VMAT2 imaging agent, D6-[18F]FP-(+)-DTBZ: Improved radiosynthesis, purification by solid-phase extraction and characterization

Article abstract of DOI:10.1016/j.nucmedbio.2019.07.002

Objectives: Recently, a deuterated tracer, D6-[¹⁸F]FP-(+)-DTBZ, 9-O-hexadeutero-3-[¹⁸F]fluoropropoxyl-(+)-dihydrotetrabenazine ([¹⁸F]9), targeting vesicular monoamine transporter 2 (VMAT2) in the central nervous system, wa

Full text of DOI:10.1016/j.nucmedbio.2019.07.002

Nuclear Medicine and Biology 72–73 (2019) 26–35
Contents lists available at ScienceDirect
Nuclear Medicine and Biology
journal homepage: www.elsevier.com/locate/nucmedbio
VMAT2 imaging agent, D6-[18F]FP-(+)-DTBZ: Improved radiosynthesis,
purification by solid-phase extraction and characterization
Ruiyue Zhao a,b, Zhihao Zha b,c, Xinyue Yao a,b, Karl Ploessl c, Seok Rye Choi c, Futao Liu a,
Lin Zhu a, Hank F. Kung b,c,
⁎
a
College of Chemistry, Beijing Normal University, Beijing 100875, PR China
Department of Radiology, University of Pennsylvania, Philadelphia, PA 19104, USA
Five Eleven Pharma Inc., Philadelphia, PA 19104, USA
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Objectives: Recently, a deuterated tracer, D6-[18F]FP-(+)-DTBZ, 9-O-hexadeutero-3-[18F]fluoropropoxyl-(+)-
dihydrotetrabenazine ([18F]9), targeting vesicular monoamine transporter 2 (VMAT2) in the central nervous sys-
tem, was reported as a useful imaging agent for the diagnosis of Parkinson's disease (PD). The production of [18F]
9 was optimized and simplified by using solid-phase extraction (SPE) purification.
Received 3 April 2019
Received in revised form 6 June 2019
Accepted 6 July 2019
Available online xxxx
Methods: Three major nonradioactive impurities were synthesized and characterized. The preparation of [18F]9
was optimized by using different labeling conditions, and an SPE purification method was evaluated. The influ-
ence of chemical impurities in the final dose of [18F]9 was assessed by an in vitro binding assay, an assay of the
in vivo biodistribution in mice, and ex vivo and in vitro autoradiography of brain sections.
Keywords:
VMAT2
18F
Results: Optimized fluorination conditions for [18F]9 were found – heating at 130 °C for 10 min in DMSO, and a
high radiochemical yield and three major chemical impurities were observed. An SPE method involving a Sep-
Pak® tC18 Plus Light cartridge with a two-step elution process was successfully implemented. This process
gave a good radiochemical yield (38.7 10.5%, decay corrected; radiochemical purity N99%) and low chemical
impurities. An in vivo biodistribution study and autoradiography of brain sections showed that there was no sig-
nificant difference between HPLC-purified and SPE-purified [18F]9.
Deuterium
Tetrabenazine
SPE purification
Parkinson's disease
Conclusion: A VMAT2 targeting imaging agent, D6-[18F]FP-(+)-DTBZ, [18F]9, was prepared by optimized labeling
conditions and an easy SPE purification. This method offers a short preparation time and operational simplicity. In
conjunction with PET imaging, this new VMAT2 agent might be a useful clinical tool for diagnosing PD.
© 2019 Elsevier Inc. All rights reserved.
1. Introduction
Tetrabenazine (TBZ, Xenazine), used to treat hyperkinetic move-
ment disorders, was approved by the FDA in 2008 [3,4]. Relative to
Vesicular monoamine transporter 2 (VMAT2) plays an important
role in neuronal transmission and is critical in the mechanism of the
packing and storing of monoamine neurotransmitters in vesicles. Unlike
the environment on the synaptic membrane, where there are specific
and distinct transporters for the reuptake of three monoamines (dopa-
mine, serotonin and norepinephrine), VMAT2 nonselectively transports
all three monoamines from the cytosol into the vesicular lumen and it a
common ATP-dependent transporter. The VMAT2 transporters are ac-
tive only when VMAT2-containing neurons are active. Therefore, imag-
ing VMAT2 in the brain can elucidate the neuronal integrity (total
number) of all three monoaminergic neurons, which is important for
understanding neurological and psychiatric disorders [1,2].
TBZ, its metabolite, dihydrotetrabenazine (DTBZ), showed higher bio-
logical stability and less sensitivity to drugs affecting dopamine levels
in the brain. DTBZ has been labeled with the radionuclide 11C,
dihydrotetrabenazine ([11C]-(+)-DTBZ), which was successfully ap-
plied in the diagnosis of Parkinson's disease (PD). However, the natural
short half-life of 11C (T1/2 = 20 min) limits the widespread use of [11C]-
(+)-DTBZ [5–7]. An alternative VMAT2 binding agent,
a 9-
fluoropropoxyl derivative of dihydro-tetrabenazine ([18F]FP-(+)-
DTBZ, a.k.a. AV-133, florbenazine) labeled with 18F, which has a longer
half-life (T1/2 = 110 min), has been successfully evaluated as a PET im-
aging agent for mapping VMAT2 in the brain. Clinical studies have sug-
gested that [18F]FP-(+)-DTBZ PET imaging is useful because the
reduction in VMAT2 binding in the striatum is correlated with neuron
deficits, which is valuable in the diagnosis of PD [8–13]. Neurodegener-
ation of dopamine neurons can be readily detected prior to the expres-
sion of PD symptoms, allowing early diagnosis [14–16]. [18F]FP-(+)-
⁎
Corresponding author at: Five Eleven Pharma Inc., Philadelphia, PA 19104, USA.
E-mail address: kunghf@gmail.com (H.F. Kung).
https://doi.org/10.1016/j.nucmedbio.2019.07.002
0969-8051/© 2019 Elsevier Inc. All rights reserved.

R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
27
Fig. 1. Chemical structures of VMAT2 targeting agents: (+)-tetrabenazine (TBZ, Xenazine), (+)-dihydrotetrabenazine (DTBZ) and tetrabenazine-D6 (SD-809, Austedo).
DTBZ may be a useful biomarker for measuring dopaminergic degener-
ation in vivo and monitoring the severity of disease [17]. [18F]FP-(+)-
DTBZ shows a significant correlation between VMAT2 density and cog-
nitive performance in patients with Lewy body dementia (DLB) [18].
Although [18F]FP-(+)-DTBZ has demonstrated clinical utility and,
potentially, can be used as a clinical tool to facilitate the diagnosis of
PD, ten years after this agent was first tested in humans, it has not re-
ceived FDA approval. Currently, there is no potential replacement for
18F-labeled DTBZ for imaging VMAT2 binding sites in the brain. We de-
veloped a suitable “new chemical entity” – a deuterated DTBZ deriva-
tive, D6-[18F]FP-DTBZ ([18F]9), which was recently reported as an
alternative VAMT2 imaging agent [19]. This new agent, which has “free-
dom of operation”, can be commercially developed to assist in the diag-
nosis of PD by visualizing reductions in viable neurons in the basal
ganglia regions of the brain.
Deuterium (D) is a stable isotope of hydrogen (H), providing a novel
strategy for developing new drugs. The physical and chemical proper-
ties of deuterium are similar to those of hydrogen. Because carbon-
deuterium (C-D) bonds are generally stronger than corresponding
carbon hydrogen (C\\H) bonds, selective deuterium substitution may
lead to beneficial changes in the pharmacology and pharmacokinetics
of drugs by decreasing their rate of metabolism [20–23]. Recently, the
FDA has approved a deuterated derivative of TBZ, tetrabenazine-D6
(SD-809, Austedo, Fig. 1), for the treatment of Huntington's disease.
Tetrabenazine-D6, in which two methyl groups have been substituted
with deuterated methyl groups, provides improved drug efficacy [24].
Similarly, a deuterated derivative of D6-[18F]FP-(+)-DTBZ, [18F]9, has
been prepared as a VMAT2 PET imaging agent (Fig. 2) [19]. After iv in-
jection into rats, [18F]9 showed excellent regional brain signals for PET
imaging of VMAT2 and improved plasma stability as demonstrated by
a reduction in bone uptake due to defluorination [19].
main process used for daily routine preparations of this compound
[25–27]. In recent years, an increasing number of PET radiotracers
have been purified by SPE. [18F]FMZ ([18F]Flumazenil, ethyl 8-fluoro-
5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5-a][1,4]benzodiazepine-
3-carboxylate) [28], a well-known radioligand for assessing GABAA re-
ceptors, was synthesized based on a two-step SPE procedure. The SPE
purification method for
[
18F]FES ([18F]Fluoroestradiol, 16α-
fluoroestradiol-17β-sulfamate) [29], a radiotracer for diagnosing and
monitoring the treatment of breast cancer, using Oasis Wax 3cc and
Sep-Pak QMA light cartridges, afforded [18F]FES in 15% RCY. Therefore,
implementing radiolabeling and purification via SPE is a common tactic
for simplifying purifications and improving yields.
Molar activity, particularly for neuroimaging, has a dramatic impact
on PET imaging. Small amounts of nonradioactive chemicals (chemical
byproducts) do not affect the specific distribution and binding. In the
package inserts of FDA-approved neuroimaging agents, the recom-
mended chemical doses for [18F]AV-45 (florbetapir, Amyvid) and [18F]
AV-1 (florbetaben, Neuraceq) are “mass dose” no higher than 50 μg
and 30 μg, respectively [30]. An SPE preparation of [18F]FP-(+)-DTBZ
was reported previously [31]. The report indicated that there was a sig-
nificant amount of pseudocarrier (OH-derivative, 200–600 μg) remain-
ing in the final dose. Although PET imaging studies in normal and
unilaterally lesioned monkey brains showed no significant difference
in striatal localization between HPLC- and SPE-purified [18F]FP-(+)-
DTBZ. Ex vivo inhibition studies suggested that the highest dose
(3.5 mg/kg) of pseudocarrier (OH-derivative) via SPE purification
might inhibit total binding by 20% [32].
Previously, the preparation of D6-[18F]FP-(+)-DTBZ,
[
18F]9,
employed HPLC to remove the hydroxyl side product, which is pro-
duced after the hydrolysis of the O-tosylate (-OTs), 7, to the correspond-
ing hydroxyl derivative, 6 [19]. To reduce the preparation time and
simplify the production process, we report herein a streamlined SPE
process with a combination of ethanol-water as the eluent, and the ma-
jority of hydroxyl side product 6 and other related impurities were re-
moved from the product using the developed method. Herein, we
reduced the “mass dose” to no higher than 50 μg and confirmed that
the amount of nonradioactive chemicals will not affect the specific dis-
tribution at low activity. This simplified SPE method might produce
multiple doses of D6-[18F]FP-(+)-DTBZ, [18F]9, facilitating its wide-
spread clinical application in the diagnosis of PD.
One major issue of using D6-[18F]FP-(+)-DTBZ, [18F]9, for wide-
spread clinical application is that its preparation required the use of
semipreparative radio-HPLC. This relatively long and laborious process
may be accompanied by considerable losses in the radioactive product.
In the development of a method for the large-scale clinical use for [18F]
FDG, HPLC purification was replaced by solid-phase extraction (SPE).
Multiple sequential solid-phase cartridges were adopted for hydrolysis
of the protecting groups and the final purification. Since the 1990s,
this simplified process for the preparation of [18F]FDG has been the
Fig. 2. Chemical structures of VMAT2 imaging agents: [11C]-(+)-DTBZ and [18F]FP-(+)-DTBZ (AV-133) and D6-[18F]FP-(+)-DTBZ, [18F]9.

28
R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
2. Experimental section
4 mmol) and 10% dry Pd/C (30 mg) was stirred in tetrahydrofuran
(THF, 20 mL) and ethanol (EtOH, 20 mL) under H2 at room temperature
overnight. The reaction mixture was filtered and washed with EtOH
(30 mL) and THF (30 mL). The solvent was removed under vacuum to
give 1.18 g of compound 1 (yield 96%) as a yellow solid. 1H NMR
(400 MHz, CDCl3) δ 6.68 (s,1H), 6.67 (s, 1H), 3.87 (s, 3H), 3.44–3.38
(m, 1H), 3.16–2.97 (m, 4H), 2.66–2.56 (m, 2H), 2.49–2.42 (m, 1H),
1.99 (t, J = 2.01 Hz, 1H), 1.79–1.68 (m, 2H), 1.57–1.45 (m, 3H),
1.12–1.05 (m, 1H), 0.97–0.93 (m, 6H). HRMS calcd. for C18H27NO3 [M
+ H]+ 306.1991, found 306.2100.
2.1. General
All reagents were obtained commercially and used without further
purification unless indicated. All deuterated chemicals were purchased
from Toronto Research Chemicals Inc. Solvents were dried using a mo-
lecular sieve system (Pure Solve Solvent Purification System; Innovative
Technology, Inc.). The 1H NMR and 13C NMR spectra were recorded on
an Avance spectrometer at 400 and 100 MHz, respectively, and refer-
enced to NMR solvents as indicated. High-resolution mass spectrometry
(HRMS) data were obtained with an Agilent (Santa Clara, CA) G3250AA
LC/MSD TOF system. Generally, crude compounds were purified by flash
column chromatography (FC) over silica gel (Aldrich). All SPE cartridges
used were purchased from Waters Corporation, USA, and Macherey-
Nagel, Germany. Prior to use, Sep-Pak® Oasis HLB, Sep-Pak® tC2 Plus
Light, Sep-Pak® C8 Plus Short, Sep-Pak® tC18 Plus Light, Sep-Pak® CN
Plus Short, Chromafix C4, Chromafix C18 and Chromafix C18 Hydrawere
cartridges were activated by flushing with 5 mL of ethanol followed by
10 mL of water. Sep-Pak Accell Plus QMA Light cartridges were activated
with 10 mL of 0.5 М K2CO3 followed by 10 mL of water.
2.2.2. Allyl-d5-4-methylbenzenesulfonate, (2)
To a suspension of sodium hydride (60% in mineral oil, 7 mg,
0.26 mmol) and ethyl ether (Et2O 2 mL) under an argon atmosphere
was added prop-2-en-d5-1-ol (11 mg, 0.17 mmol) at room temperature.
After gas evolution ceased, the reaction mixture was cooled to 0 °C. p-
Toluenesulfonyl chloride (TsCl, 67 mg, 0.35 mmol) was dissolved in
diethyl ether (Et2O, 10 mL), and the resulting solution was slowly
added to the reaction mixture. After the addition was complete, the re-
action mixture was warmed to room temperature and allowed to stir for
3 h. The reaction mixture was added to a saturated solution of ammo-
nium chloride (NH4Cl, 10 mL) at 0 °C. Then, the mixture was extracted
with ethyl acetate (EtOAc, 10 mL × 3). The combined the organic layer
was dried over MgSO4 and filtered. The filtrate was concentrated, and
the residue was purified by FC (hexane/EtOAc = 85/15) to give
22.8 mg of compound 2 (yield: 62%) as a colorless oil. 1H NMR
(400 MHz, CDCl3) δ 7.81–7.79 (m, 2H), 7.36–7.34 (m, 2H), 2.45 (s, 3H).
2.2. Chemistry
Initially, we planned to investigate the radiolabeling of desired D6-
[
18F]FP-(+)-DTBZ, [18F]9, following a previously reported procedure in-
volving a direct SN2 fluorination reaction (Scheme 1). By carefully ex-
amining the reaction, displacement of -OTs by fluoride anion, it was
predicted that the reaction may lead to several side products. We rea-
soned that three side products would be produced by a. hydrolysis of
the -OTs group, leading to the hydroxy side product (compound 6); b.
elimination leading to the vinyl side product (compound 3); and c. the
nucleophilic substitution of the phenol DTBZ (cleavage of the carbon
chain from the phenolic hydroxy group at high temperature) by the
-OTs precursor leading to the dimer side product (compound 8). To
fully investigate these possible side products, we carefully performed
LC/MS analysis of the reaction mixture after the radiolabeling reaction,
and the LC/MS results clearly showed the three expected side products
listed above (Fig. S1A). All these side products were then synthesized
for use as standards for validation and generating standard calibration
curves for HPLC analysis (Scheme 2).
2.2.3.
(2R,3R,11bR)-9-((Allyl-d5)oxy)-3-isobutyl-10-methoxy-
1,3,4,6,7,11b-hexahydro-2H-pyrido[2,1-a]isoquinolin-2-ol, (3)
To a mixture of compound 1 (30 mg, 0.1 mmol) in acetone (10 mL)
was added compound 2 (22.8 mg, 0.1 mmol) followed by cesium car-
bonate (Cs2CO3, 130 mg, 0.4 mmol). The mixture was stirred at reflux
for 2 h. After cooling to room temperature, 10 mL of water was added,
and the mixture was extracted with ethyl acetate (EtOAc, 10 mL × 3).
The combined the organic layer was dried over MgSO4 and filtered.
The filtrate was concentrated, and the residue was purified by FC
(DCM/MeOH/NH3·H2O = 95/5/0.5) to give 11 mg of compound 3
(yield 30%) as a yellow solid. 1H NMR (400 MHz, CDCl3) δ 6.50 (s, 1H),
6.41 (s, 1H), 3.66 (s, 3H), 3.24–3.19 (m, 1H), 2.99–2.83 (m, 4H),
2.46–2.39 (m, 2H), 2.31–2.26 (m, 1H), 1.83–1.78 (m, 2H), 1.57–1.52
(m, 2H), 1.43–1.32 (m, 2H), 0.91–0.85 (m, 1H), 0.77–0.73 (m, 6H). 13
C
2.2.1. (2R,3R,11bR)-3-Isobutyl-10-methoxy-1,3,4,6,7,11b-hexahydro-2H-
pyrido[2,1-a]isoquinoline-2,9-diol, (1)
A mixture of (2R,3R,11bR)-9-(benzyloxy)-3-isobutyl-10-methoxy-
2,3,4,6,7,11b-hexahydro-2H-pyrido[2,1-a]isoquinolin-2-ol (1.59 g,
NMR (100 MHz, CDCl3) δ 147.87, 146.69, 129.95, 126.55, 113.82,
108.55, 74.85, 61.12, 60.26, 56.25, 52.14, 41.86, 40.75, 39.89, 29.33,
25.57, 24.37, 21.97. HRMS calcd. for C21H26D5NO3 [M + H]+ 351.2618
found 351.2706
Scheme 1. Chemical reaction for the preparation of D6-[18F]FP-(+)-DTBZ, [18F]9, and the three chemical impurities, 3, 6 and 8, observed after [18F]fluorination.

R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
29
Scheme 2. Synthesis of compounds 3, 6, 7 (precursor) and 8. a: Pd/C, H2, EtOH/THF, rt, overnight; b: NaH, TsCl, Et2O, 3 h, rt; c: Cs2CO3, acetone, reflux, 2 h; d: Et3N, DMAP, TsCl, DCM, rt,
overnight.
2.2.4. 3-Hydroxypropyl-1,1,2,2,3,3-d6 4-methylbenzenesulfonate, (4)
To a solution of propane-d6-1,3-diol (500 mg, 6.5 mmol) in dichloro-
methane (DCM, 10 mL) was added triethylamine (Et3N, 888 mg,
8.8 mmol) and 4-dimethylaminopyridine (50 mg, 0.44 mmol). After
the mixture was stirred at 0 °C for 10 min, a solution of p-
toluenesulfonyl chloride (TsCl, 840 mg, 4.4 mmol) in DCM (20 mL)
was added dropwise. After the addition was complete, the reaction mix-
ture was warmed to room temperature and allowed to stir overnight. To
the reaction mixture was added 15 mL of brine, and then it was ex-
tracted with DCM (10 mL × 3). The combined the organic layer was
dried over MgSO4 and filtered. The filtrate was concentrated, and the
residue was purified by FC (hexane/EtOAc = 50/50) to give 500 mg of
compound 4 (yield 50%) as a colorless oil. 1H NMR (400 MHz, CDCl3)
δ 7.80–7.79 (m, 2H), 7.37–7.35 (m, 2H), 2.45 (s, 3H). 13C NMR
(100 MHz, CDCl3) δ 145.09, 133.19, 130.11, 128.10, 21.87. HRMS calcd.
for C10H8D6O4S [M + H]+ 237.0989 found 237.1022.
dimethylaminopyridine (61 mg, 0.5 mmol), p-toluenesulfonyl chloride
(TsCl, 2.1 g, 11 mmol) and dichloromethane (DCM, 10 mL) following
the same procedure described for compound 4. FC with hexane/EtOAc
(60/40) gave 1.56 g of compound 5 (yield 90%) as a white solid. 1H
NMR (400 MHz, CDCl3) δ 7.78–7.76 (m, 4H), 7.38–7.36 (m, 4H), 2.48
(s, 6H). 13C NMR (100 MHz, CDCl3) δ 145.28, 132.85, 130.18, 128.10,
21.88. HRMS calcd. for C17H14D6O6S2 [M + H]+ 391.1078. found
391.1140.
2.2.6. (2R,3R,11bR)-9-(3-Hydroxypropoxy-1,1,2,2,3,3-d6)-3-isobutyl-10-
methoxy-1,3,4,6,7,11b-hexahydro-2H-pyrido[2,1-a]isoquinolin-2-ol, (6)
Compound 6 was prepared from compound 1 (50 mg, 0.16 mmol),
compound 4 (50 mg, 0.2 mmol), cesium carbonate (Cs2CO3, 208 mg,
0.64 mmol) and acetone (15 mL) following the same procedure de-
scribed for compound 3. The concentrated residue was purified by FC
(DCM/MeOH/NH3·H2O = 90/9/1) to give 30 mg of compound 6 (yield
52%) as a pale-yellow solid. 1H NMR (400 MHz, CDCl3) δ 6.69 (s, 1H),
6.63 (s, 1H), 3.82 (s,3H), 3.48–3.40 (m, 1H), 3.16–3.03 (m, 4H),
2.65–2.57 (m, 2H), 2.47 (t, J = 1.6 Hz, 1H), 2.02–1.97 (m, 1H),
1.71–1.75 (m, 2H), 1.63–1.49 (m, 2H), 1.07–1.12 (m, 1H), 0.95–0.97
2.2.5. Propane-1,3-diyl-d6 bis(4-methylbenzenesulfonate), (5)
Compound 5 was prepared from propane-d6-1,3-diol (0.36 g,
4.46 mmol), triethylamine (Et3N, 1.8 g, 18 mmol), 4-

30
R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
(m, 6H). 13C NMR (100 MHz, CDCl3) δ 147.70, 146.71, 130.02, 126.49,
113.59, 108.29, 74.55, 60.93, 60.06, 55.97, 51.83, 41.57, 40.52, 39.69,
29.07, 25.37, 24.11, 21.77. HRMS calcd. for C21H27D6NO4 [M + H]+
370.2786 found 370.3505.
The samples were analyzed by LC/MS with an Agilent LC/MSD TOF
instrument in positive ESI mode and an Agilent HPLC 1100 series sys-
tem. The HPLC conditions were as follows: stationary phase: Agilent
Poroshell 120 EC-C18 2.7 μ 3.0 × 50 mm; mobile phase: 0.4 mL/min:
UV detection: 280 nm; and ESI positive mode. The following gradient
was used: from 0 to 15 min, gradient 0.1% formic acid 100–0%, ACN
0–100%; and from 15 to 20 min, 100% ACN.
2.2.7. 3-(((2R,3R,11bR)-2-Hydroxy-3-isobutyl-10-methoxy-1,3,4,6,7,11b-
hexahydro-2H-pyrido[2,1-a]isoquinolin-9-yl)oxy)propyl-1,1,2,2,3,3-d6 4-
methylbenzenesulfonate, (7)
2.5. Solid-phase extraction (SPE) purification
Compound 7 was prepared from compound 1 (100 mg, 0.33 mmol),
compound 5 (140 mg, 0.35 mmol), cesium carbonate (Cs2CO3, 427 mg,
1.3 mmol) and acetone (30 mL) following the same procedure de-
scribed for compound 3. The concentrated residue was purified by FC
(DCM/MeOH/NH3·H2O = 95/5/0.5) to give 100 mg of compound 7
(yield 58%) as a pale-yellow solid. 1H NMR (400 MHz, CDCl3) δ
7.78–7.75 (m, 2H), 7.25–7.27 (m, 2H), 6.66 (s, 1H), 6.51 (s, 1H), 3.77
(s, 3H), 3.39–3.38 (m, 1H), 3.13–2.99 (m, 4H), 2.63–2.56 (m, 2H),
2.47–2.40 (m, 4H), 2.00–1.95 (m, 1H), 1.70–1.72 (m, 2H), 1.53–1.48
(m, 2H), 1.13–1.15 (m, 1H), 0.96–0.94 (m, 6H). 13C NMR (100 MHz,
CDCl3) δ 147.80, 146.80, 130.00, 126.61, 113.79, 108.76, 74.58, 60.91,
60.08, 56.15, 51.88, 41.62, 40.55, 39.70, 29.10, 25.36, 24.12, 21.78.
HRMS calcd. for C39H52D6N2O6 [M + H]+ 524.2875 found 524.2927.
The reaction mixture was cooled to room temperature, and 7 mL of
water was added. The mixture was loaded onto different cartridges
(eight different cartridges were tested: Sep-Pak Oasis, tC2, C8, tC18,
CN, Chromafix C18 Hydrox, Chromafix C18, and Chromfix C4). The
loaded cartridges were first washed with 10 mL of water and then
eluted with 2 mL of gradient concentrations of ethanol-water or
acetonitrile-water (10%, 20%, 30%, 40%, 50% and 60%). The resulting elu-
ates were analyzed by HPLC. To optimize the elution volume, the se-
lected cartridges were washed with water, followed by different
concentrations (25%, 30% or 35%) of ethanol-water in 2 mL portions.
The flow rate was approximately 10 mL/min.
2.6. Semipreparative HPLC purification method
2.2.8. (2R,2′R,3R,3′R,11bR,11b'R)-9,9′-((Propane-1,3-diyl-d6)bis(oxy))bis
(3-isobutyl-10-methoxy-1,3,4,6,7,11b-hexahydro-2H-pyrido[2,1-a]
isoquinolin-2-ol), (8)
The reaction mixture was cooled to room temperature, and 7 mL of
water was added to this mixture. The mixture was loaded onto an
Oasis HLB (3 cc) cartridge and then washed with 10 mL of water. [18F]
9 was eluted with 1 mL of acetonitrile. This solution was diluted with
1 mL of 10 mM AFB and injected into the HPLC system (Gemini 250
× 10 mm, ACN/10 mM AFB = 45/55, 3 mL/min). The eluate from 14.0
to 15.5 min was collected and diluted with 25 mL of water. This mixture
was loaded onto a tC18 cartridge, from which the desired product, [18F]
9, was eluted with 1 mL of ethanol.
The residue from the synthesis of compound 7 was purified by FC
(DCM/MeOH/NH3·H2O = 90/9/1) to give 5.8 mg of compound 8
(yield 5%) as a pale-yellow solid. 1H NMR (400 MHz, CDCl3) δ
6.67–6.34 (m, 2H), 3.81 (s,6H), 3.39–3.43 (m, 2H), 3.15–2.99 (m, 8H),
2.62–2.59 (m, 4H), 2.48–2.43 (m, 2H), 1.99 (t, J = 10.8, 2H), 1.73–1.63
(m, 4H), 1.63–1.46 (m, 6H), 1.46–1.03 (m, 2H), 0.97–0.93 (m, 12H).
13C NMR (100 MHz, CDCl3) δ 147.88, 147.14, 129.81, 126.67, 113.84,
108.67, 74.85, 61.10, 60.24, 56.29, 52.05, 41.78, 40.74, 39.90, 29.28,
25.55, 24.37, 21.98. HRMS calcd. for C39H52D6N2O6 [M + H]+
657.4671 found 657.4100.
2.7. In vitro binding assay for IC50 determination
Tissue homogenates of striatum (dissected from mouse brain) were
prepared in buffer 1 (50 mM of HEPES, 0.32 M sucrose, pH = 7.4). Com-
pounds were examined for their ability to compete with [18F]9 for bind-
ing at concentrations ranging from 10−6 to 10−11 M. The binding assays
were performed in glass tubes (12 × 75 mm) with final volumes of
0.25 mL. In the nonspecific binding assay, 50 μM ( )-tetrabenazine
(TBZ), was used as the positive control. After incubation for 1 h at
room temperature, the bound ligand was separated from the free ligand
by filtration through glass fiber filters. The filters were washed three
times with 4 mL of ice-cold PBS buffer at pH 7.4, and the radioactivity
of the material remaining on the filters was counted with a gamma
counter (WIZARD2, Perkin-Elmer). Data were analyzed using the non-
linear least-square curve fitting program LIGAND to determine IC50
values.
2.3. Nucleophilic [18F]fluorination of 9 under different conditions
A solution containing [18F]fluoride (37–74 MBq) was loaded onto an
activated QMA light cartridge and eluted with 1 mL of K222/K2CO3 solu-
tion (2 mg K2CO3, 10 mg K222 in 0.15 mL of water and 0.85 mL of aceto-
nitrile) into a glass test tube. The solution was dried under a flow of
argon at 90 °C and azeotropically dried with 2 mL of acetonitrile. The
precursor (1 mg) was dissolved in 1 mL of solvent (DMSO, DMF, DMA
or ACN) and added to the dried [18F]F−/K222/K2CO3 complex. The reac-
tion mixture was heated to different temperatures (70 °C, 75 °C, 90 °C,
110 °C or 130 °C). Aliquots (20 μL) were removed at different time
points (5, 10 and 15 min), and the aliquots were quenched in ice
water prior to analysis by HPLC.
2.4. Analysis of the radiochemical yield, radiochemical purity and amount
of chemicals by HPLC
2.8. In vivo biodistribution in mice
Five normal CD-1 mice per group were used for each biodistribution
study. The assay used 0.15 mL of solutions containing different concen-
trations of 3, 6, and 8. HPLC-purified [18F]9 in 0.1% bovine serum albu-
min was injected into the tail vein of CD-1 mice (33–37 g, male) while
they were under isoflurane anesthesia (1.11 MBq per mouse). The
mice (n = 5) were sacrificed by cervical dislocation at 30 min postinjec-
tion. The organs or tissues of interest were removed and weighed, and
radioactivity was counted with an automatic gamma counter. The per-
centage of the dose per organ was calculated by comparing the tissue
counts with suitable diluted aliquots of the injection material. Different
regions of the brain corresponding to the cerebellum (CB), striatum
(ST), hippocampus and cortex were dissected from the brains and
counted separately.
The radiochemical yield (RCY) and radiochemical purity (RCP) were
analyzed by HPLC with a gamma ray radio detector and a UV/Vis detec-
tor (Agilent 1200 series, Ascentis C18 column, 150 × 4.6 mm; 280 nm)
with the following mobile phase (1 mL/min) gradient. From 0 to 2 min,
isocratic 10 mM ammonium formate buffer (AFB) 95%, ACN 5%; from 2
to 5 min, gradient AFB 95–30%, ACN 5–70%; from 5 to 10 min, gradient
AFB 30–0%, ACN 70–100%; from 10 to 15 min, gradient AFB 0–95%, ACN
100–5%; and from 15 to 20 min, 95% AFB (retention times: [18F]fluo-
ride = 1.4 min, [18F]9 = 7.8 min). The amounts of compounds 3, 6, 7,
and 8 were quantified using this HPLC system; standard calibration
curves were plotted using authentic sample solutions at least 6 different
concentrations from 0.01 μg to 1 μg (Fig. S1B).

R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
31
Table 2
2.9. Ex vivo autoradiography
Results of the fluorination reaction for producing D6-[18F]FP-(+)-DTBZ, [18F]9, in DMSO at
different temperatures and times. The goal was to produce the highest RCY and the lowest
amount of O-tosylated precursor, 7. (Avg
Two normal CD-1 mice were injected with HPLC- or SPE-purified
18.5 MBq [18F]9 via the tail vein and sacrificed at 30 min postinjection.
The brain of each mouse was harvested and frozen. Coronal sections
20 μm thick were cut on a cryostat, thaw-mounted onto slides and
dried at room temperature. The dried sections were exposed to imaging
plates for 30 min, and images were acquired using a Typhoon FLA 7000
(GE Healthcare Life Sciences).
SD, n = 3).a
6 (μg)
Time
(min)
Temp.
(°C)
RCY (%)
8 (μg)
7 (μg)
(precursor)
5
70
70
70
70
90
52.0
52.7
56.3
49.3
71.3
69.0
68.7
79.3
79.0
78.7
88.3
89.7
81.0
14.7 383
123 50.1
15.4
7.48
2.56
9.08
22.1
44.4
25.3
18.1
6.72
11.2
23.5
17.1
17.2
230
82.2
51.7
20.2
97.7
12.9
0.00
0.00
0.00
0.00
0.00
0.00
0.00
206
10
15
20
5
10
15
5
10
15
5
10
15
12.3 415
15.9 465
14.0 503
15.0 364
21.7 354
19.9 411
4.90 305
3.90 354
3.50 370
5.80 358
6.60 357
7.50 395
69.4 49.6
57.4 56.7
19.3 58.1
48.7 109
112 120
28.6 135
25.6 175
20.1 186
7.10 187
29.4 205
29.3 195
60.4 196
142
89.5
35.1
132
22.3
0.00
0.00
0.00
0.00
0.00
0.00
0.00
90
90
2.10. In vitro autoradiography
110
110
110
130
130
130
All 20-μm-thick coronal sections of rats and monkeys were stored at
−20 °C until use in vitro autoradiography. Prior to the assays, the sec-
tions were thawed, dried at room temperature and preincubated for
20 min in an ice-cold incubator. The assay was then carried out in Coplin
jars in ice-cold incubation buffer 1 for 15 min at room temperature. [18F]
9 was diluted in buffer 1 to yield approximately 30,000 CPM/mL, and
compounds 3, 6 and 8 were diluted in buffer 1 to 5.7 nmol/mL,
2.7 nM/mL and 6.0 nM/mL, respectively. The brain sections were incu-
bated at room temperature for 30 min with 1 mL of [18F]9 solution
and then rinsed with PBS twice and water for 3 min. The sections
were dried and then exposed to imaging plates for 30 min.
a
Aliquots (10 μL) were removed from the reaction mixture at different time points (5,
10 and 15 min) were quenched in ice water. The RCYs and chemical impurities were de-
termined by HPLC.
OTs-precursor, 7. At a higher temperature, the amount of OTs- precursor
7 remaining was below the detection limit, which simplified the subse-
quent SPE purification. The amounts of all chemicals in the crude mix-
ture from the fluorination reaction were analyzed by HPLC compared
with the calibration curves of authentic samples (Fig. S1). Three major
chemical products were produced in the reaction mixture: the OH–
side product, 6; the OTs-precursor, 7; and the dimer side product, 8
3. Results and discussion
3.1. Optimization of the 18F fluorination conditions
The fluorination reaction for the preparation of [18F]FP-(+)-DTBZ,
[
[
18F]9, is an SN2 reaction with an O-tosyl leaving group in which the
18F]fluoride ion is activated by complexing with K+/K222. Three param-
eters of this fluorination were investigated to maximize the RCY: the
solvent, reaction time and reaction temperature. The reaction was per-
formed in 1 mL of DMF, DMA or DMSO solution containing 1 mg
(1.9 μmol) of compound 7 at 110 °C for 10 min. This labeling reaction
was also performed in 1 mL of ACN at 75 °C for 10 min in a closed vial
(due to the lower boiling point of ACN). As shown in Table 1, when
the labeling was performed in DMF, DMA, ACN and DMSO, the RCYs
were 50.7%, 14.3%, 73.0% and 79.0%, respectively. Fluorination in
DMSO at 110 °C for 10 min provided the highest RCY (79%); therefore,
subsequent reactions were all carried out in DMSO.
The effects of temperature and time in the production of D6-[18F]FP-
(+)-DTBZ, [18F]9, were evaluated in DMSO as the solvent. The RCYs im-
proved with increasing temperature; the RCYs in DMSO for 5 min were
52%, 71%, 79%, and 88% at 70 °C, 90 °C, 110 °C and 130 °C, respectively
(Table 2). The labeling yield did not improve significantly with a longer
heating time, suggesting that the SN2 fluorination reaction was com-
plete in 10 min. Additionally, the RCYs decreased with longer heating
times (15 min), suggesting that more side reactions occurred.
The time and temperature of the fluorination reaction in DMSO were
further optimized. Initially, the goal was to maximize the labeling yield
of D6-[18F]FP-(+)-DTBZ, [18F]9, and reduce or eliminate the remaining
Table 1
Fluorination reactions in different solvents at different temperatures for producing D6-
[
18F]FP-(+)-DTBZ, [18F]9. (Avg
SD, n = 3).
Solventa
Time (min)
Temp. (°C)
RCYb (%)
Fig. 3. Results of the initial SPE purifications using different cartridges to remove hydroxyl
side product 6 and retain the desired product, [18F]FP-(+)-DTBZ, [18F]9. Eight different
cartridges were tested (Sep-Pak Oasis, tC2, C8, tC18, CN, Chromafix C18 Hydrox,
Chromafix C18, and Chromfix C4). After loading the crude 18F-labeled reaction mixture
on the cartridges, they were eluted with 2 mL of incremental concentrations ethanol-
water and acetonitrile-water mixtures. The goal was to maximize the removal of
residual hydroxyl side product 6 and retain the desired product, D6-[18F]FP-(+)-DTBZ,
DMF
DMA
ACN
10
10
10
10
110
110
75
50.7
14.3
73.0
79.0
13.6
1.20
6.90
3.90
DMSO
110
a
DMA = N,N-dimethylacetamide; DMF = N,N-dimethylformamide;
ACN = acetonitrile; DMSO = dimethyl sulfoxide.
[
18F]9. The initial evaluation suggested that tC18 most effectively removed hydroxyl side
product 6 and retained the desired product (solid arrow indicates the best performance).
b
Determined by radio-HPLC analysis of the crude product.

32
R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
(Table 2). The residual OTs-precursor, 7, was completely hydrolyzed
and transformed into hydroxyl side product 6 and dimer side product
8 when the reaction was heated above 90 °C. More dimer side product
8 was produced at higher temperatures. This dimeric side product, 8,
derived from the fluorination of the O-tosyl-proproxy-phenyl group
has not been reported in the literature. It is likely that a similar dimeric
side product might have been produced in other similar reactions (the
fluorination of O-tosyl-proproxy-phenyl groups). Since this type of fluo-
rinated PET tracer is quite common in the field, one may wonder how
often this type of side product was produced but not investigated or re-
ported. By careful examination of the LC/MS data, another side product,
vinyl derivative 3, was found in the reaction mixture. The amount of
vinyl derivative 3 was within the detection range; however, the total
amount of 3 in the reaction mixture was extremely low (b 0.5 μg),
which is often below the detection limit. All of these side products
were authentically prepared and characterized. Based on the results
listed in Table 2, DMSO was selected as the optimal solvent, and the re-
action was heated at 130 °C for 10 min. The resulting crude reaction
mixture was optimized via SPE purification.
radiochemical purity, maximize the radiochemical yield and minimize
residual chemical impurities.
3.2.1. Gradient elution on different cartridges
Building on the reported SPE purification method for the preparation
[
18F]FP-(+)-DTBZ, we aimed to reduce the major side product, hy-
droxyl side product, 6. The choice of SPE cartridge was determined by
the difference in the lipophilicities of [18F]9 and hydroxyl side product
6. Therefore, we only selected and evaluated reversed-phase SPE car-
tridges. The fluorination was carried out using the optimized conditions
described above (DMSO as the solvent, heating at 130 °C for 10 min).
After cooling to room temperature, the crude reaction mixture (approx-
imately 1 mL) was diluted with 7 mL of water and loaded onto different
cartridges. Then, 2 mL portions of eluent (from 10% to 60% acetonitrile-
water or ethanol-water) were used to elute the products. The purpose
of this gradient elution was to find the suitable eluent concentration
range when the majority of 6 was removed while retaining the desired
product, [18F]FP-(+)-DTBZ, [18F]9, on the cartridge (Fig. 3; more de-
tailed graphs are shown in Fig. S2). When the amount of residual 6
was b10% (red dotted line), the losses in activity of [18F]9 were 26.4%,
20.1%, 30.6%, 33.9% and 47.8% in Sep-Pak tC2 acetonitrile-water elution,
Sep-Pak tC18 ethanol-water elution, Sep-Pak tC18 acetonitrile-water
elution, Chromafix C18 acetonitrile-water elution and Chromafix C4
ethanol-water elution, respectively. We focused on Sep-Pak tC18 with
ethanol-water elution and optimized the concentration and volume as
this system showed the smallest loss in activity (Fig. 3, solid red arrow
indicates the best combination).
3.2. SPE purification for the production of [18F]9
Currently, “HPLC-free” purifications utilizing disposable SPE car-
tridges are widely used in PET radiochemistry. An SPE purification of
[
18F]FP-(+)-DTBZ, [18F]9, has been reported [31]. However, the re-
ported SPE purification did not eliminate nonpolar chemical impurities,
such as the hydroxyl side product. The final dose contained 200–600 μg
of the pseudocarrier, 9-OH-DTBZ. Different cartridges and eluents were
tested in the optimization of the SPE purification of D6-[18F]FP-(+)-
DTBZ, [18F]9. The optimization had three objectives: achieve excellent
3.2.2. Gradient elution and volume tests with the tC18 cartridge
Since the tC18 cartridge most effectively removed hydroxyl side
product 6 and retained the desired product, [18F]9, additional
Fig. 4. Further refinement using an Sep-Pak tC18 cartridge to remove hydroxyl side product 6 and retain the desired product, [18F]9, by using different ratios of ethanol-water as the eluent.
A: Eluted with 2 mL portions of incremental concentrations of ethanol-water. B, C and D: Eluted with 2 mL portions of incremental concentrations of ethanol-water at 25%, 30% and 35%,
respectively.

R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
33
brain, different concentrations of impurities were added to [18F]9 (puri-
fied by HPLC), and the solutions were tested in normal mice. In these in-
jections, the average amounts of 6 and 8 in the SPE-purified product
were 35 μg and 5 μg (Table 3), and the amount of 3 was b0.5 μg. The
amounts of chemical impurities were estimated on a moles per kg
basis from human to mouse by the following formula: (Human equiva-
lent dose, HED, assumes that the average human weight is 70 kg and the
average weight of a mouse is 35 g) [33].
Table 3
Results of the optimized production of D6-[18F]FP-(+)-DTBZ, [18F]9, and the residual
amounts of chemical impurities per batch (Avg
SD, n = 9).a
Radiochemical yield, RCY (%)b
Radiochemical purity, RCP (%)
Residual amount of 6
Residual amount of 8
Residual amount of 3
38.7
10.5
N99%
34.7
4.36
12.8 μg
4.86 μg
b 0.5 μg
a
The optimal conditions for the SPE purification were loading the crude fluorination
reaction mixture on a Sep-Pak tC18 cartridge and eluting with 10 mL of 25% ethanol-water
followed by 2 mL of 50% ethanol-water to give the final product, D6-[18F]FP-(+)-DTBZ,
HED ðmg=kgÞ ¼ animal dose in mg=kg
ꢀ ðanimal weight in kg=human weight in kgÞ0:33
[
18F]9.
b
RCY (%) = activity of product/initial activity (%) (with decay correction).
Therefore, the amounts of 3, 6 and 8 for mice were 8.6 × 10−5 mg/kg,
6.1 × 10−3 mg/kg and 8.6 × 10−4 mg/kg, respectively. We prepared
mixtures with 10×, 100× and 1000× higher doses of 3, 6 and 8 and
added them to the HPLC-purified D6-[18F]FP-(+)-DTBZ, [18F]9. HPLC-
purified [18F]9 without added chemical impurities was used as the con-
trol dose. Whole body and brain distribution in mice were evaluated
using the doses listed above (Table 5). The results of the biodistribution
study showed that there was no difference between the control (no
added chemical impurities) and the 10× group. The 100× and 1000×
groups showed a significant reduction in brain uptake. Comparing the
10× group and the control, there was no significant difference in re-
gional brain distribution, i.e., striatum, cerebellum and cortex. Differ-
ences were observed between the 1000× group and the control in the
pancreas (p b 0.05) and brain distribution (cerebellum p b 0.05; stria-
tum, hippocampus, cortex and hypothalamus p b 0.01). It appears that
the 10-fold higher content of chemical impurities in the final product
did not influence the striatum uptake where VMAT2 binding sites are
optimization experiments were performed. Using ethanol-water mix-
tures ranging from 25%, 30% and 35% led to good results (Fig. 4A).
After multiple experiments, 10 mL of 25% ethanol-water eluent was
chosen for the first stage of the elution (solid red arrow, Fig. 4B).
Under these conditions, dimer side product 8 was retained on the car-
tridge due to its high lipophilicity. Using 2 mL of 50% ethanol-water
for the second stage of the elution produced the highest (89.1%) yield
of the desired product, D6-[18F]FP-(+)-DTBZ, [18F]9, accompanied by
34.7 12.8 μg of hydroxyl side product 6 and 4.36 4.86 μg of dimer
side product 8 (see Table 3).
The optimal conditions for SPE purification were finally selected:
loading the crude fluorination reaction mixture to a Sep-Pak tC18 car-
tridge and eluting with 10 mL of 25% ethanol-water and then with
2 mL of 50% ethanol-water to give the final product, D6-[18F]FP-(+)-
DTBZ, [18F]9. With these optimized conditions, the desired product,
D6-[18F]FP-(+)-DTBZ, [18F]9, was obtained in 35 min with an average
radiochemical yield, RCY, of 38.7 10.5% (decay corrected, Avg SD,
n = 9) and a radiochemical purity, RCP, of N99% (Table 3). There was
a significant reduction in the impurities during the SPE purification
(for the HPLC profiles before and after purification, see Fig. S3). This
SPE purification method took only 5 min and resulted in an activity
loss of 25.8 16.4%. The activity lost in the HPLC purification method
was 25.8 16.4% (with decay correction), and it was a 30 min process.
The activity losses with the SPE and HPLC purification methods were
similar, but the SPE method was fast and simple.
Table 4
Comparison of in vitro binding affinities (IC50, nM) of ( )-TBZ, FP-(+)-DTBZ (AV-133),
D6-FP-(+)-DTBZ, 9 and other impurities (3, 6 and 8) for vesicular monoamine transporter
2 (VMAT2).a
Compound
Structure
IC50 (nM)
(Avg
SD,
Currently, the preparation has only been carried out at a radioactiv-
ity level of 37–370 MBq in the laboratory. For clinical applications in
hospitals or radiopharmacies, the method should be automated in a
hot cell and conducted at a higher radioactivity level (N37 GBq; suffi-
cient for multiple patient doses).
n = 3)
(
)-Tetrabenazine
21.2
1.74
FP-(+)-DTBZ
(AV-133)
3.24
7.23
2.70
792
1.57
0.20
0.56
113
3.3. In vitro binding assay
Using [18F]9 as the “hot” ligand and homogenates of striatum (dis-
sected from mouse brain), the binding affinities of ( )-TBZ, FP-(+)-
DTBZ, (AV-133), hydroxyl side product 6, vinyl side product 3, dimer
D6-FP-(+)-DTBZ,
9
side product
8 and “cold” D6-FP-(+)-DTBZ 9 were measured
(Table 4). The results suggested that the binding affinities of deuterated
AV133 and nondeuterated AV133 were similar (IC50 = 7.23 0.20 and
3.24 1.57 nM, respectively). Both showed higher binding affinity than
3
6
(
)-TBZ. Among all the chemical impurities, 3 showed a comparably
high binding affinity, whereas 6 and 8 showed much lower binding af-
finities. The results suggested that the impurity profile was very favor-
able. The impurity with the highest binding affinity, 3 (IC50 = 2.70
0.56 nM), was not detectable below b0.5 μg per batch. The other impu-
rities, 6 and 8, which were present in higher quantities, showed
hundred- or thousand-fold lower binding affinities for VMAT2 binding
sites; therefore, they are unlikely to affect the binding of VAMT2 in vivo.
8
1341
62.0
3.4. Biodistribution study in normal mice
a
In vitro binding studies were performed using mouse striatum homogenates, and the
radio tracer was D6-[18F]FP-(+)-DTBZ ([18F]9) purified by HPLC (all chemical impurities
were removed).
To explore the potential effects of chemical impurities (compounds
3, 6 and 8) in the binding of [18F]9 (the final product) to VMAT2 in the

34
R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
Table 5
3.5. Ex vivo and in vitro autoradiography
Biodistribution of D6-[18F]FP-(+)-DTBZ, [18F]9, in normal mice at 30 min postinjection: a
comparison of different amounts of added chemical impurities.a (n = 5, Avg
SD).
Ex vivo and in vitro autoradiography both showed that there were
no differences between SPE- and HPLC-purified D6-[18F]FP-(+)-DTBZ,
0×
10×
100×
1000×
[
18F]9, and the intensity of binding in the striatum matched well with
Organ
Blood
Heart
Muscle
Lung
Kidney
Spleen
Pancreas
Liver
1.19
1.90
1.22
2.52
3.06
3.34
15.2
10.5
1.34
2.28
2.11
0.16
0.18
0.50
0.20
0.37
1.07
2.61
0.76
0.25
0.40
0.28
1.56
1.93
0.99
2.50
3.32
2.90
15.2
11.6
1.17
2.17
2.32
0.45
0.08
0.10
0.18
0.24
0.26
1.00
0.71
0.09
0.42
0.25
1.22
1.89
0.98
2.38
3.27
2.67
15.6
12.7
1.23
2.28
1.85
0.16
0.10
0.04
0.16
0.33
0.12
2.19
0.74
0.09
0.33
0.09
1.52
1.89
0.99
2.44
3.56
2.36
12.3
12.5
1.19
2.80
1.07
0.24
0.14
0.07
0.21
0.53
0.46
0.98
0.93
0.05
0.22
0.06
the known VMAT2 distribution in the brain (Fig. 5). The resulting im-
ages corroborated the regional brain tissue dissection data and showed
high tracer uptake and retention in regions of brain with higher VMAT2
densities. The competitive binding studies in autoradiography showed
that 3 at 5.7 nmol/mL will block [18F]9 binding. However, 6 and 8 (at
2.7 nmol/mL and 6.0 nmol/mL, respectively) will not block VMAT2
binding, most likely due to their lower binding affinities.
Skin
Bone
Brain
It is important to note that the major difference between HPLC- and
SPE-purified D6-[18F]FP-(+)-DTBZ, [18F]9, is the amount of residual
chemical impurities in the dose. HPLC purification will remove all of
the chemical impurities, including compounds 3, 6 and 8, while the
SPE will only remove the majority of these chemicals, and residual
chemical impurities remained (b50 μg of total chemical impurity per
production batch). It is important to note that the total chemical
amount listed in the package insert per dose of Amyvid ([18F]AV-45,
florbetapir F18) is b50 μg of total chemical [34]. Amyvid is manufactured
via an HPLC purification method and is commercially distributed. The
SPE preparation of D6-[18F]FP-(+)-DTBZ, [18F]9, as described in this
paper, clearly contained b50 μg of total chemical per dose. The contents
of chemical impurities after SPE purification were low; therefore, they
are unlikely to influence VMAT2 binding in vivo. The results of the
in vivo biodistribution study and autoradiography studies in normal
mice also suggested that the impurities at their expected low doses
did not affect the biodistribution in the brain, and no significant dif-
ferences in striatal localization were observed. These results support
the use of this SPE purification method to prepare D6-[18F]FP-(+)-
DTBZ, [18F]9. Future implementation of this method for the routine
preparation of multiple doses will facilitate the diagnosis of PD
based on changes in VMAT2 binding sites in the basal ganglia regions
of the brain.
Regional brain distribution
Cerebellum
Striatum
Hippocampus
Cortex
Hypothalamus
Remainder
1.16
5.19
2.07
1.42
3.59
1.87
0.14
0.83
0.45
0.19
0.42
0.23
1.30
5.33
2.23
1.57
3.81
2.09
0.09
0.75
0.16
0.16
0.35
0.22
1.24
4.76
1.49
1.30
2.51
1.59
0.10
0.59
0.13
0.15
0.35
0.06
0.94
1.66
0.98
0.93
1.13
1.03
0.08
0.10
0.09
0.07
0.07
0.07
Ratio (vs. CB)
Striatum
Hippocampus
Cortex
Hypothalamus
Remainder
4.53
1.78
1.23
3.09
1.62
0.91
0.29
0.11
0.23
0.07
4.10
1.72
1.21
2.94
1.61
0.45
0.07
0.05
0.14
0.08
3.85
1.20
1.05
2.02
1.29
0.55
0.04
0.07
0.20
0.08
1.77
1.03
0.99
1.19
1.09
0.22
0.08
0.09
0.05
0.05
a
The concentrations of the impurities added for the 0×, 10×, 100× and 1000× groups
were listed proportionally: 0 mg/kg, 0.071 mg/kg (compound 3, 8.6 × 10−4 mg/kg, com-
pound 6, 6.1 × 10−2 mg/kg and compound 8, 8.6 × 10−3 mg/kg), 0.71 mg/kg and
7.1 mg/kg, respectively.
most abundant. The low levels of impurities of the purified product will
likely show no effect on the in vivo binding to VMAT2 target binding
sites in humans.
Fig. 5. Autoradiography results. A: Ex vivo autoradiography of mouse brain sections using SPE- and HPLC-purified D6-[18F]FP-(+)-DTBZ, [18F]9. Normal CD-1 mice were injected with
19 MBq of SPE- or HPLC-purified [18F]9. B: In vitro autoradiography of rat brain sections; VMAT2 binding sites were blocked with 3, 6 and 8 (the concentrations of competitive
compounds 3, 6 and 8 are 5.7 nmol/mL, 2.7 nmol/mL and 6.0 nmol/mL, respectively). Sections were incubated at room temperature for 30 min with 4.5 MBq of D6-[18F]FP-(+)-DTBZ,
[
18F]9, in 20 mL of buffer solution. C and D: In vitro autoradiography of rat and monkey brain sections using SPE- or HPLC-purified D6-[18F]FP-(+)-DTBZ, [18F]9. Same procedure with
B. HY: Hypothalamic nucleus, SN: substantia nigra.

R. Zhao et al. / Nuclear Medicine and Biology 72–73 (2019) 26–35
35
[7] Avendaño-Estrada A, Ávila-Rodríguez MA. Reference tissue models in the assess-
ment of 11C-DTBZ binding to the VMAT2 in rat striatum: a test-retest reproducibil-
ity study. Synapse 2018;72:e22029.
[8] Goswami R, Ponde D, Kung M, Hou C, Kilbourn M, Kung H. Fluoroalkyl derivatives of
dihydrotetrabenazine as positron emission tomography imaging agents targeting
vesicular monoamine transporters. Nucl Med Biol 2006;33:685–94.
[9] Kilbourn M, Hockley B, Lee L, Hou C, Goswami R, Ponde D, et al. Pharmacokinetics of
[(18)F]fluoroalkyl derivatives of dihydrotetrabenazine in rat and monkey brain.
Nucl Med Biol 2007;34:233–7.
[10] Kung M, Hou C, Goswami R, Ponde D, Kilbourn M, Kung H. Characterization of opti-
cally resolved 9-fluoropropyl-dihydrotetrabenazine as a potential PET imaging
agent targeting vesicular monoamine transporters. Nucl Med Biol 2007;34:239–46.
[11] Okamura N, Villemagne V, Drago J, Pejoska S, Dhamija R, Mulligan R, et al. In vivo
measurement of vesicular monoamine transporter type 2 density in Parkinson dis-
ease with 18F-AV-133. J Nucl Med 2010;51:223–8.
4. Conclusion
An improved preparation of D6-[18F]FP-(+)-DTBZ, [18F]9, using op-
timal reaction conditions and solid-phase extraction (SPE) was demon-
strated. The desired product was prepared in 35 min in 38.9% RCY
(decay corrected) and a high RCP (N99%). This SPE purification method
is well suited to automatic synthesis for routine clinical applications.
This new VMAT2 imaging agent, D6-[18F]FP-(+)-DTBZ, [18F]9, in con-
junction with PET imaging, might provide a routine clinical tool for im-
proving the diagnosis of Parkinson's disease (PD).
[12] Villemagne VL, Okamura N, Pejoska S, Drago J, Mulligan RS, Chetelat G, et al. In vivo
assessment of vesicular monoamine transporter type 2 in dementia with lewy bod-
ies and Alzheimer disease. Arch Neurol 2011;68:905–12.
[13] Alexander PK, Lie Y, Jones G, Sivaratnam C, Bozinovski S, Mulligan RS, et al. Manage-
ment impact of imaging brain vesicular monoamine transporter type-2 (VMAT2) in
clinically uncertain Parkinsonian syndrome (CUPS) with 18F-AV133 and PET. J Nucl
Med 2017;58:1815–20.
Abbreviations
AFB
ammonium formate buffer
DTBZ
FDG
FES
dihydrotetrabenazine
2-fluoro-2-dexoy-D-glucose
16α-fluoroestradiol-17β-sulfamate
ethyl 8-fluoro-5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5-
a][1,4]benzodiazepine-3-carboxylate
[14] Stoessl AJ, Martin WW, McKeown MJ, Sossi V. Advances in imaging in Parkinson's
FMZ
disease. Lancet Neurol 2011;10:987–1001.
[15] de la Fuente-Fernandez R, Schulzer M, Kuramoto L, Cragg J, Ramachandiran N, Au
WL, et al. Age-specific progression of nigrostriatal dysfunction in Parkinson's dis-
ease. Ann Neurol 2011;69:803–10.
[16] Strafella AP, Bohnen NI, Perlmutter JS, Eidelberg D, Pavese N, Van Eimeren T, et al.
Molecular imaging to track Parkinson's disease and atypical parkinsonisms: new im-
aging frontiers. Mov Disord 2017;32:181–92.
[17] Hsiao IT, Weng YH, Hsieh CJ, Lin WY, Wey SP, Kung MP, et al. Correlation of
Parkinson disease severity and 18F-DTBZ positron emission tomography. JAMA
Neurol 2014;71:758–66.
[18] Siderowf A, Pontecorvo MJ, Shill HA, Mintun MA, Arora A, Joshi AD, et al. PET imag-
ing of amyloid with Florbetapir F 18 and PET imaging of dopamine degeneration
with 18F-AV-133 (florbenazine) in patients with Alzheimer's disease and Lewy
body disorders. BMC Neurol 2014;14:79.
FP-(+)-DTBZ 9-fluoropropyl-(+)-dihydrotetrabenazine
HED
HEPES
HPLC
HRMS
LC/MS
PD
human equivalent dose
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
high-performance liquid chromatography
high-resolution mass spectrometry
liquid chromatography/mass spectrometry
Parkinson's disease
PET
RCP
positron emission tomography
radiochemical purity
[19] Liu F, Choi SR, Zha Z, Ploessl K, Zhu L, Kung HF. Deuterated (18)F-9-O-hexadeutero-
3-fluoropropoxyl-(+)-dihydrotetrabenazine (D6-FP-(+)-DTBZ): a vesicular mono-
amine transporter 2 (VMAT2) imaging agent. Nucl Med Biol 2018;57:42–9.
[20] Meanwell NA. Synopsis of some recent tactical application of bioisosteres in drug de-
sign. J Med Chem 2011;54:2529–91.
RCY
radiochemical yield
SPE
solid-phase extraction
TBZ
tetrabenazine
VMAT2 vesicular monoamine transporter 2
[21] Gant TG. Using deuterium in drug discovery: leaving the label in the drug. J Med
Chem 2014;57:3595–611.
[22] Kuchar M, Mamat C. Methods to increase the metabolic stability of (18)F-radio-
tracers. Molecules 2015;20:16186–220.
[23] Pirali T, Serafini M, Cargnin S, Genazzani AA. Applications of deuterium in medicinal
chemistry. J Med Chem 2019;62(11):5276–97.
Acknowledgments
[24] Frank S, Testa CM, Stamler D, Kayson E, Davis C, Edmondson MC, et al. Effect of
deutetrabenazine on chorea among patients with Huntington disease: a random-
ized clinical trial. JAMA 2016;316:40–50.
This work was supported in part by Five Eleven Pharma Inc. and the
Key Field Research and Development Program of Guangdong Province
(2018B030337001).
[25] Ido T, Wan CN, Fowler JS, Wolf AP. Fluorination with F2: a convenient synthesis of 2-
deoxy-2-fluoro-D-glucose. J Org Chem 1977;42:2341.
[26] Kuge Y, Nishijima K, Nagatsu K, Seki K, Ohkura K, Tanaka A, et al. Chemical impuri-
ties in [18F]FDG preparations produced by solid-phase 18F-fluorination. Nucl Med
Biol 2002;29:275–9.
Appendix A. Supplementary data
[27] Nakao R, Ito T, Yamaguchi M, Suzuki K. Simultaneous analysis of FDG, ClDG and
Kryptofix 2.2.2 in [18F]FDG preparation by high-performance liquid chromatogra-
phy with UV detection. Nucl Med Biol 2008;35:239–44.
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.nucmedbio.2019.07.002.
[28] Vaulina D, Nasirzadeh M, Gomzina N. Automated radiosynthesis and purification of
[(18)F]flumazenil with solid phase extraction. Appl Radiat Isot 2018;135:110–4.
[29] Fedorova O, Nikolaeva V, Krasikova R. Automated SPE-based synthesis of 16α-[18F]
fluoroestradiol without HPLC purification step. Appl Radiat Isot 2018;141:57–63.
[30] Trembath L, Newell M, Devous MD. Technical considerations in brain amyloid PET
imaging with 18F-florbetapir. J Nucl Med Technol 2015;43:175–84.
[31] Zhu L, Liu Y, Plossl K, Lieberman B, Liu J, Kung H. An improved radiosynthesis of
[(18)F]AV-133: a PET imaging agent for vesicular monoamine transporter 2. Nucl
Med Biol 2010;37:133–41.
[32] Zhu L, Qiao H, Lieberman BP, Wu J, Liu Y, Pan Z, et al. Imaging of VMAT2 binding sites
in the brain by (18)F-AV-133: the effect of a pseudo-carrier. Nucl Med Biol 2012;39:
897–904.
[33] Food and Administration D. Guidance for industry: estimating the maximum safe
starting dose in initial clinical trials for therapeutics in adult healthy volunteers. Cen-
ter for Drug Evaluation and Research (CDER); 2005; 7.
References
[1] Kilbourn MR. In vivo radiotracers for vesicular neurotransmitter transporters. Nucl
Med Biol 1997;24:615–9.
[2] Albin R, Koeppe R. Rapid loss of striatal VMAT2 binding associated with onset of
Lewy body dementia. Mov Disord 2006;21:287–8.
[3] Hayden MR, Leavitt BR, Yasothan U, Kirkpatrick P. Tetrabenazine. Nat Rev Drug
Discov 2009;8:17.
[4] Brossi A, Lidlar H, Walter M, Schnider O. Syntheseversuche in der Emetin-Reihe. 2-
Oxo-hydrobenzo[a]chinoline. Helv Chim Acta 1958;41:119–39.
[5] Frey KA, Koeppe RA, Kilbourn MR, Vander Borght TM, Albin RL, Gilman S, et al. Pre-
synaptic monoaminergic vesicles in Parkinson's disease and normal aging. Ann
Neurol 1996;40:873–84.
[6] Albin R, Koeppe R, Wernette K, Zhuang W, Nichols T, Kilbourn M, et al. Striatal [11C]
dihydrotetrabenazine and [11C]methylphenidate binding in Tourette syndrome.
Neurology 2009;72:1390–6.
[34] FDA. Amyvid (Florbetapir F18 injection); 2012.