“In-loop” [11C]CO2 fixation: Prototype and proof of concept

Article abstract of DOI:10.1002/jlcr.3528

Carbon-11-labeled carbon dioxide is the most common feedstock for the synthesis of positron emission tomography radiotracers and can be directly used for ¹¹C-carbonylation. Herein, we report the development of an apparatus that takes advantage of “in-loop” technologies to facilitate robust and reproducible syntheses of ¹¹C-carbonyl-based radiotracers by [¹¹C]CO₂-fixation. Our “in-loop” [¹¹C]CO₂-fixation method is simple, efficient, and proceeds smoothly at ambient pressure and temperature. We selected model ¹¹C-carbonyl-labeled carbamates as well as symmetrical and unsymmetrical ureas based on their widespread use in radiotracer design and our clinical research interests for proof of concept. Utility of this method is demonstrated by the synthesis of a reversible radiopharmaceutical for monoamine oxidase B, [¹¹C]SL25.1188, and 2 novel fatty acid amide hydrolase inhibitors. These radiotracers were isolated and formulated (>3.5?GBq; 100?mCi) with radiochemical purities (>99%) and molar radioactivity (≥80?GBq/μmol; ≥2162?mCi/μmol).

Full text of DOI:10.1002/jlcr.3528

Full-Paper
“In-Loop” [11C]CO2-Fixation: Prototype and
Proof-of-Concept
Kenneth Dahl1, Thomas L. Collier1, 2, Ran Chang1, Xiaofei Zhang1,
Oleg Sadovski3, Steven H. Liang1* and Neil Vasdev1*
1Division of Nuclear Medicine and Molecular Imaging, Massachusetts General Hospital, and
Department of Radiology, Harvard Medical School, 55 Fruit Street, Boston, MA 02114,
USA.
2Advion Inc., 10 Brown Road, Ithaca, NY 14850, USA.
3Research Imaging Centre, Centre for Addiction and Mental Health, 250 College Street,
Toronto, Ontario, Canada MST 1R8.
*E-mail: liang.steven@mgh.harvard.edu; vasdev.neil@mgh.harvard.edu; Fax: +1 617-726-
6165; Tel: +1 617-643-4736
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.3528
This article is protected by copyright. All rights reserved.

Abstract
Carbon-11 labelled carbon dioxide is the most common feedstock for the synthesis of
11
positron emission tomography radiotracers, and can be directly used for C-carbonylation.
Herein, we report the development of an apparatus that takes advantage of “in-loop”
technologies to facilitate robust and reproducible syntheses of 11C-carbonyl-based
radiotracers by [11C]CO2-fixation. Our “in-loop” [11C]CO2-fixation method is simple,
efficient, and proceeds smoothly at ambient pressure and temperature. We selected model
11C-carbonyl labelled carbamates as well as symmetrical and unsymmetrical ureas based on
their widespread use in radiotracer design and our clinical research interests for proof-of-
concept. Utility of this method is demonstrated by the synthesis of a reversible
radiopharmaceutical for monoamine oxidase B, [11C]SL25.1188, as well as two novel fatty
acid amide hydrolase inhibitors. These radiotracers were isolated and formulated (>3.5 GBq;
100 mCi) with radiochemical purities (>99%) and molar radioactivity (≥80 GBq/µmol;
≥2162 mCi/mol).
Key words: CO2-fixation; FAAH, MAO-B, Carbonylation; Carbon-11, Loop
Introduction
Carbon-11 (11C; t1/2 = 20.4 min) is a valuable radionuclide in positron emission
tomography (PET).1-4 The high positron emission decay mode (β+; 99.8%) and low energy
11
protons (0.96 keV) allow for high resolution in-vivo imaging. Another advantage is that C
can be incorporated into bioactive or endogenous molecules without any structural
modification (isotopologue of carbon-12) and with minimal effect on the (bio)chemical
11
properties of the compound. Moreover, the short half-life of C allows for multiple in vivo
PET imaging studies to be undertaken in the same subject or animal on a single day. The
most widely used reaction in 11C-radiopharmaceutical production is 11C-methylation, via
[11C]CH3I or [11C]CH3OTf. However 11C-carbonylation is also a viable strategy using
[11C]CO2, [11C]CO, [11C]COCl2 or [11C]HCN.5
This article is protected by copyright. All rights reserved.

[11C]CO2 is a highly attractive building block for radiolabelling, since it is produced
directly in target. However, due to low chemical reactivity and trapping efficiency, the direct
incorporation of [11C]CO2 into organic molecules poses a significant challenge. The
traditional methods for “[11C]CO2-fixation” rely on Grignard reagents, organolithiums or
organosilanamines.5 All such reagents require the rigorous exclusion of atmospheric moisture
and CO2 during storage and manipulation. The direct formation of 11C-labelled carbonyl
groups by [11C]CO2-fixation, has been recently established in radiopharmaceutical
production,5-6 where strong organic bases, e.g. 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU)7
or
2-tert-Butylimino-2-diethylamino-1,3-dimethylperhydro-1,3,2-diazaphosphorline
(BEMP),8-9 are utilized to reversibly capture [11C]CO2 in solution. While initial work focused
11
on the synthesis of C-labelled carbamates, the scope of the method rapidly extended to
syntheses of unsymmetrical [11C]ureas and [11C]oxazolidinones via POCl3 dehydration to
generate 11C-labelled isocyanates as a key intermediates,9 as well as 11C-carboxylic acids via
metal-catalyzed aryl boronation reactions.10-11
Despite evident advantages of [11C]CO2-fixation for radiotracer production, the
radiochemistry has been generally restricted to semi-manual vial-based methods.
Automation of this methodology would be highly desirable, not only from a user’s
perspective, as it would simplify the radiosynthesis process and reduce exposure to
radioactivity, but also to improve its reproducibility and compliance of routine
radiopharmaceutical production under current good manufacturing practices (cGMP). The
aim of this study was two-fold: 1) to develop a simple and easily automated [11C]CO2-
fixation platform; and 2) to carry out proof-of-concept radiosyntheses of 11C-carbonyl
labelled compounds with diverse scaffolds.
The traditional vial-based procedure was abandoned and we sought to take advantage
of the captive solvent or “in-loop” technologies, that have been established for Grignard
11
reactions12-15, or C-methylation,16-17 specifically, ease of automation, reliability, efficiency
(minimal radioactive loss to the surrounding atmosphere, needles, septa, or to transfer from
vessels), and in particular, its high surface-to-volume ratio, and should be ideally suited for
11
[11C]CO2-fixation radiochemistry. We selected model C-carbonyl labelled carbamates as
well as symmetrical and unsymmetrical ureas based on their widespread use in radiotracer
design and our clinical research interests for the proof-of-concept radiolabelling by “in loop
[11C]CO2-fixation” (Table 1 and Scheme 1). The radiotracers that were isolated in this study
11
include, two new and structurally similar C-isotopologues of a fatty acid amide hydrolase
This article is protected by copyright. All rights reserved.

(FAAH) inhibitor, [11C]JNJ1661010 ([11C]5) and [11C]7, as well as the monoamine oxidase B
(MAO-B) radiopharmaceutical, [11C]SL25.1188 ([11C]2). 18-20
Results and Discussion
The “in-loop” [11C]CO2-fixation apparatus is designed as a standalone system, where
crude 11C-carbonylated products are generated for subsequent purification by
chromatography. The method described herein consists of trapping and reacting [11C]CO2
directly in a stainless steel loop (as commonly used for sample injection in HPLC). The
system is divided into two main parts: 1) the gas-phase entrapment and purification of in-
target produced [11C]CO2; and 2) the “in-loop” reactor system, where [11C]CO2 is fixated and
reacted to furnish the final 11C-carbonyl labelled product (Figure 1). For easy adoption of this
method into a standard hot cell, we integrated the “in loop” system to a commercial
radiosynthesis module (GE TracerLab FXFN) to carry out automated HPLC purification and
formulation of the labelled compounds.
Our initial focus was to optimize conditions for the “in-loop” [11C]CO2-fixation
reaction. The synthesis of [11C]methyl-N-benzylcarbamate ([11C]11) was selected as a model
reaction (Table 1). The experiments were carried out by bubbling [11C]CO2 in a stream of N2
(20 mL/min) into the reactor loop with benzylamine (4.6 µmol) and the CO2 fixating base,
BEMP (2.5 µL, 8.6 µmol), dissolved in DMF, which was pre-loaded in section A of the
reactor loop (Figure 1). The radiochemical yields (RCY) shown in Table 1 is non-isolated
and determined by radio-HPLC. Initially, the intermediate [11C]carbamate ion ([11C]9) was
11
treated with methyl iodide (Table 1, Route 2) and to our delight, the desired C-labelled
carbamate product was obtained in nearly quantitative yield (Table 1, entry 1). Next, we set
to explore the [11C]isocyanide formation ([11C]10, Table 1, Route 1). The [11C]11 formation
was studied using different amounts of benzylamine (0.15 - 4.6 µmol), with POCl3 (0.2% v/v
in MeCN) as the dehydrating reagent, and the reaction mixture was quenched with an excess
of methanol (50% in DMF) pre-loaded in section B of the reactor loop (Figure 1). As
anticipated, the reaction was strongly dependent on the amount of benzylamine (8), as larger
amounts of 8 led to formation of the undesirable symmetrical product, [11C]dibenzylurea
([11C]12). For example, using 4.6 µmol of 8, only 16% of [11C]11 was obtained with [11C]12
as the main 11C-labelled product (Table 1, entry 2). However, by reducing the amount of 8 to
This article is protected by copyright. All rights reserved.

0.3 µmol, a high and reproducible non-isolated RCY of 83 ± 2% (n = 3) was obtained.
Further reduction of 8 by 1-fold resulted in a reduced yield (Table 1, entry 6).
11
One potential concern with C-reactions via captive solvent methods is the general
dependency on solvents with high viscosity or boiling points, for example, DMF and DMSO.
MeCN is the most common solvent in [11C]CO2-fixation radiochemistry.21-25 Our “in-loop”
reactions demonstrated excellent compatibility with MeCN as a solvent in the reaction to
prepare [11C]11, which resulted in excellent non-isolated RCY (91%, Table 1, entry 7).
Quantitative trapping efficiency was achieved for all examined solvents (MeCN, DMSO and
DMF). However, a large variation in selectivity to the desired product was observed under
these conditions (Table 1, entries 6-8). Despite superior yield obtained with MeCN, DMF
was selected as the optimal solvent for the subsequent reactions, due to evaporation of MeCN
that occurred while degassing with helium during our solution preparation. It is noteworthy
11
that this degassing step was crucial in our hands for isolating C-labelled compounds with
high molar radioactivity. Finally, in a last attempt to further increase the yield, BEMP was
replaced with another commonly used fixating base, DBU. High trapping efficiency of
11
[11C]CO2 was obtained with DBU (>99%); however, the overall yield to the desired C-
labelled product was reduced. In summary, our model reaction showed that ideal “in-loop”
11C-labelling conditions were achieved with BEMP (8.6 µmol), benzylamine (0.3 µmol)
dissolved in DMF (40 µL), followed by elution with 0.2% v/v POCl3 in MeCN (100 µL) and
finally quenching with an excess of 50% v/v methanol in DMF (50 µL).
To further exemplify the utility of this newly developed method, using the established
conditions for the “in-loop” [11C]CO2-fixation reaction (Table 1, entry 5), we set out to
prepare three radiotracers (vide supra), namely [11C]SL25.1188 ([11C]2), [11C]JNJ1661010
([11C]5), and [11C]7 (Scheme 1). The radiotracers were synthesized using both the standard
vial-based method and our novel “in-loop” method (Table 2). In all cases, the “in-loop”
methodology attained higher non-isolated RCYs compared with conventional reactions in a
vial. For example, [11C]SL2511.88 ([11C]2), was obtained in a two-fold higher non-isolated
RCY using the “in-loop” methodology (Table 2, entry 1). It is important to note that our
previous synthesis of [11C]2 resulted in comparable RCYs (60-70%) when using a vial and
MeCN as the solvent.15 However, the experimental conditions (including the solvent) for both
methods were kept constant, to facilitate a direct comparison. Moreover, a staggering 15-fold
improvement in non-isolated RCY was obtained for both FAAH inhibitor radioligands,
[11C]JNJ1661010 ([11C]5) and [11C]7, using the “in-loop” method (Table 2, entries 2 and 3).
This article is protected by copyright. All rights reserved.

It is noteworthy that higher amounts of the sparingly reactive primary amines (3 and 6, 55
µmol and 75 µmol, respectively) were necessary to reach the high RCYs (Table 2, entries 2
and 3), indicating that the [11C]carbamate ion formation is promoted by higher amine
concentration. The difference in RCYs between the two methods could be explained by the
high surface-to-volume ratio inside the loop. Consequently, the “in-loop” method generates
an ideal environment for efficient [11C]CO2 trapping and promotes the formation of the
[11C]carbamate ion.
Carbon-11 labelled PET radiopharmaceuticals for in-vivo human studies are typically
produced in GBq quantities. Therefore, to establish the utility of the current method, all three
compounds were prepared on a large scale, starting from 110 GBq (3 Ci) of [11C]CO2.
Compounds were isolated using semi-preparative HPLC followed by solid-phase extraction
(SPE) to generate the final formulated product (for HPLC traces of [11C]2, [11C]5 and [11C]7
see Figures 2 – 4, respectively).
Table 2 shows that all compounds were produced in high RCP (>99%) and molar
radioactivities (80-134 GBq/µmol (2162-3621 mCi/mol)). Carbon-11 labelled 7 was
isolated and formulated in sufficient quantities for in-vivo PET imaging studies (3.8 GBq;
103 mCi; 18% decay-corrected RCY, relative to starting [11C]CO2). Compounds [11C]2 and
[11C]5 were obtained in very high (>50%) isolated RCYs, with 10.2 and 14.3 GBq (276 and
387 mCi), respectively, ready for injection. PET imaging studies with all of these radiotracers
are underway in our laboratories, and we are presently designing a fully-automated version of
the “in-loop” apparatus for cGMP radiopharmaceutical production. We anticipate that the
simplicity of this “in-loop” methodology will facilitate widespread use and will provide
comparable radiochemical yields and specific radioactivities to our previously developed 11C-
11
urea- and carbamate-based radiotracers.21-23 Our future work will focus on C-carboxylation
and amidation as innovative methodologies to achieve these reactions in radiopharmaceutical
production are unveiled.
This article is protected by copyright. All rights reserved.

Conclusion
We describe herein the development and design of a novel “in-loop” [11C]CO2-
fixation apparatus, which offers an efficient method for the production of a wide range of
11
11
11
11C-carbonylated radiotracers including C-oxazolidinones, C-ureas, and C-carbamates.
The “in-loop” methodology was applied in the preparation of three 11C-carbonyl labelled PET
radiotracers ([11C]2, [11C]5, [11C]7) in high isolated yields (>3.5 GBq) and radiochemical purity
(>99%). The molar radioactivity of all 3 compounds were above 80 GBq/µmol (≥2162
Ci/mmol) at the end-of-synthesis. Given its simplicity, we anticipate the current “in-loop”
[11C]CO2 fixation methodology to be widely implemented in the development and routine
production of PET radiopharmaceuticals.
Experimental section
Materials and general methods
All reagents and solvents were obtained from commercially available sources and
used without further purification, unless specified otherwise. Anhydrous grade MeCN,
MeOH and DMF was purchased from Acros organics (USA); BEMP, DBU, aniline (3),
benzylamine, methyl iodide, 1,2-benzisoxazol-3-amine (6), 3-phenyl-5-piperazino-1,2,4-
thiadiazole (4), JNJ-1661010 (5) as well as phosphorus(V) oxychloride (POCl3) were all
obtained from Aldrich (USA). A Tracerlab FXFN radiofluorination module (GE, Sweden)
installed with a C18 semi-preparative HPLC column (ACE 5, C18-HL, 250 x 10 mm,
Waters) was used for automated purification of the radiolabelled compounds. A HLB C18
light (30 mg, Waters) cartridge was used for SPE. An analytical HPLC system included a
high-pressure gradient pump (Shimadzu LC-10AD pump), variable wavelength UV-detector
(Shimadzu SPD-10AV), and a radioactivity detector. The system was controlled by ClarityTM
6.1 chromatography software. The radiochemical purity (RCP) of the isolated products was
determined by reverse phase HPLC. Analysis was monitored using 254 nm as the detection
wavelength. Identification of all radioactive products was confirmed by co-elution with the
corresponding non-radioactive compound. The specific radioactivity (SA) was determined by
HPLC.
This article is protected by copyright. All rights reserved.

Prototype design
A schematic of the [11C]CO2 “in-loop” synthesizer is shown in Figure 1. All
components were purchased from commercial sources. The operation of the module was
controlled remotely via electrical switches from outside the hot cell. All reagents and solvents
were conveniently stored in loops and transferred using controlled N2 gas flow (2 Bar)
generated by a standard mass-flow controller (MFC, 0-2000 mL/min, OMEGA). The
radioactive entrapment and release were monitored using scintillation detectors (Caroll &
Ramsey). [11C]CO2 was trapped from the target using a stainless steel coil (0.5-m, 1/16” o.d.,
0.04” i.d.) immersed into a vertically driven (air-pressure) Dewar vessel filled with nitrogen.
Two 3-port, two-way valves (V1 and V2, P/N 009-0933-900, Parker) were used to direct N2
flow or target gas flow onto the liquid nitrogen trap. An 8-port, two-way HPLC injection
valve (V3, P/N 170-0160H, Vici) was equipped with a reagent loop (1 mL (2-m), TefzelTM,
1/16” o.d., 0.03” i.d.) as well as a reactor loop (3.5 mL (7-m), stainless-steel, 1/16” o.d.,
0.03” i.d.), which was applied as reaction vessel. Importantly, the reactor loop is dividable
into two different sections: section A (1 mL (2-m)) and section B (2.5 mL (5-m)). Two union-
connectors (U1 and U2, P/N ZU1M, Valco) were used to combine the sections. This dual-
loop system makes it possible to pre-load two solutions at the same time. A sodium
hydroxide-coated silica trap (Ascarite II, 20-30 mesh) was connected to the waste port of the
HPLC valve and placed in a well counter to measure unreacted [11C]CO2.
Model reaction for the “in-loop” [11C]CO2-fixation
(a) Set-up. A schematic of the reactor loop is displayed in Figure 5. Prior to the start-
of-synthesis with the injection valve in position A and by disconnecting the finger tight luer
adaptors (Figure 5, U1, P-655 and P-660, Idex Health & Science), a solution of alkylating (10
µL, methyl iodide) or dehydrating (0.2 µL, POCl3) agent in MeCN (100 µL) were loaded into
the reagent loop using a 1 mL syringe. Furthermore, to the same loop, additional amount of
MeCN (800 µL) was added. A solution of appropriate amine, BEMP (2.5 µL) in
dimethylformamide (40 µL, DMF), was loaded onto the reactor loop (Figure 5, section A). If
“Route 1” (Table 1, scheme) are to be used, a solution of methanol (25 µL) dissolved in DMF
(25 µL) was also loaded onto the reactor loop (Figure 5, section B). To enable reagent
injection to the different sections of the reactor loop, the inlet nut on union-connectors
(Figure 5, U2 and U3) were replaced by a needle injection port (#VISF-2, Valco). It is
important to note that prior to use, all solutions were de-gassed with helium flow for 15 min
and kept over 4Å molecular sieves to minimize atmospheric CO2.
This article is protected by copyright. All rights reserved.

(b) Trapping and reaction. No-carrier-added [11C]CO2 was prepared using a PET
trace cyclotron (GE medical system, Sweden) via irradiation of a nitrogen target using 16.4
MeV protons. After 1-min irradiation at 50 µA, typically 5 GBq (140 mCi) of [11C]CO2 was
obtained at end-of-bombardment (EOB). The [11C]CO2 was transferred from the target
chamber by the target gas (1% O2 in N2) and concentrated in the liquid nitrogen trap of the
[11C]CO2-fixation apparatus (Figure 1). The stainless-steel tube was removed from the liquid
nitrogen by an air-pressure driven lift, whereby the [11C]CO2 was passed through the coated
reaction loop (Figure 5, Step 1) in a controlled stream of N2 (20 mL/min). When radioactivity
peaked in the reaction loop, as measured by the proximal radiation detector (typically within
2 min), the flow of N2 was stopped and the reaction was allowed to proceed for an additional
2 min. Following this reaction, the position of the injection valve was switched to position B
(Figure 5, Step 2) and the content of the reactor loop was eluted out with a second reagent
(alkylating or dehydration reagent in MeCN) by a stream of nitrogen (80 mL/min) into an
empty receiving vial (5 mL, glass v-vial, Alltech). Moreover, the pre-loaded MeCN was
immediately flushed through the reactor loop to maximize recovery. The crude radioactive
product was dissolved in 3.5 mL of water and injected onto a GE TracerLab FXFN
radiofluorination module for HPLC purification and formulation.
(c) Clean-up. The [11C]CO2-fixation module was easily cleaned by passing 5 mL
volumes of water, MeCN and acetone through both loops (reactor and reagent loop) with the
injection valve in position B (Figure 5), and flushed using N2 gas (500 mL/min) for 15 min.
Precursor and non-radioactive standard synthesis
Precursors and the non-radioactive standards for the [11C]CO2-fixation reactions were
either synthesized in-house (1, 2 and 7) or were obtained from commercially available
sources (3, 4, 5 and 6). Compounds included in this study are depicted in Scheme 1.
SL25.1188 reference standard (2) and the corresponding precursor compound (1) were
synthesized in 5- and 6-steps, respectively, starting from 4-(benzyloxy)-2-fluorobenzonitrile
via reduction, alkylation, and hydrolysis (Scheme 2).18, 26 Moreover, the non-radioactive
standard 7 was synthesized in two-steps starting from 6 (Scheme 3).
This article is protected by copyright. All rights reserved.

Synthesis of SL25.1188 reference (2) and precursor (1) compounds
6-(Benzyloxy)benzo[d]isoxazol-3-amine (13). Acetone oxime (20 mmol) was added to
a solution of 1 M potassium tert-butoxide in THF (20 mL, 20 mmol). The solution was kept
at room temperature (r.t.) for 20 min, after which a solution of 4-(benzyloxy)-2-
fluorobenzonitrile (18.5 mmol) in THF (15 mL) was slowly added using an addition funnel.
After stirring for 3 h at r.t., the mixture was heated at 60 ºC overnight. A dark brown solution
was obtained. The reaction mixture was quenched with water (10 mL) and partitioned
between saturated NaHCO3 solution (5 mL) and ethyl acetate (75 mL). The organic layers
were separated and washed with water (3 x 30 mL). A brown solid was obtained upon solvent
removal using a rotary evaporator. The solid was then treated with a mixture of EtOH (40
mL), H2O (26 mL) and HCl (12 N, 14 mL) at reflux for 2 h. After cooling, the reaction
mixture was basified with solid sodium carbonate and NaOH (1 N, 15 mL), followed by
extraction with ethyl acetate (2 x 100 mL). The combined organic layers were washed with
brine (50 mL), dried over MgSO4, and the solvent was removed under reduced pressure. The
residue was purified using chromatography on silica gel by eluting with a 1:3 mixture of
1
ethyl acetate and hexane. The product was obtained as a white solid. Yield: 62%; H NMR
(300 MHz, CDCl3) δ 7.31 – 7.48 (m, 6H), 6.90 – 6.97 (m, 2H), 5.12 (s, 2H), 4.30 (s, 2H).
Ethyl 6-(benzyloxy)benzo[d]isoxazol-3-ylcarbamate (14). Ethyl chloroformate (1.2
mmol) was added dropwise to a solution of 13 (1 mmol) and pyridine (2 mmol) maintained at
º
º
0 C. The mixture was stirred at 0 C for 4 h, poured into water and extracted with ethyl
acetate (2 x 20 mL). The combined organic layers were washed with brine (20 mL), dried
over MgSO4, and the solvent was evaporated in vacuo. The residue was purified by
chromatography on silica gel by eluting with a 1:3 mixture of ethyl acetate and hexane. The
1
product was obtained as a white solid. Yield: 93%; H NMR (300 MHz, CDCl3) δ 8.08 (d, J
= 9.6 Hz, 1H), 7.86 (s, 1H), 7.30 – 7.49 (m, 5H), 6.95 – 7.03 (m, 2H), 5.13 (s, 2H), 4.32 (q, J
13
= 7.1 Hz, 2H), 1.36 (t, J = 7.1 Hz, 3H). C NMR (75 MHz, CDCl3) δ 166.1, 161.5, 153.1,
153.0, 135.9, 128.7, 128.3, 127.5, 125.2, 114.6, 108.9, 93.5, 70.5, 62.4, 14.4.
(S)-3-(6-(Benzyloxy)benzo[d]isoxazol-3-yl)-5-(methoxymethyl)oxazolidin-2-one (15).
A solution of 14 (0.9 mmol) in 1 mL of DMF was added to a solution of (R)-4-
(methoxymethyl)-1,3-dioxolan-2-one (1.7 mmol) and potassium carbonate (0.085 mmol) in 2
mL DMF at 140 ºC. The mixture was stirred at 140 ºC for 2 h. After cooling the solution was
evaporated in vacuo. The residue was purified by chromatography on silica gel by eluting
This article is protected by copyright. All rights reserved.

with a 1:3 mixture of ethyl acetate and hexane. The product was obtained as a white solid.
1
Yield: 62%; H NMR (300 MHz, CDCl3) δ 8.41 (d, J = 9.7 Hz, 1H), 7.30 – 7.49 (m, 5H),
6.95 – 7.02 (m, 2H), 5.14 (s, 2H), 4.85 – 4.96 (m, 1H), 4.23 (dd, J = 9.9, 8.9 Hz, 1H), 4.11
13
(dd, J = 10.0, 6.5 Hz, 1H), 3.69 (qd, J = 10.8, 4.1 Hz, 2H), 3.45 (s, 3H). C NMR (75 MHz,
CDCl3) δ 166.4, 161.5, 153.8, 152.9, 135.9, 128.7, 128.3, 127.5, 126.5, 114.5, 108.2, 93.4,
73.8, 72.4, 70.5, 59.6, 46.6.
(S)-3-(6-Hydroxybenzo[d]isoxazol-3-yl)-5-(methoxymethyl)oxazolidin-2-one (16). A
solution of 15 (8.47 mmol) and Pd/C (500 mg) in 100 mL EtOH/THF (1:1) under hydrogen
atmosphere was stirred at r.t. for 60 min. Upon completion of the reaction, the mixture was
filtered through Celite® and the filtrate was evaporated in vacuo. The residue was purified by
chromatography on silica gel by eluting with a 1:2 mixture of ethyl acetate and hexane. The
1
product was obtained as a white solid. Yield: 95%; H NMR (300 MHz, DMSO) δ 10.46 (s,
1H), 8.19 (d, J = 8.9 Hz, 1H), 6.89 (d, J = 2.0 Hz, 1H), 6.83 (dd, J = 8.9, 2.1 Hz, 1H), 4.88 –
5.06 (m, 1H), 4.16 (t, J = 9.3 Hz, 1H), 3.85 (dd, J = 9.6, 6.3 Hz, 1H), 3.54 – 3.72 (m, 2H). 13C
NMR (75 MHz, DMSO) δ 166.3, 161.0, 154.2, 153.1, 126.5, 114.4, 106.8, 95.0, 74.5, 72.6,
59.1, 46.5.
(S)-5-(Methoxymethyl)-3-(6-(4,4,4-trifluorobutoxy)benzo[d]isoxazol-3-yl)oxazolidin-
2-one (SL25.1188, 2). 16 (7.6 mmol) was added to a solution of 1,1,1-trifluorobutane (10.5
mmol) and potassium carbonate (15.2 mmol) in 28 mL MeCN. The mixture was heated to
reflux for 2 h. The mixture was than cooled, filtered and evaporated under reduced pressure.
The residue was purified by chromatography on silica gel by eluting with a 1:3 mixture of
1
ethyl acetate and hexane. The product was obtained as a white solid. Yield: 76%. H NMR
(300 MHz, CDCl3) δ 8.41 (d, J = 9.7 Hz, 1H), 6.82 – 6.95 (m, 2H), 4.84 – 5.00 (m, 1H), 4.20
– 4.27 (m, 1H), 4.03 – 4.17 (m, 3H), 3.69 (qd, J = 10.8, 4.1 Hz, 2H), 3.45 (s, 3H), 2.25 – 2.45
(m, 2H), 2.04-2.20 (m, 2H). 19F NMR (282 MHz, CDCl3) δ -66.3 (t, J = 10.8 Hz).
(S)-1-Methoxy-3-(6-(4,4,4-trifluorobutoxy)benzo[d]isoxazol-3-ylamino)propan-2-ol
(1). To a solution of 2 (1.4 mmol) in EtOH (7 mL), aq. NaOH (1.3 N, 5.4 mL, 7 mmol) was
added. The mixture was stirred at r.t. overnight. The resulting solution was diluted with water
(50 mL) and EtOAc (50 mL). The combined organic layers were washed with brine (20 mL),
dried over MgSO4, and the solvent was evaporated in vacuo. The residue was purified by
chromatography on silica gel by eluting with a 1:1 mixture of ethyl acetate and hexane. The
This article is protected by copyright. All rights reserved.

product was obtained as a white solid. Yield: 90%; 1H NMR (300 MHz, DMSO) δ 7.73 (d, J
= 8.7 Hz, 1H), 6.99 (d, J = 1.9 Hz, 1H), 6.78 – 6.88 (m, 2H), 4.96 (d, J = 5.2 Hz, 1H), 4.08 (t,
J = 6.2 Hz, 2H), 3.89 (dq, J = 10.5, 5.3 Hz, 1H), 3.21 – 3.34 (m, 5H), 3.04 – 3.16 (m, 1H),
2.30 – 2.46 (m, 2H), 1.86 – 2.01 (m, 2H). 13C NMR (75 MHz, DMSO) δ 164.1, 161.1, 158.7,
128.1 (q, J = 277.0 Hz), 122.5, 112.4, 110.1, 94.0, 75.5, 67.8, 66.8, 58.8, 46.8, 29.9 (q, J =
30.0 Hz), 21.9 (d, J = 3.0 Hz). 19F NMR (282 MHz, DMSO) δ -64.8 (t, J = 11.6 Hz).
Synthesis of reference compound (7)
N-(benzo[d]isoxazol-3-yl)-4-(3-phenyl-1,2,4-thiadiazol-5-yl)piperazine-1-
carboxamide (7). 2,2,2-Trichloroethyl chloroformate (15mg, 72 µmol) was added to a
solution of benzo[d]isoxazol-3-amine (6, 8.4 mg, 60 µmol) and pyridine (5.7 mg, 72 µmol) in
º
º
THF (0.6 mL) maintained at 0 C. The mixture was kept at 0 C for 1.5 h under stirring. The
resulting solution was quenched with water and subsequently extracted with EtOAc (3 x 5
mL). The combined organic layers were washed with brine (20 mL), dried over MgSO4, and
the solvent was removed in vacuo to yield the crude intermediate product (17). Next, N-
ethyldiisopropylamine (4.2 mg, 32 µmol) was added to a solution of 17 (8.0 mg, 26 µmol)
and 3-phenyl-5-(piperazin-1-yl)-1,2,4-thiadiazole (4, 8.5 mg, 34 µmol) in DMSO (0.6 mL).
º
The mixture was stirred at 70 C for 2 h, quenched with water and subsequently extracted
with EtOAc (2 x 5 mL). The combined organic layer was washed with brine and dried over
anhydrous MgSO4 and concentrated in vacuo. The residue was purified by preparative TLC
by eluting with a 10:1 mixture of dichloromethane and methanol. The product was obtained
1
as a yellow solid. Yield: 32% over two steps; H NMR (300 MHz, DMSO-d6) δ 10.01 (s,
1H), 8.14 – 8.09 (m, 2H), 7.85 (d, J = 7.8 Hz, 1H), 7.65 – 7.55 (m, 2H), 7.48 – 7.45 (m, 3H),
13
7.33 – 7.27 (m, 1H), 3.72 (s, 4H), 3.68 (s, 4H). C NMR (75 MHz, DMSO-d6) δ 185.2,
169.5, 163.4, 155.8, 154.3, 133.3, 130.7, 130.5, 129.1, 128.0, 125.1, 123.3, 117.3, 110.2,
48.8, 43.7
Acknowledgement
We would like to thank the staff at the Radiochemistry Program, Massachusetts General
Hospital and Harvard Medical School for their support. R.C. is supported by China
Scholarship Council (201506250036). S.H.L. is a recipient of an NIH career development
award (DA038000). N.V. thanks National Institute on Ageing of the NIH for funding this
work (R01AG054473).
This article is protected by copyright. All rights reserved.

References
[1]
[2]
[3]
[4]
[5]
PW. Miller, NJ. Long, R. Vilar, AD. Gee, Angew Chem Int Ed. 2008, 47, 8998.
SM. Ametamey, H. Honer, PA. Schubiger, Chem. Rev, 2008, 108, 1501.
VW. Pike, Curr. Med. Chem, 2016, 23, 1818.
K. Dahl, C. Halldin, M. Schou, Clin Transl Imaging. 2017, 5, 257.
BH. Rotstein, SH. Liang, MS. Placzek, JM. Hooker, AD. Gee, F. Dollé, AA. Wilson,
N. Vasdev, Chem Soc Rev. 2016, 45, 4708.
[6]
[7]
BH. Rotstein, SH. Liang, JP. Holland, TL. Collier, JM. Hooker, AA. Wilson, N.
Vasdev, Chem Comm. 2013, 49, 5621.
JM. Hooker, A. Reibel, S. Hill, M. Schueller, JS. Fowler, Angew Chem Int Ed. 2009,
48, 3482.
[8]
[9]
AA. Wilson, A. Garcia, S. Houle, N. Vasdev, Org Biomol Chem. 2010, 8, 428.
AA. Wilson, A. Garcia, S. Houle, O. Sadovski, N. Vasdev, Chem Eur J. 2011, 17,
259.
[10] PJ. Riss, S. Lu, S. Telu, FI. Aigbirhio, VW. Pike, Angew Chem Int Ed, 2012, 51,
2698.
[11] BH. Rotstein, JM. Hooker, J. Woo, TL. Collier, TJ. Brady, SH. Liang, N. Vasdev,
ACS Med Chem Lett, 2014, 5, 668.
[12] RJ. Davenport, VW. Pike, K. Dowsett, DR. Turton, K. Poole, Appl Radiat Isot. 1997,
48, 917.
[13] RJ. Davenport, K. Dowsett, VW. Pike, Appl Radiat Isot. 1997, 48, 1117.
[14] M. Matarrese, F. Sudati, D. Soloviev, S. Todde, EA. Turolla, MG. Kienle, F. Fazio,
Appl Radiat Isot. 2002, 57, 675.
[15] M-R. Zhang, M. Ogawa, Y. Yoshida, K. Suzuki, Appl Radiat Isot. 2006, 64, 216.
[16] AA. Wilson, A Garcia, L Jin, S Houle, Nucl Med Biol. 2000, 27, 529.
[17] AA. Wilson, A. Garcia, S. Houle, N. Vasdev, J Label Compd Radiopharm. 2009, 52,
490.
[18] W. Saba, H. Valette, M. Peyronnea, Y. Bramoulle, C. Coulon, O. Curet, P. George, F.
Dolle, M. Bottlaender Synapse. 2010, 64, 61.
[19] N. Vasdev, O. Sadovski, A. Garcia, F. Dollé, JH. Meyer, S. Houle, AA. Wilson, J
Label Compd Radiopharm. 2011, 54, 678.
[20] P Rusjan, AA. Wilson, L. Miler, I. Fan, R. Mizrahi, S. Houle, N. Vasdev, JH. Meyer,
J Cereb Blood Flow Metab. 2014, 34, 883.
[21] JW. Hicks, J. Parkes, O. Sadovski, J. Tong, S. Houle, N. Vasdev, AA. Wilson, Nucl
Med Biol. 2013, 40, 740.
[22] JW. Hicks, AA. Wilson, EA. Rubie, JR. Woodgett, S. Houle, N. Vasdev, Bioorg Med
Chem Lett. 2012, 22, 2099.
[23] AA. Wilson, JW. Hicks, O. Sadovski, J. Parkes, J. Tong, S. Houle, C. Fowler, N.
Vasdev, J Med Chem. 2013, 56, 201.
[24] A. Dheere, N. Yusuf, AD. Gee, Chem Comm. 2013, 49, 8193.
[25] AV. Mossine, AF. Brooks, IM. Jackson, CA. Quesada, P. Sherman, EL. Cole, DJ.
Donnelly, PJH. Scott, X. Shao, Bioconjugate Chem. 2016, 27, 1382.
[26] S. Jegham, F. Puech, P. Burnier, PCT Int. Appl. 1996: WO9638444 A1.
This article is protected by copyright. All rights reserved.

Scheme 1: Compounds radiolabelled at room temperature using the “in-loop” [11C]CO2-fixation module.
This article is protected by copyright. All rights reserved.

Scheme 2: Synthesis procedure for compounds 1 and 2.
Scheme 3: Synthesis procedure for compound 7. (32% yield over 2 steps).
This article is protected by copyright. All rights reserved.

Figure 1: Schematic drawing of the “in-loop” [11C]CO2-fixation module.
This article is protected by copyright. All rights reserved.

Figure 2: A) Semi-preparative HPLC trace (55:45 MeCN:10 mM NH4HCO2, at a flow rate 7 mL/min, ACE 5
C18, 5 µm, 10 x 250 mm) of a typical radiosynthesis of [11C]2, purified using a GE TracerLab FXFN
radiosynthesis module. Gamma (tR = 14.7 minutes; top), and UV (λ = 254 nm; bottom). Units on the y-axis are
arbitrary. B) Analytical HPLC trace (50:50 MeCN:H2O, at a flow rate of 1 mL/min, XBridge C-18, 3.5 µm, 150
x 4.6 mm) of a typical analysis of [11C]2. Gamma (tR ~ 5 minutes; top), and UV (λ = 254 nm; bottom). Units on
the y-axis are arbitrary.
This article is protected by copyright. All rights reserved.

Figure 3: A) Semi-preparative HPLC trace (55:45 MeCN:10 mM NH4CH3CO2, at a flow rate of 5 mL/min,
ACE 5-C18, 5 µm, 10 x 250 mm) of a typical radiosynthesis of [11C]5, purified using a GE TracerLab FXFN
radiosynthesis module. Gamma (tR = 11.8 minutes; top), and UV (λ = 254 nm; bottom). Units on the y-axis are
arbitrary. B) Analytical HPLC trace (50:50 MeCN:H2O, at a flow rate of 1.5 mL/min, XBridge C-18, 3.5 µm,
150 x 4.6 mm) of a typical analysis of [11C]5. Gamma (tR ~ 6.5 minutes; top), and UV (λ = 254 nm; bottom).
Units on the y-axis are arbitrary.
This article is protected by copyright. All rights reserved.

Figure 4: A) Semi-preparative HPLC trace (55:45 MeCN:10 mM NH4CH3CO2, at a flow rate 5 mL/min, ACE
5-C18, 5 µm, 10 x 250 mm) of a typical radiosynthesis of [11C]7, purified using a GE TracerLab FXFN
radiosynthesis module. Gamma (tR = 12.7 minutes; top), and UV (λ = 254 nm; bottom). Units on the y-axis are
arbitrary. B) Analytical HPLC trace (50:50 MeCN:H2O, at a flow rate of 1.5 mL/min, XBridge C-18, 3.5 µm,
150 x 4.6 mm) of a typical analysis of [11C]7. Gamma (tR ~ 8 minutes; top), and UV (λ = 254 nm; bottom). Units
on the y-axis are arbitrary.
This article is protected by copyright. All rights reserved.

Figure 5: General “in-loop” [11C]CO2-fixation radiosynthesis procedure. Step 1: The “in-loop” [11C]CO2
fixation in a pre-loaded trapping-solution (BEMP, amine dissolved in DMF) to form the intermediate product,
[11C]carbamate ion (reactor loop, section A). Step 2: Elution and “in-loop” reaction using a pre-loaded solution
(POCl3 or methyl iodide in reagent loop) in section A of the reactor loop, followed by a second in-loop reaction
with a pre-loaded nucleophile (alcohol or amine) in section B of the reactor loop.
This article is protected by copyright. All rights reserved.

Table 1. “In-loop” formation of [11C]11 and [11C]12
Benzylamine (8)
Trapped
RCY of
RCY of
Entry
(µmol)
[11C]CO2 (%)c
[11C]11 (%)d
[11C]12 (%)d
1b
2a
4.6
4.6
>99
>99
98±1h
-
16
68
3a
4a
1.3
0.6
>99
>99
37
52
46
29
5a
6a
7a, e
0.3
0.15
0.3
>99
>99
>99
83±2h
4
1
4
70
91
8a, f
9a, g
0.3
0.3
>99
>99
2
10
52
13
a Reaction conditions: Using Route 1, benzylamine (0.15-4.6 µmol), BEMP (2.5 µL, 8.6 µmol),
DMF (40 µL), 0.2% POCl3 (v/v) in MeCN (100 µL, 1.1 µmol), 50% MeOH (v/v) in DMF (50
µL, 1.2 mmol). bUsing Route 2, with 10% iodomethane (v/v) in MeCN (100 µL) as alkylating
agent. Percentage [11C]CO2 left in loop after entrapment. Non-isolated radiochemical yield
p
d
e
(RCY), based on radio-HPLC analysis of the crude product. MeCN was used as solvent.
fDMSO was used as solvent. DBU was used as fixation base. (n = 3); all other entries unless
g
h
noted represent single experiments.
This article is protected by copyright. All rights reserved.

Table 2. Radiotracers synthesized using vial and “in-loop” [11C]CO2-fixation
Vial
“In-loop”
Radioactivity at
EOS
Molar
methoda
methodb
RCP
Synthesis
Entry Compound
radioactivity
(%)e time (min)
RCY(%)c,
(GBq (mCi))f
(GBq/µmol)g
RCY(%)c, d
d
1
2
3
[11C]2
[11C]5
[11C]7
33
5
70 (43)h
72 (54)h
30 (18)h
>99
>99
>99
35
33
38
10.2±2.8 (276±74)
14.3 (387)
134±21
84
2
3.8 (102)
80
aConditions for vial reaction. (1) Precursor 1 (0.6 µmol); primary amines 3 and 6 (110 and 150
mol, respectively); BEMP (5 µL, 34.6 µmol), DMF (80 µL); (2) 0.2% POCl3 (v/v) in MeCN (100
µL, 1.1 µmol); (3) secondary amine (4) in DMF (100 µL). bConditions for “in-loop” reaction. (1)
Precursor 1 (0.3 µmol); primary amines 3 and 6 (55 and 75 mol, respectively); BEMP (2.5 µL,
17.3 µmol), DMF (40 µL); (2) 0.2% POCl3 (v/v) in MeCN (100 µL, 1.1 µmol); (3) secondary
c
amine (4) in DMF (50 µL). Non-isolated radiochemical yield (RCY), based on radio-HPLC
d
e
analysis of the crude product. Experiments were performed once (n=1). Radiochemical purity
(RCP) of the isolated product. Total amount of the isolated product at the end-of-synthesis. gMolar
radioactivity of the isolated product established by HPLC. The decay-corrected radiochemical
f
h
yield of the isolated product obtained after semi-preparative HPLC and formulation.
This article is protected by copyright. All rights reserved.