A new concept for the production of 11C-labelled radiotracers

Background The GMP-compliant production of radiopharmaceuticals has been performed using disposable units (cassettes) with a dedicated synthesis module. To expand this “plug ‘n’ synthesize” principle to a broader scope of modules we developed a pressure controlled setup that offers an alternative to the usual stepper motor controlled rotary valves. The new concept was successfully applied to the synthesis of N-methyl-[11C]choline, L-S-methyl-[11C]methionine and [11C]acetate. Results The target gas purification of cyclotron produced [11C]CO2 and subsequent conversion to [11C]MeI was carried out on a TRACERlab Fx C Pro module. The labelling reactions were controlled with a TRACERlab Fx FE module. With the presented modular principle we were able to produce N-methyl-[11C]choline and L-S-methyl-[11C]methionine by loading a reaction loop with neat N,N'-dimethylaminoethanol (DMAE) or an ethanol/water mixture of NaOH and L-homocysteine (L-HC), respectively and a subsequent reaction with [11C]MeI. After 18 min N-methyl-[11C]choline was isolated with 52% decay corrected yield and a radiochemical purity of > 99%. For L-S-methyl-[11C]methionine the total reaction time was 19 min reaction, yielding 25% of pure product (> 97%). The reactor design was used as an exemplary model for the technically challenging [11C]acetate synthesis. The disposable unit was filled with 1 mL MeMgCl (0.75 M) in tetrahydrofuran (THF) bevore [11C]CO2 was passed through. After complete release of [11C]CO2 the reaction mixture was quenched with water and guided through a series of ion exchangers (H+, Ag+ and OH−). The product was retained on a strong anion exchanger, washed with water and finally extracted with saline. The product mixture was acidified and degassed to separate excess [11C]CO2 before dispensing. Under these conditions the total reaction time was 18 ± 2 min and pure [11C]acetate (n = 10) was isolated with a decay corrected yield of 51 ± 5%. Conclusion Herein, we described a novel single use unit for the synthesis of carbon-11 labelled tracers for preclinical and clinical applications of N-methyl-[11C]choline, L-S-methyl-[11C]methionine and [11C]acetate. Supplementary Information The online version contains supplementary material available at 10.1186/s41181-022-00159-y.

Page 2 of 14 Wenz et al. EJNMMI Radiopharmacy and Chemistry (2022) 7:6 Background Nearly 100 years ago, the production of carbon-11 ( 11 C) was reported for the first time (Crane and Lauritsen 1934). The ability to label and metabolically track almost every biologically relevant carbon-containing compound by 11 C makes this radionuclide an indispensable marker in molecular imaging by positron emission tomography (PET) in modern nuclear medicine. A common access to 11 C-labelled radiopharmaceuticals is the direct activation of [ 11 C]CO 2 with organometallic reagents. Additionally, it can be converted into [ 11 C] CH 3 I (Langstrom et al. 1987;Larsen et al. 1997) or [ 11 C]CH 3 OTf (Jewett 1992) and then used for radiolabelling. In both cases, the radioactive building block is used in gas form with an inert carrier gas (usually helium). The actual labelling step is performed by bubbling the gas mixture into a reactor, a reaction loop or on a solid matrix. Tracer specific work-up by solid phase extraction (SPE) and cartridge purification has become widely accepted often eliminating the need for time-consuming HPLC separation. A schematic overview is depicted in Fig. 1.
The used materials can have a significant influence on the success or failure of the radio synthesis. To minimize problems, GMP-compliant production of radiopharmaceuticals is preferably performed using very expensive disposable units (cassettes). The synthesis then proceeds according to a "plug 'n' synthesize" principle and delivers reliable results. In addition, cross-contamination with other in-house produced pharmaceuticals and handling errors caused by personnel are minimized.
The major clinical benefit of N-methyl-[ 11 C]choline in PET has been in the evaluation of patients with biochemically recurrent prostate cancer (Hara et al. 1998;Scattoni et al. 2007;Reske et al. 2008;Rinnab et al. 2008;Mitchell et al. 2012), and currently, for localization of parathyroid adenoma (Orevi et al. 2014;Liu et al. 2020;Noltes et al. 2021). L-S-methyl-[ 11 C]methionine is a 11 C-labelled analogue of the essential amino acid methionine. It enters several metabolic pathways and has been used in various applications such as imaging brain tumours (Sato et al. 1992), hyperparathyroidism (Sundin et al. 1996;Rubello et al. 2006), neck and head tumours (Lindholm et al. 1993). Currently it is used for the evaluation of multiple myeloma including the staging, the prognostication and the assessment of therapy's response (Lapa et al. 2017;Morales-Lozano et al. 2020;Lückerath et al. 2015). The synthesis Fig. 1 General description of 11 C-tracer synthesis. 1. Reaction step, 2. Workup and transport of the product mixture to SPE purification cartridge(s), 3. Washing step and 4. Product extraction Page 3 of 14 Wenz et al. EJNMMI Radiopharmacy and Chemistry (2022) 7:6 of both N-methyl-[ 11 C]choline and L-S-methyl-[ 11 C]methionine are quite robust and they are produced routinely in many institutions with a cyclotron.
[ 11 C]Acetate has been used as imaging agent for studying myocardial oxidative metabolism (Henes et al. 1989;Grassi et al. 2012;Nesterov et al. 2015) and regional myocardial blood flow (Gropler et al. 1991) as well as for prostate cancers (Oyama et al. 2002).
The most common method for the production of [ 11 C]CO 2 is the irradiation of a 14 N 2 target with protons following the reaction [ 14 N(p,α) 11 C]. The unavoidable presence of [ 16 O]oxygen ( 16 O 2 ) traces in the target gas leads to the formation of undesired [ 13 N]nitrogen during the irradiation, that is present in form of nitrogen oxides (NO x ) and has to be removed before further conversions (Ache and Wolf 1966). Modern synthesis units use molecular sieves for this purpose, as they can be regenerated and used over a long period of time with low maintenance. This process starts by binding [ 11 C] CO 2 selectively onto the molecular sieve. Subsequent purging with an inert carrier gas removes the unwanted by-products. At high temperatures (> 250 °C) pure [ 11 C]CO 2 is released and can either be used directly for the synthesis (CO 2 Bypass) or be converted to [ 11 C]methyl-iodide. In the so-called "gas phase" conversion it is reduced with hydrogen at 360 °C to [ 11 C]methane on a heterogeneous nickel catalyst and then reacted with iodine at 720 °C to produce [ 11 C]CH 3 I (Larsen et al. 1997) (Fig. 2 left). The individual syntheses for N-methyl-[ 11 C]choline (Pascali et al. 2000;Jinming et al. 2006;Kuznetsova et al. 2003), L-S-methyl-[ 11 C]methionine (Lodi et al. 2008) and [ 11 C]acetate (Berridge et al. 1995;Kruijer et al. 1995;Roeda et al. 2002;Soloviev and Tamburella 2006;Kang Se et al. 2016;Maurer et al. 2019;Mitterhauser et al. 2004) have been optimised in the last decades and are reviewed in detail (Oleksiy et al. 2013;Gomzina et al. 2015;Dahl et al. 2017). A general overview is shown in Fig. 2.
While the requirements toward materials and equipment for 11 C methylation reactions are undemanding, as no aggressive solvents or sensitive reagents are required. The conditions for carboxylation reactions, however, require the use of organometallic compounds. These are stable in organic solvents such as tetrahydrofuran or diethyl ether which are not compatible with many polymers used in regular disposable cassettes due to possible corrosion of the material. Furthermore, organometallic reagents are sensitive to air and moisture, which also leads to difficulties in handling.
Despite the important applications, there is currently no cassette-like system that can fulfil all these requirements. This prompted us to develop a novel flexible disposable setup for the synthesis of tracers according to the synthesis principle shown in Fig. 1.

C-precursor production
For the production of 11 C-labelled compounds from [ 11 C]CO 2 a TRACERlab FX C Pro (GE Healthcare) system was used as synthesis module. To use purified [ 11 C]CO 2 directly valve (V x ) was installed (CO 2 bypass, see Fig. 3). No further modifications to the module were made (for details see Additional file 1: Tables S1, S2).

General modifications to the TRACERlab FX FE Pro
The synthesis module was cleaned and dried before each synthesis and used as a control unit for the disposable setup. The 2-way valve V4 was replaced by a 3-way valve and connected to an ascarit ® (II) trap. This allowed pressure exhaust during the carboxylation reaction and trapping of unreacted [ 11 C]CO 2 . The V18 and V19 lines which are usually connected to the module reactor, were connected to the waste vessel instead. The product vessel was filled via the usually closed second neck with a tube or cannula reaching to the bottom of the vial (See Fig. 4). The reactors activity detector was dismounted and placed close to the disposable unit to monitor the activity profile during the reaction.

Disposable materials
For the assembly of the disposable synthesis units the following materials were used: Check valves (

N-Methyl-[ 11 C]choline
The solvent reservoirs of the TRACERlab FX FE module were prepared as following: Vial 1: 4 mL 0.9% Saline (Extract), vial 2: Empty, vial 3: Empty, vial 4: Empty, vial 5: Empty, vial 6: 5 mL water (Wash), vial 7: 5 mL ethanol (Wash), product vessel: Empty, reaction loop: 100 µl DMAE. Process description: The disposable unit was assembled as described in Fig. 5 and the loop was filled with DMAE. The SPE cartridge was placed in front of the activity detector of the reactor (B) and the reaction loop was positioned close to the product detector (A). The solvent reservoirs were connected and the synthesis program was started. Monitoring and semi-automatic control was performed by using the pressure and activity detectors (for details see Additional file 1: Table S3).

[ 11 C]Acetate
The disposable unit was assembled as described in Fig. 7 and the solvent reservoirs of the TRACERlab FX FE module were prepared as following: Vial 1: Empty, vial 2: 2 mL water (to quench the reaction), vial 3: Empty, vial 4: 15 mL Water (for transport and washing), vial 5: Empty. Vial 6: 15 mL Water (Washing), Vial 7: 4 mL 0.9% Saline (Product extraction from anion exchange cartridge), product vessel: 1 mL citrate buffer (for removal of [ 11 C]CO 3 2− ). Process description: The ion exchangers were pre-conditioned and the parts were assembled as described in Fig. 7. The solvent reservoirs were connected and the synthesis program was started. A helium overpressure was generated via V5 and CV A was closed. The reactor was purged with dry argon for 5 min via the "normally CO 2 " input. Then the MeMgCl solution was carefully added to the reactor via the septum. Afterwards the argon line was exchanged for the [ 11 C]CO 2 line. The anion exchange cartridge was placed in front of the reactor activity detector (B) and the reaction vessel was positioned close to the product detector (A) (for a detailed description see Additional file 1: Table S5).

Results
The loop  Since clogging during hydrolysis of the MeMgCl THF solution is a major problem in [ 11 C]acetate synthesis, a reactor design was chosen to prevents this issue. The reaction was optimised by irradiation of the gas target for 2 min with 10 µA beam current yielding A EOB ([ 11 C]CO 2 ) = 2000 MBq. After 18 ± 2 min, we were able to isolate the desired product in 52 ± 5% (n = 9) yield. And to simulate clinically relevant conditions, a higher starting activity of [ 11 C]CO 2 (A EOB = 25 GBq) by irradiating the target for 5 min with 60 µA beam current was used as well. The exemplarily diagrams shown in Table 1 were taken from this experiment. The total troduction time was 17 min and 6.6 GBq (decay corrected 51%) of the product could be isolated.
In order to estimate the activity distribution, all the components of the synthesis apparatus were dismounted and analysed after the reaction. The decay corrected activity was 8% for the exhaust ascarite ® (II) trap of the reactor, 4% for the PS-H + cartridge, 7% for the 3 Ag + cartridges and 3% for the anion exchanger. 14% of the total activity was measured in the waste vessel. The exhaust line of the product vessel was also connected to an ascarite ® (II) trap, the received activity is in direct correlation to the carbonate content (8%) of the saline extraction solution. Eight percent of the total activity was not found. We suspect that these traces remained on the remaining parts.
Quality control was performed according to literature and was in accordance with monograph of Pham Eu. for [ 11 C]acetate and L-S-methyl-[ 11 C]methionine injection solutions. The results are summarized in Table 2.

Discussion
The main objective was to use disposable, commercially available materials without complicated modifications. A summarized general setup is shown Fig. 8. The individual reaction paths were controlled externally by pressure regulation-and the special feature is the use of check valves or even 3-way valves as control units.
The basic principle of this assembly consists of pluggable elements that allow convenient customisation. T-shaped luer adapters, which are available in various designs serve as the framework. A "loop" or a "reactor" setup is possible in both cases. For the Page 9 of 14 Wenz et al. EJNMMI Radiopharmacy and Chemistry (2022) 7:6 latter, a disposable syringe body with septum is used. Luer check valves are used to attach the desired SPE cartridges for cleaning. Additionally, a T-shaped valve is connected at the end to separate wash and extraction solution. Solvent and gas supply were controlled by the synthesis module. A typical reaction proceeds as following:  Radiopharmacy and Chemistry (2022) 7:6 The check valves CV A , CV B and CV C are closed by a gas overpressure (via the orange adapter) and the gaseous radio-precursor is then passed through a check valve into the apparatus. The gas flow is automatically directed into the reactor or reaction loop where the chemical reaction takes place. Pressure equalisation is permitted through the connected exhaust line. The gas supply is then terminated and the waste vessel evacuated. This opens CV A and CV B and the reaction mixture can be guided through the SPE cartridge and be washed. Once all washing steps are completed, purging with Helium continues until the generated overpressure closes CV B again. By opening the exhaust valve at the product vessel CV C is opened. The product can then be extracted, sterile filtered and dispensed.
[ 11 C]Choline and [ 11 C]methionine are produced on a regular basis in our clinic. The total yield for methylation reactions is largely dependent on the efficient production of [ 11 C]methyliodide. This process produced about 60% decay corrected [ 11 C] MeI with respect to the produced [ 11 CO] 2 in the period of our study. Since the general reaction conditions were not changed, both the reaction times and the yields obtained are within the range of the syntheses described in literature (Oleksiy et al. 2013;Gomzina et al. 2015;Dahl et al. 2017). Due to the high sensitivity to water, the use of disposable products is especially desirable for Grignard reagents. The first attempts for the [ 11 C]acetate synthesis were performed with a loop setup were as well. Despite a very precise work, we were not able to achieve a stable and reliable reaction. Fluctuating yields, high carbonate content in the reaction mixture and occasionally clogging of the reaction loop was observed. Therefore, further experiments were carried out in a disposable syringe as a reactor.
With the selected helium gradient (Additional file 1: Table S1) for the delivery of the purified [ 11 C]CO 2 from the molecular sieve an efficient trapping in 1 mL THF (even without Grignard) was achieved. After 180 s of gas injection, less than 10% of the total activity was monitored on the exhaust ascarite trap.
Reliable and constant results were obtained using similar conditions reported by Kang Se et al. (2016). Methyl magnesium chloride (1 mL, 0.75 M) in tetrahydrofuran was used as an alkylation reagent. To guarantee a constant concentration a stock solution was used for all experiments. Under these conditions, about 60% [ 11 C]acetate was formed. The remaining activity was split among the known by-products [ 11 C]acetone, [ 11 C]tert-butanole and [ 11 C]carbonate. A lower Grignard concentration led to a higher formation of [ 11 C]carbonate, an increase resulted in an elevated formation of [ 11 C]acetone and [ 11 C]tert-butanole.
Product trapping on an anion exchange resin can only be effective when all the chloride (0.75 mmol) is removed from the reaction mixture (Kruijer et al. 1995). In order to verify the latter, we added an aqueous silver nitrate solution (1%) to the reaction mixture after the cation exchanger procedure. The absence of a colourless precipitate indicated complete chloride removal. Thus, the ion exchangers were adjusted to the corresponding amount of Grignard reagent in "cold" pre-tests.
As the reaction mixture needed to pass through five ion exchanger cartridges, the waste vessel was evacuated during the washing process. The uncharged by-products [ 11 C]acetone and [ 11 C]tert-butanole were separated completely within an acceptable period of time. After extraction with saline, the mixture of anionic components [ 11 C] acetate and [ 11 C]carbonate was transferred into aqueous citrate buffer solution (1 mL, 139 mM). The resulting solution had a pH between 5 and 6 and by helium bubbling for 90 s [ 11 C]carbonate was efficiently removed as [ 11 C]CO 2 (Maurer et al. 2019). In summary, we were able to produce [ 11 C]acetate within 17 ± 2 min and decay corrected yield of 51 ± 5% pure (n = 10).

Conclusion
The used materials are sterile, affordable and single-use products. Furthermore, the key materials are available in inert materials, such as polyethylene (PE) or polypropylene (PP) which is essential for the use of sensitive chemicals and organic solvents. As a large variety of modules allow pressure/vacuum control, this design can easily be adapted for specific application. This concept may not only help in the establishment of new 11 C-tracers for research and routine clinical applications but also to improve the synthesis of established 11 C-labelled PET-tracers, especially, when no cassette based modules are available or reactive chemicals have to be used. We are convinced that this setup