Direct Incorporation of [11C]CO2 into Asymmetric [11C]Carbonates.

A novel carbon-11 radiolabelling methodology for the synthesis of the dialkylcarbonate functional group has been developed. The method uses cyclotron-produced short-lived [11C]CO2 (half-life 20.4 min) directly from the cyclotron target in a one-pot synthesis. Alcohol in the presence of base trapped [11C]CO2 efficiently forming an [11C]alkylcarbonate intermediate that subsequently reacted with an alkylchloride producing the di-substituted [11C]carbonate (34% radiochemical yield, determined by radio-HPLC) in 5 minutes from the end of [11C]CO2 cyclotron delivery.

Introduction

Positron emission tomography (PET) is an imaging technique able to detect and monitor specific target proteins in vivo [1-5]. The use of PET imaging has advanced in the last few decades to become a valuable tool in clinical diagnostics, medical research, and drug discovery [6-8]. PET relies on the use of tracer amounts of imaging probes (radiotracers). The administration of radiotracers allows the biochemical process to be imaged and quantified in vivo without manifestation of pharmacological or toxicological effects [9-13].

Carbon-11 (11C) is one of the most common radionuclides used for the synthesis of PET radiotracers. The short half-life of 11C (20.4 min) makes it an attractive radionuclide as it enables the collection of a sufficient amount of PET data while keeping the subject radiation dose and exposure time to minimum. Furthermore, it allows orthologous substitution with carbon-12 in biologically active molecules with no alteration of the parent molecule’s physicochemical and pharmacological properties. Carbon-11 is commonly produced in the form of [11C]carbon dioxide ([11C]CO2) [14, 15]. [11C]CO2 is usually converted into more reactive secondary precursors such as [11C] methyl iodide ([11C]CH3I), [11C]carbon monoxide ([11C] CO), and [11C]phosgene ([11C]COCl2) [16-19]. As these multistep conversion processes are time-consuming, the use of [11C]CO2 for directly radiolabelling functional groups is highly attractive.

[11C]CO2 is a weak electrophile with an affinity for electron-donating reagents such as amines and organometallics [20]. However, due to the thermodynamic and kinetic properties of [11C]CO2, it has high activation energy which requires the use of highly reactive reagents, temperatures, pressures, or the presence of a catalyst [21-23]. Nevertheless, the primary synthon, [11C]CO2, has been deployed successfully for the synthesis of 11C-compounds that contain carbonyl groups such as [11C]carbamates [24, 25], amide [26], and [11C]ureas [23, 27–29]. However, the radiolabelling of the carbonyl group of carbonates from [11C]CO2 has not yet been established. To date, the synthesis of [11C]carbonates has relied on the use of [11C]COCl2 which is produced from a multistep process starting from cyclotron-produced [11C]CO2 or [11C]CH4, conversion to [11C]CCl4 and then to [11C]COCl2 [30, 31]. Although this 11C-carbonate reaction is rapid and efficient, routine production of [11C]COCl2 requires multistep syntheses and specialized equipment, thereby restricting its widespread use [30, 31].

As the carbonate functional group is found in prodrug compounds as well as being an intermediate in organic synthesis [32-35], we aimed at developing a simple and robust radiolabelling methodology that uses [11C]CO2 for the synthesis of [11C]carbonates. Here we present a rapid, one-pot radiosynthetic strategy using [11C]CO2 directly from the cyclotron, avoiding the need for specialized equipment and multistep syntheses.

Materials and Methods

All purchased chemicals were used without further purification. Chemicals were purchased in highest available purity from Sigma-Aldrich and Alfa Aesar and used as received (>99 % purity). All solvents were purchased as anhydrous in highest available purity (>99.8 % purity) from Sigma-Aldrich.

[11C]CO2 was produced by a Siemens RDS112 cyclotron (St Thomas’ Hospital, London, United Kingdom) via the 14N(p,α)11C nuclear reaction. Typical irradiation time for exploratory work was 1 minute, 10 μA, bombardment typically yielding ca. 300 MBq [11C]CO2 at end of cyclotron bombardment. Radiolabelling reactions were performed in a 1.5 mL screw top vial with a “V” internal shape. HPLC analysis was performed on an Agilent 2060 Infinity HPLC system with a variable wavelength detector (254 nm was used as default wavelength) [10] An Agilent Eclipse XDB-C18 reverse-phase column (4.6 × 150 mm, 5 μm) was used at a flow rate of 1 mL/min and H2O/MeOH (HPLC-grade solvents with 0.1 % TFA) gradient elution (flow rate: 1 mL/min, 0–2 min: 5 % MeOH, 2–11 min: 5 to 95% MeOH linear gradient, 11–13 min: 90 % MeOH, 13–14 min: 90% to 5% MeOH linear gradient, and 14–15 min: 5 % MeOH). The RCY was estimated by radio-HPLC and defined as the area under the 11C-product peak expressed as a percentage of the total 11C labelled peak areas observed in the chromatogram. Molar radioactivity was calculated from analytical HPLC sample of 25 μL. A calibration curve of known mass quantity versus HPLC peak area (254 nm) was used to calculate the mass concentration of the 25 μL radiolabelled compound. The identity of the radiolabelled compound peak was confirmed by HPLC coinjection of a nonradioactive reference compound and yielded a single peak.

Results and Discussion

As the starting point, we selected the method developed by Salvatore et al. [21-23] (Figure 1) for the synthesis of carbonates. The established method used nonradioactive CO2, an alcohol derivative, and benzylchloride (BzCl) in the presence of Cs2CO3, TBAI in DMF to produce the corresponding carbonate derivative efficiently. By substituting CO2 with [11C]CO2 and applying the same reaction conditions, the synthesis of di-substituted [11C]carbonates was investigated.

Figure 1

Method by Salvatore et al. [21–23] for the synthesis of carbonates using nonradioactive CO2.

[11C]CO2 was trapped in isopropyl alcohol in the presence of Cs2CO3, forming an [11C]isopropylcarbonate intermediate that subsequently reacted with BzCl to produce [11C]benzyl isopropyl carbonate ([11C]1) in a moderate radiochemical yield (RCY). The RCY is the nonisolated radiochemical yield determined by radio-HPLC analysis of the crude product of 24% (Table 1, entry 1). Interestingly, almost all the cyclotron-produced [11C]CO2 was trapped within the reaction mixture at room temperature (>95%); any unreacted radioactive [11C]CO2 was immobilized on an ascarite trap connected to the vial vent needle. The trapping efficiency is the amount of radioactivity trapped in the reaction vial as a percentage of the overall radioactivity produced by the cyclotron.

Table 1

Optimisation of [11C]1 synthesis.

Entrya	Base	Trapping efficiency (%)	Temperature (°C)	Solvent	RCY (%)b
1	Cs2CO3	95.2	25	DMF	24
2	Cs2SO4	1.5	25	DMF	0
3	CsI	4.3	25	DMF	5
4	CsF	33.5	25	DMF	0
6	K2CO3	10	25	DMF	0
7	CaCO3	0	25	DMF	0
8	Cs2CO3	20	25	CH3CN	0
9	Cs2CO3	65	25	DMSO	0
10	Cs2CO3	>95%	65	DMF	33
11c	Cs2CO3	>95%	100	DMF	82, 74

Reaction conditions: isopropanol (22 μmol), Cs2CO3 (66 μmol), TBAI (66 μmol), and organohalide (66 μmol) in 500 μL DMF, 10 mins from end of delivery (EOD) (n = 1).

The nonisolated radiochemical yield determined by radio-HPLC analysis of the crude product.

n = 2.

In an attempt to increase the RCY, Cs2CO3 was replaced with Cs2SO4 (Table 1, entry 2). The trapping efficiency of [11C]CO2 dropped significantly from 95.2% to 1.5%. Since Cs2CO3 contributed towards the trapping of [11C]CO2 efficiently, we investigated whether the Cs+ or the CO3 2- ion was responsible for the high [11C]CO2-trapping efficiency. Of a number of caesium bases explored (Table 1, entries 3–5), CsI and CsF trapped only minute amounts of [11C]CO2 (4% and 34%, respectively), indicating that the basicity of the reaction mixture had a major effect on trapping efficiency. The results can be explained by the ability of a strong base to deprotonate the alcohol present in the reaction mixture enabling it to react with [11C]CO2 to form a 11C radiolabelled intermediate. The importance of CO3 2− was then explored by comparing Cs2CO3 with other carbonate bases (K2CO3 and CaCO3, Table 1, entries 6 and 7). The trapping efficiencies were extremely low for both reagents. High trapping in the reaction mixture with Cs2CO3 is therefore most likely due to its superior solubility in organic solvents.

In a further attempt to increase the RCY of [11C]1, a number of aprotic solvents were screened (CH3CN and DMSO, Table 1, entries 8 and 9). However, these solvents did not produce [11C]1, and the trapping efficiency was poor (20% and 65%, respectively). Reaction dependency on temperature was subsequently examined. The RCY of [11C]1 improved from 24% to 33% by increasing the reaction temperature from 25°C to 65°C (Table 1, entry 10). Increasing the temperature to 100°C promoted the product formation and resulted in the highest observed RCY (82%, Table 1, entry 11). This might be rationalised by an increase in Cs2CO3 solubility at higher temperatures. However, due the presence of Cs2CO3 as a reagent, low molar activities (A m) were observed. The low A m (2 GBq/μmol in this case) is likely due to release of nonradioactive CO2 from Cs2CO3. CO3 2− deprotonates the alcohol to form HCO3 −, which at high temperature has the potential to decompose releasing nonradioactive CO2 causing isotopic dilution and low A m of the [11C]CO2. We therefore focused on improving A m by substituting Cs2CO3 with an alternative base.

1,8-diazabicyclo[5.4.0]undecene (DBU) is a basic amine that has been shown to retain [11C]CO2 in organic solutions [26]. Replacing Cs2CO3 with DBU (Table 2, entry 1) resulted in [11C]1 formation, but with low RCY (6%). The low RCY could be due to DBU being unable to deprotonate isopropyl alcohol efficiently. We opted for a stronger base, NaH, which was able to deprotonate the isopropyl alcohol. Using a ratio of 1 : 1 NaH : isopropanol (equiv.) at 100°C, [11C]1 was obtained with an RCY of 26% (Table 2, entry 3). Decreasing the temperature from 100°C to 60°C slightly improved the RCY (31%, Table 2, entry 4). [11C]1 was produced with a molar activity (Am) of 10–20 GBq/μmol. This is because short cyclotron bombardments (1 minute) and low beam currents (5–10 μA) were used (0.3 GBq). In clinical productions at our facility, cyclotron bombardment times of 50 minutes and beam currents of 30 μA are used to produce higher amounts of radioactivity (typically 60 GBq). It is therefore estimated that this would increase the Am to > 50 GBq/μmol at end of synthesis. Decreasing the ratio of NaH : isopropanol (from 1 : 1 to 0.5 : 1) reduced the RCY further to 18% (Table 2, entry 5). Increasing the ratio NaH : isopropanol 2 : 1 did not produce the target product (Table 2, entry 6). Increasing the amount of TBAI to 3 equiv. or removing it completely also did not improve the RCY (Table 2, entries 7 and 8).

Table 2

Optimisation of [11C]1 synthesis using alternative bases.

Entrya	Base (eq)	TBAI (eq)	Temp (°C)	RCY (%)yb
1c	DBU (3)	3	100	6
2c	DBU (3)	—	100	0
3	NaH (1)	1	100	26
4d	NaH (1)	1	60	31 ± 2
5c	NaH (0.5)	1	60	18
6	NaH (2)	1	60	0
7c	NaH (0.5)	—	60	6
8	NaH (1)	3	60	7

Isopropanol (1 equiv., 22 μmol), BzCl (3 equiv.), TBAI (1−3 equiv.), and base (1−3 equiv.) in 500 μL DMF reaction time 5 mins from EOD.

The nonisolated radiochemical yield determined by radio-HPLC analysis of the crude product.

Reaction time of 10 mins from EOD.

n = 3.

Conclusions

In conclusion, we have developed a radiolabelling methodology for the synthesis of [11C]carbonates using [11C]CO2 directly from the cyclotron. The carbonate [11C]1 was synthesized by bubbling [11C]CO2 into a solution containing alkylchloride, alcohol, and a base in DMF. The choice of the base was critical for maximising the RCY and A m. The first protocol uses Cs2CO3 and produces the target 11C radiolabelled product in a high RCY and low A m. The second strategy, which uses NaH, produced [11C]1 in high A m and moderate RCY. These methodologies are a simple and practical alternative to 11C-phosgene for the synthesis of 11C-carbonates. 11C-phosgene synthesis is technically challenging to implement and requires the use of specialist equipment. The developed strategies described here use readily available labware and converts [11C]CO2 directly to [11C]carbonates in rapid synthesis times.