it was possible to generate the Lnordeprenyl-D 2 precursor with an overall yield of 33% in five synthetic steps and purity of 99,1% by HPLC and 1 H-NMR analysis. The structure of the compounds synthesized was confirmed using analytical and spectroscopic techniques such as 1 H NMR mono and bidimentional (COSY), 13 C NMR and HETCOR (HSCQ and HMBC) experiments, IR and MS spectroscopy.
We have recently described the fully automated synthesis of [ 11 C]D-deprenyl tracer by one-step N-alkylation with [ 11 C]MeOTf using the commercially platform GE TRA-CERlab® FX C Pro (Scheme 4) (Buccino et al. 2016). This methodology initially provided us great potential of [ 11 C]MeOTf for reducing the amount of precursor and synthesis time, as well as for increasing radiochemical yields and reproducibility.
The use of the free base version of the precursor D-nordeprenyl had a positive impact in the radiochemical yield of [ 11 C]D-deprenyl. Because of these results, in the present Scheme 3 Organic synthesis of (1,1-d 2 )Propargyl p-toluenesulphonate 6 Scheme 4 Radiosynthesis of [ 11 C]D-deprenyl by one-step N-alkylation with [ 11 C]MeOTf work we proposed the use of the precursor L-nordeprenyl-D 2 as free base for its labelling with [ 11 C]MeOTf. Using the commercially available hydrochloride salt of Lnordeprenyl-D 2 , (Buccino et al. 2016), the overall radiochemical yield was 24 ± 9% (n = 10) (end of synthesis, decay corrected from [ 11 C]MeI), but it increased to 44 ± 13% (n = 12) with the employment of the L-nordeprenyl-D 2 free base (yields are referred to [ 11 C]MeI, even when [ 11 C]MeOTf is the radioactive precursor in the labelling reaction; TRACERlab®FX C Pro allows to measure activities of [ 11 C]MeI but not those of [ 11 C]MeOTf ). The use of the aqueous NaOH to neutralize the hydrochloride salt is no longer necessary, and losses of radioactivity in the form of [ 11 C]MeOH (possible product of hydrolysis of the radioactive precursor [ 11 C]MeOTf ) are diminished. This fact can be appreciated in the radioactivity profile trapped in the reactor during the labelling step (Fig. 1). We could observe an increased amount of [ 11 C]L-deprenyl-D 2 (peak at t R = 7.5 min) in the semipreparative gamma chromatograph when free base precursor was used, being this compound more than 80% of the injected radioactivity. When the salt is used, this value decreased to less than 50%, and one major 11 C-containing impurity at t R = 4.0 min was found (Fig. 2). In order to confirm the identity of these radiochemical impurities observed during the radiosynthesis of [ 11 C]L-deprenyl-D 2 using the different precursors, a series of blank experiments were performed (Fig. 3). When bubbling [ 11 C]MeOTf in anhydrous MEK (350 uL), after heating to 80°C for 1 min., a major compound (95%) eluted at t R = 2.6 min, which was assigned to unreacted [ 11 C]MeOTf, and a minor compound (5%) at t R = 3.0 min. That could correspond to [ 11 C]MeOH (hydrolysis product of [ 11 C]MeOTf). This minor peak increased its We hypothesized that the major impurity observed for precursor L-nordeprenyl-D 2 .HCl might be [ 11 C]MeCl, product of the nucleophilic attack of chloride anion to [ 11 C]MeOTf. In order to confirm this assumption, [ 11 C]MeOTf was collected in MEK containing 5 uL of 80% Benzalconium chloride (organic-soluble chloride salt). After heating to 80°C for one minute, the chromatogram showed a major peak at t R = 4.1 min, which validated our original hypothesis. Volatilisation of [ 11 C]MeCl (Boiling point −23,8°C, at 1 atm) during heating could explain the loss of radioactivity observed in this step when the precursor is in its hydrochloride form.
Radiochemical purity of [ 11 C]L-deprenyl-D 2 obtained using this methodology was 99.7 ± 0.6% (n = 12) and Specific activity was 170 ± 116 GBq/μmol (n = 12). Other QC parameters (such as ethanol and residual solvents concentrations, pH, half-life and radionuclidic purity) were in agreement with United States or European Pharmacopeas for all the batches produced with this methodology (n = 12).
These results are in concordance with those presented by Wilson et al. 2000; in which radiochemical yields of [ 11 C]raclopride (from [ 11 C]MeI) were very poor (<10%) when HBr salt of the radiolabelling precursor was used. These authors identified the major product as [ 11 C]MeBr, which is less reactive than [ 11 C]MeI for nucleophilic attack. Langer et al. 1999; also reported a similar finding when desmethyl-raclopride.HBr salt was used. In that case, when [ 11 C]MeOTf is used as 11 C-methylating agent, HBr salt of the precursor of [ 11 C]raclopride only yielded [ 11 C]MeBr as labelled product. These findings allow us to conclude that the use of the free base form of the precursor of [ 11 C]L-deprenyl-D 2 presents many advantages in comparison to the hydrochloride salt, fundamentally in terms of radiochemical yield. Losses of radioactivity are decreased and radiochemical purity of crude [ 11 C]L-deprenyl-D 2 is increased, which affect dramatically the overall yield of the radiopharmaceutical process.

Conclusions
A facile and efficient synthesis of L-nordeprenyl-D 2 precursor of [ 11 C]L-deprenyl-D 2 as free base has been developed in five synthetic steps with an overall yield of 33%. The precursor as free base has been stable for 9 months stored at low temperature (−20°C). An efficient automated synthetic method for [ 11 C]L-deprenyl-D 2 has been performed using L-nordeprenyl-D 2 free base and [ 11 C]MeOTf as methylating agent. This methodology offers a short preparation time (about 35 min) and simplicity in operation for routine preclinical and clinical studies.