Synthesis and evaluation of two new candidate high-affinity full agonist PET radioligands for imaging 5-HT1B receptors

Article abstract of DOI:10.1016/j.nucmedbio.2019.01.005

Introduction: The serotonin 1B receptor subtype is of interest in the pathophysiology and treatment of depression, anxiety, and migraine. Over recent years 5-HT_1B receptor binding in human brain has been examined with PET using radioligands that are partial but not full agonists. To explore how the intrinsic activity of a PET radioligand may affect imaging performance, two high-affinity full 5-HT_1B receptor agonists (AZ11136118, 4; and AZ11895987, 5) were selected from a large compound library and radiolabeled for PET examination in non-human primates. Methods: [¹¹C]4 was obtained through Pd(0)-mediated insertion of [¹¹C]carbon monoxide between prepared iodoarene and homochiral amine precursors. [¹¹C]5 was obtained through N-¹¹C-methylation of N-desmethyl precursor 6 with [¹¹C]methyl triflate. [¹¹C]4 and [¹¹C]5 were studied with PET in rhesus or cynomolgus monkey. [¹¹C]4 was studied with PET in mice and rats to measure brain uptake and specific binding. Ex-vivo experiments in rats were performed to identify whether there were radiometabolites in brain. Physiochemical parameters for [¹¹C]4 (pKa, logD and conformational energetics) were evaluated. Results: Both [¹¹C]4 and [¹¹C]5 were successfully produced in high radiochemical purity and in adequate amounts for PET experiments. After intravenous injection of [¹¹C]4, brain radioactivity peaked at a low level (0.2 SUV). Pretreatment with tariquidar, an inhibitor of the brain P-gp efflux transporter, increased brain exposure four-fold whereas pretreatment with a high pharmacological dose of the 5-HT_1B antagonist, AR-A000002, had no effect on the binding. Ex-vivo experiments in rats showed no radiometabolites entering brain. [¹¹C]5 also failed to enter monkey brain under baseline conditions. Conclusions: [¹¹C]4 and [¹¹C]5 show too low brain uptake and specific binding to be useful PET radioligands. Low brain uptake is partly ascribed to efflux transporter action as well as unfavorable conformations.

Full text of DOI:10.1016/j.nucmedbio.2019.01.005

Accepted Manuscript
Evaluation of agonist 5-HT1B PET radioligands
Anton Lindberg, Shuiyu Lu, Sangram Nag, Magnus Schou, Jeih-
San Liow, Sami S. Zoghbi, Michael P. Frankland, Robert L.
Gladding, Cheryl L. Morse, Akihiro Takano, Nahid Amini,
Charles S. Elmore, Yong Sok Lee, Robert B. Innis, Christer
Halldin, Victor W. Pike
PII:
S0969-8051(18)30349-4
DOI:
Reference:
https://doi.org/10.1016/j.nucmedbio.2019.01.005
NMB 8052
To appear in:
Nuclear Medicine and Biology
Received date:
Revised date:
Accepted date:
28 November 2018
17 January 2019
21 January 2019
Please cite this article as: A. Lindberg, S. Lu, S. Nag, et al., Evaluation of agonist
5-HT1B PET radioligands, Nuclear Medicine and Biology, https://doi.org/10.1016/
j.nucmedbio.2019.01.005
This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.

ACCEPTED MANUSCRIPT
Synthesis and evaluation of two new candidate high-affinity full agonist PET
radioligands for imaging 5-HT1B receptors
Evaluation of agonist 5-HT1B PET radioligands
Anton Lindberg[a,b]*, Shuiyu Lu[b], Sangram Nag[a], Magnus Schou[a,c], Jeih-San Liow[b], Sami S.
Zoghbi[b], Michael P. Frankland[b], Robert L. Gladding[b], Cheryl L. Morse[b], Akihiro Takano[a],
Nahid Amini[a], Charles S. Elmore[d], Yong Sok Lee[e], Robert B. Innis[b], Christer Halldin[a],
Victor W. Pike[b]
[a] Department of Clinical Neuroscience, Center for Psychiatry Research, Karolinska Institutet and
Stockholm County Council, SE-17176 Stockholm, Sweden.
[b] Molecular Imaging Branch, National Institute of Mental Health, National Institutes of Health,
Bethesda, Maryland 20892-1003, USA.
[c] PET Science Centre, Precision Medicine and Genomics, IMED Biotech Unit, AstraZeneca, SE-17176
Stockholm, Sweden.
[d] Isotope Chemistry, Early Chemical Development, Pharmaceutical Sciences, IMED Biotech Unit,
AstraZeneca, SE-43250 Göteborg, Sweden.
[e]
Center for Molecular Modeling, Center for Information Technology, National Institutes of Health,
Bethesda, Maryland 20892-5624, USA.
*anton.lindberg@ki.se
Keywords: Serotonin subtype 1B receptor, Radioligand, Agonist, Carbon-11, Carbonylation
1

ACCEPTED MANUSCRIPT
Abstract
Introduction: The serotonin 1B receptor subtype is of interest in the pathophysiology and treatment
of depression, anxiety, and migraine. Over recent years 5-HT1B receptor binding in human brain has
been examined with PET using radioligands that are partial but not full agonists. To explore how the
intrinsic activity of a PET radioligand may affect imaging performance, two high-affinity full 5-HT1B
receptor agonists (AZ11136118, 4; and AZ11895987, 5) were selected from a large compound library
and radiolabeled for PET examination in non-human primates.
Methods: [11C]4 was obtained through Pd(0)‐mediated insertion of [11C]carbon monoxide between
prepared iodoarene and homochiral amine precursors. [11C]5 was obtained through N-11C-
methylation of N-desmethyl precursor 6 with [11C]methyl triflate. [11C]4 and [11C]5 were studied with
PET in rhesus or cynomolgus monkey. [11C]4 was studied with PET in mice and rats to measure brain
uptake and specific binding. Ex-vivo experiments in rats were performed to identify whether there
were radiometabolites in brain. Physiochemical parameters for [11C]4 (pKa, logD and conformational
energetics) were evaluated.
Results: Both [11C]4 and [11C]5 were successfully produced in high radiochemical purity and in
adequate amounts for PET experiments. After intravenous injection of [11C]4, brain radioactivity
peaked at a low level (0.2 SUV). Pretreatment with tariquidar, an inhibitor of the brain P-gp efflux
transporter, increased brain exposure four-fold whereas pretreatment with a high pharmacological
dose of the 5-HT1B antagonist, AR-A000002, had no effect on the binding. Ex-vivo experiments in rats
showed no radiometabolites entering brain. [11C]5 also failed to enter monkey brain under baseline
conditions.
Conclusions: [11C]4 and [11C]5 show too low brain uptake and specific binding to be useful PET
radioligands. Low brain uptake is partly ascribed to efflux transporter action as well as unfavorable
conformations.
2

ACCEPTED MANUSCRIPT
Introduction
The brain serotonin subtype 1B (5-HT1B) receptor is implicated in depression,
anxiety, and migraine.[1] Therefore, there is interest to study this receptor in living subjects
with positron emission tomography (PET) to gain insight into neuropsychiatric disorders.
PET with suitable radioligands might also assist in developing drugs targeting the 5-HT1B
receptor.[2] Effective radioligands, labeled with positron-emitting carbon-11 (t1/2 = 20.4
min) or fluorine-18 (t1/2 = 109.8 min), are therefore required for imaging and quantifying
these receptors in living human brain with PET.[3]
The 5-HT1B receptor is a G-protein coupled receptor (GPCR) that is expressed
on axons for serotonin throughout the central nervous system (CNS).[4] Receptor
concentrations are high in occipital cortex and globus pallidus, and negligible in cerebellum.
The G-protein is negatively coupled to adenylate cyclase such that when the 5-HT1B receptor
is activated the cyclase activity is depressed. 5-HT1B receptors serve as presynaptic
autoreceptors and postsynaptic heteroreceptors.[5] 5-HT1B autoreceptors serve to control
release of serotonin into the synaptic cleft. Agonists inhibit release of serotonin whereas
antagonists promote release. 5-HT1B heteroreceptors play a regulatory role in the release of
other neurotransmitters, like dopamine.[6, 7]
According to a long-standing two-state model, GPCRs generally exist in either
of two affinity states for endogenous ligand, namely G-protein-coupled and G-protein-
uncoupled, with these two states in dynamic equilibrium. How this equilibrium affects the
binding of exogenous agonists or antagonists to GPCRs in vivo is not precisely known.[8]
According to the two-state model, the binding of 5-HT1B receptor agonists is expected to be
more sensitive to G-protein coupling than that of antagonists, with agonists likely binding
only to the high affinity state and antagonists to both states.[9] However, an increasing
3

ACCEPTED MANUSCRIPT
body of evidence indicates that this model is insufficient, and that multiple conformational
states of GPCRs can be involved in the binding of ligands.[10, 11]
The few effective PET radioligands ([11C]1−[11C]3) that have been reported so
far for imaging 5-HT1B receptors are now considered to be full or partial antagonists (Fig.
1).[12-14] The influence of the intrinsic activity of a PET radioligand on 5-HT1B receptor
imaging performance has not been explored and therefore is not well understood.[15] An
agonist is expected to have selectivity for binding to the G-protein coupled state of the 5-
HT1B receptor and to be more useful for studying physiologically or pharamacologically
induced changes in synaptic serotonin levels.
In the present study, we aimed to develop a full agonist PET radioligand for
imaging brain 5-HT1B receptors to compare with previously reported radioligands that are
not full antagonists for their relative sensitivities to changes in synaptic serotonin levels.[16]
From an AstraZeneca library of over 3000 high-affinity 5-HT1B receptor ligands, we identified
AZ11136118 (Ki = 0.25 nM, EC50 < 0.17 nM, agonist GTPS efficacy 112%; 4) and AZ11895987
(Ki = 0.46 nM, EC50 < 0.98 nM, agonist GTPS efficacy 113%; 5) as high-affinity full agonists
with other favorable properties for PET radioligand development, including amenability for
labeling with a positron-emitter and properties expected to be conducive to good brain
entry (4, clogP 3.0, PSA 81 Å2; 5, clogP 3.1, PSA 45 Å2).[17] Here we report the syntheses of
[11C]4 and [11C]5 (Fig. 1) and their evaluations as candidate PET radioligands in rodents and
monkey.
Materials and Methods
General
4

ACCEPTED MANUSCRIPT
All work in this study involving [11C]4 was performed at the Molecular Imaging Branch,
NIMH, Bethesda, MD, USA.
All work in this study involving [11C]5 was performed at the PET centre, Karolinska Institutet,
Stockholm, Sweden.
1H (400.13 MHz) and 13C (100.62 MHz) NMR spectra were recorded on an Avance 400
instrument (Bruker) at room temperature in deuterated solvents. NMR signals are reported
as δ (ppm) downfield from the signal of tetramethylsilane. HRMS data (ESI-TOF) were
obtained at the Bioorganic Chemistry Laboratory of NIDDK (NIH). Optical rotations were
measured on a P-1010 polarimeter (JASCO).
Chemistry
N-Benzyl-8-methoxy-1,2,3,4-tetrahydronaphthalen-2-amine (8). 8-Methoxy-tetralone (7,
2.5 g, 14 mmol) was taken up in methanol (100 mL) at 0 C and benzylamine (1.5 g, 14 mmol)
and acetic acid (1.5 mL) were added. The mixture was stirred on ice for 30 min before
addition of sodium cyanoborohydride (1.3 g, 21 mmol) and then allowed to reach ambient
temperature overnight. Water (50 mL) was added and crude product was extracted into
ethyl acetate (100 mL × 3) and dried (Na2SO4). Product was isolated with flash column
chromatography (silica gel; 0−10% methanol in DCM). Desired fractions were pooled and
1
concentrated in vacuo to give 8 as a white solid (3.2 g, 12 mmol, 85%). H NMR (400 MHz,
CDCl3) δ 7.53 (d, J= 7.4 Hz, 2H), 7.34 (m, 3H), 7.10 (t, J = 7.9 Hz, 1H), 6.65 (t, J = 7.3 Hz, 2H),
4.23 (d, J = 13.2 Hz, 1H), 4.12 (d, J = 13.2 Hz, 1H), 3.75 (s, 3H), 3.26 (m, 2H), 2.77 (m, 3H), 2.27
(t, J = 6.1 Hz, 1H), 1.88 (m, 1H). 13C NMR (101 MHz, CDCl3) δ 157.1, 135.8, 130.2, 129.8,
129.6, 129.3, 127.2, 120.7, 107.3, 55.2, 53.9, 48.4, 27.8, 25.9, 25.7, 18.3. HRMS (TOF ES+)
m/z [M+H]+ calculated for C18H22NO 268.1701, found 268.1706.
5

ACCEPTED MANUSCRIPT
2-Isopropyl-5-methylcyclohexyl benzyl(8-methoxy-1,2,3,4-tetrahydronaphthalen-2-
yl)carbamate (9). Amine 8 (1.3 g, 4.8 mmol) was taken up in acetonitrile (10 mL). TEA (1.0
mL) and R-menthyl chloroformate (1.1 g, 5.0 mmol) were added sequentially. The reaction
mixture was stirred for 2 h at ambient temperature and then concentrated in vacuo. The
product was isolated with flash column chromatography (silica gel; 10−50% ethyl acetate in
hexane). The desired fractions were pooled and concentrated to give 9 as a thick oil (2.0 g,
1
4.4 mmol, 92%). H NMR (400 MHz, CDCl3) δ 7.27 (m, 5H), 7.08 (t, J = 7.8 Hz, 1H), 6.66 (dd, J
= 8.1, 11.4 Hz, 2H), 4.39 (t, J= 86.0 Hz, 4H), 3.79 (s, 3H), 3.01 (s, 1H), 2.82 (s, 1H), 2.61 (m,
1H), 2.10 (d, J = 10.6 Hz, 1H), 1.80 (q, J = 9.3 Hz, 2H), 1.61 (d, 3H), 1.48 (s, 1H), 1.14 (m, 3H),
0.89 (d, J = 6.1 Hz, 6H), 0.80 (s, 7H). 13C NMR (101 MHz, CDCl3) δ 157.4, 140.1, 137.0, 128.3,
126.7, 126.3, 124.4, 120.8, 106.9, 75.2, 55.2, 47.2, 46.8, 41.5, 41.4, 34.3, 31.4, 29.9, 26.3,
23.4, 23.1, 22.1, 20.9, 16.3, 16.1. HRMS (TOF ES+) m/z [M+H]+ calculated for C29H40NO3
450.3008, found 450.3012.
2-Isopropyl-5-methylcyclohexyl benzyl(8-hydroxy-1,2,3,4-tetrahydronaphthalen-2-
yl)carbamate (10). 9 (0.5 g, 1.1 mmol) was taken up in anhydrous DCM (5 mL) under argon
and cooled to between −30 C and −40 °C. BBr3 (1.0 M; 1.65 mL) was added. After 1 h, the
mixture was allowed to reach ambient temperature and then stirred for another 2 h. The
mixture was poured onto ice-water (100 mL) and then extracted with DCM (20 mL × 3). The
organic layers were combined, dried (Na2SO4), and concentrated in vacuo. Product was
isolated by flash column chromatography (silica gel 0−30% ethyl acetate in hexane). The
desired fractions were pooled and concentrated to give 10 as a thick oil (0.4 g, 0.92 mmol,
1
83%). H NMR (400 MHz, CDCl3) δ 7.27 (m, 5H), 6.91 (s, 1H), 6.61 (d, J = 7.6 Hz, 2H), 4.50 (m,
4H), 3.01 (s, 1H), 2.75 (s, 2H), 2.62 (t, J = 14.0 Hz, 1H), 2.11 (t, J = 5.5 Hz, 1H), 1.75 (m, 1H),
1.64 (d, J = 12.2 Hz, 3H), 1.47 (d, J = 3.1 Hz, 1H), 0.99 (m, 2H), 0.89 (d, J = 6.2 Hz, 6H), 0.76 (d,
6

ACCEPTED MANUSCRIPT
J= 15.7 Hz, 7H). 13C NMR (101 MHz, CDCl3) δ 171.4, 154.0, 139.8, 127.1, 126.1, 122.4, 120.5,
112.0, 75.6, 60.5, 53.3, 47.1, 41.5, 34.3, 31.5, 22.7, 22.1, 21.1, 20.9, 14.2. HRMS (TOF ES+)
m/z [M+H]+ calculated for C28H38NO3 436.2852, found 436.2850.
7-(Benzyl((2-isopropyl-5-methylcyclohexyloxy)carbonyl)amino)-5,6,7,8-
tetrahydronaphthalen-1-yl trifluoromethanesulfonate (11). 10 (0.4 g, 0.92 mmol) was
taken up in DCM (10 mL) and cooled to 0 °C. Pyridine (300 L) was added, followed by
trifluoromethanesulfonic anhydride (310 mg, 1.1 mmol). The mixture was stirred at 0 °C for
10 min, warmed to ambient temperature and stirred for a further 1 h. The mixture was
diluted with water (100 mL) and extracted with ethyl acetate (20 mL × 3). The organic layers
were combined, dried (Na2SO4) and concentrated in vacuo. The product was isolated with
flash column chromatography (silica gel; 0−20% ethyl acetate in hexane). The desired
fractions were pooled and concentrated to give 11 as a thick oil (350 mg, 0.62 mmol, 68%).
1H NMR (400 MHz, CDCl3) δ 7.32 (d, J = 7.7 Hz, 1H), 7.24 (m, 3H), 7.15 (t, J = 8.7 Hz, 2H), 7.06
(d, J = 7.0 Hz, 2H), 4.54 (t, J = 12.9 Hz, 4H), 3.02 (s, 1H), 2.86 (d, J = 17.1 Hz, 2H), 2.78 (t, J =
13.5 Hz, 1H), 2.11 (d, J = 10.8 Hz, 1H), 1.90 (s, 2H), 1.65 (d, J = 10.6 Hz, 2H), 1.49 (s, 1H), 1.26
(s, 1H), 1.04 (d, J = 8.7 Hz, 2H), 0.90 (d, J = 6.4 Hz, 6H), 0.80 (s, 7H). 13C NMR (101 MHz,
CDCl3) δ 148.1, 139.4, 139.2, 137.9, 129.1, 129.0, 128.6, 128.5, 128.3, 127.1, 126.7, 125.3,
120.2, 118.5, 117.0, 113.8, 75.6, 52.8, 47.4, 41.4, 34.3, 31.5, 29.8, 27.6, 27.3, 26.0, 23.2, 22.1,
21.5, 20.9, 16.1. HRMS (TOF ES+) m/z [M+Na]+ calculated for C29H36NO5F3SNa 590.2164,
found 590.2156.
2-Isopropyl-5-methylcyclohexyl-benzyl(8-(4-methylpiperazin-1-yl)-1,2,3,4-
tetrahydronaphthalen-2-yl)carbamate (12). 11 (510 mg, 0.90 mmol) in toluene (2 mL) was
added to a stirred solution of Pd2dba3 (25 mg), BINAP (40 mg) and Cs2CO3 (385 mg, 1.2
mmol) in toluene (5 mL) under argon. 1-Methylpiperazine (110 mg, 1.0 mmol) was added
7

ACCEPTED MANUSCRIPT
and the reaction was heated to 80 °C overnight. The mixture was then filtered through
celite, concentrated, and purified with flash column chromatography (silica gel;10%
methanol in DCM) to give a mixture of the two diastereomers. The two diastereomers were
subsequently isolated by flash column chromatography (silica gel; 3% methanol in DCM
rising to 20% methanol only after the first diastereomer had completely eluted). Fractions
for the two products were separately pooled and concentrated. No absolute configuration
was determined. Diastereomer (S)-12, 230 mg, 0.41 mmol), diastereomer (R)-12 (200 mg,
0.35 mmol) were both isolated as thick oils. The analytical data for diastereomer 2 are: 1H
NMR (400 MHz, CDCl3) δ 7.31 (d, J = 6.3 Hz, 4H), 7.24 (d, J = 8.4 Hz, 1H), 7.09 (t, J = 7.6 Hz,
1H), 6.86 (d, J = 7.7 Hz, 1H), 6.82 (d, J = 7.6 Hz, 1H), 4.73 (s, 2H), 4.27 (d, J = 128.0 Hz, 2H),
2.98 (d, J = 15.4 Hz, 1H), 2.88 (s, 4H), 2.70 (s, 2H), 2.46 (q, J = 9.2 Hz, 4H), 2.34 (s, 3H), 2.13 (t,
J = 15.1 Hz, 1H), 1.92 (s, 4H), 1.65 (s, 2H), 1.51 (s, 1H), 1.32 (d, J = 19.4 Hz, 1H), 1.06 (d, J = 8.1
Hz, 1H), 0.91 (d, J = 6.4 Hz, 6H), 0.81 (d, J = 25.6 Hz, 5H). 13C NMR (101 MHz, CDCl3) δ 151.3,
140.2, 136.6, 131.0, 128.4, 127.2, 126.8, 126.3, 123.9, 116.8, 75.3, 55.6, 53.6, 51.6, 47.3,
46.1, 41.5, 34.3, 31.4, 30.0, 28.1, 26.2, 23.3, 22.1, 21.0, 16.4. HRMS (TOF ES+) m/z [M+H]+
calculated for C33H48N3O3 518.3747, found 518.3757.
(R)-8-(4-Methylpiperazin-1-yl)-1,2,3,4-tetrahydronaphthalen-2-amine [(R)-13]. (R)-12 (200
mg, 0.35 mmol) was mixed in a 1:2 solution (3 mL) of HBr (33 wt % in acetic acid) and formic
acid in a sealed vessel. The reaction was heated to 80 °C overnight, neutralized to pH 7 with
NaOH (6 M) and extracted with DCM (30 mL × 3). The organic layers were combined, dried
and concentrated. The residue was taken up in methanol (5 mL) and Pd/C (10%, 30 mg) was
added. The vessel was put under vacuum and flushed with argon twice before being
pressurized with H2 gas at 5 bar. The mixture was stirred overnight, filtered through celite,
and concentrated to give (R)-13 as an off-white solid (80 mg, 0.32 mmol). This product was
8

ACCEPTED MANUSCRIPT
determined to be the desired eutomer by comparison of its specific optical rotation with
1
that of reference sample of known pharmacological activity/absolute configuration. H NMR
(400 MHz, CDCl3) δ 7.11 (t, J = 7.7 Hz, 1H), 6.91 (d, J = 7.6 Hz, 1H), 6.86 (d, J = 7.4 Hz, 1H),
3.15 (m, 2H), 3.03 (m, 2H), 2.91 (d, J = 4.5 Hz, 2H), 2.83 (s, 4H), 2.59 (s, 4H), 2.37 (s, 3H), 2.32
(q, J = 8.5 Hz, 1H), 2.13 (s, 4H), 2.00 (m, 1H), 1.60 (m, 1H). 13C NMR (101 MHz, CDCl3) δ
170.2, 150.9, 136.0, 129.6, 125.2, 123.1, 115.8, 59.4, 52.3, 46.6, 45.5, 34.1, 31.7, 27.6, 20.0,
13.2. HRMS (TOF ES+) m/z [M+H]+ calculated for C15H24N3 246.1970, found 246.1969.
4-Iodophenyl methanesulfonate (14). 4-Iodophenol (300 mg, 2.0 mmol) was taken up in
THF (5 mL) and triethylamine (0.2 mL) was added followed by methanesulfonyl chloride (0.5
mL, 2.9 mmol). The mixture was stirred for 3 h, then extracted with DCM (20 mL × 3), dried
(Na2SO4) and concentrated in vacuo. The product was isolated with flash chromatography
(silica gel; 0−30% ethyl acetate in hexane). The desired fraction was pooled and
1
concentrated in vacuo to give 14 as white solid (425 mg, quantitative yield). H NMR (400
MHz, CDCl3) δ 7.74 (d, J = 8.8 Hz, 2H), 7.05 (d, J = 8.9 Hz, 2H), 3.15 (s, 3H). 13C NMR (101
MHz, CDCl3) δ 148.9, 139.2, 124.1, 92.0, 37.6. HRMS (TOF ES+) m/z [M+H]+ calculated for
C7H7O3SI 298.9161, found 298.9158.
Pharmacological assays
Ligands 4 and 5 were assayed at AstraZeneca for 5-HT1B receptors agonist and
antagonist activity using a GTPS SPA (96 well) protocol. Data were analyzed by calculating
percent effect and/or IC50 (antagonist assay; 50% of stimulated response) or EC50 (agonist
assay; half maximum of curve) using an activity base template (see Supporting Information
for assay protocols).
9

ACCEPTED MANUSCRIPT
Radiochemistry
Synthesis of [11C]4: [11C]Carbon dioxide was produced with the 14N(p,α)11C
nuclear reaction in a target containing 1% oxygen in nitrogen, initially pressurized to 225 psi.
This target was bombarded with a beam (45 µA) of protons (16.5 MeV) from a PETtrace
cyclotron (GE Medical Systems; Milwaukee, WI). Irradiations were conducted for 40 min for
radiotracer production and typically provided about 110 GBq of [11C]carbon dioxide. Shorter
irradiations were used to produce lower amounts of [11C]carbon dioxide for initial
radiochemistry experiments. At the end of irradiation, radioactivity was released to a trap
containing molecular sieve (13X; 400 mg) at room temperature, which was then swept with
helium at 80 mL min-1 to remove residual oxygen.
[11C]Carbon dioxide was released from the molecular sieve trap at 280 °C in a
stream of helium (10 mL min-1) for concentration in a trap containing silica (10 mg) cooled in
liquid nitrogen. This trap was heated with a halogen lamp (150 W) to release [11C]carbon
dioxide in helium (10 mL min-1) for conversion into [11C]carbon monoxide by passage
through a heated (875 C) quartz tube (22 cm length; 0.7 cm i.d.) packed with molybdenum
wire (99.97%, 0.05 mm diameter; Strem Chemicals, Newburyport, MA). The effluent was
passed through ascarite to strip-off unconverted [11C]carbon dioxide and the [11C]carbon
monoxide was then trapped in a second silica cryotrap at –196 C in liquid nitrogen. By
heating this trap with a halogen lamp, [11C]carbon monoxide was released into helium for
transfer to an autoclave pre-loaded with a solution of Pd2dba3 (1.3 mg, 1.4 µmol), Xantphos
(3.0 mg, 5.2 µmol) and iodo-precursor (14; 2.4 mg, 8.1 µmol) in THF (80 µL). The autoclave
was then pressurized to ~3500 psi with THF by a solvent pump (SSI Isocratic; Scantech Lab,
Sweden) and heated at 150 C for 3 min. The reaction mixture was then flushed out of the
autoclave into a 5-mL glass vial (Alltech; Sigma Aldrich; St Louis, MI) that had been loaded
10

ACCEPTED MANUSCRIPT
with a solution of the R-amine [(R)-13], 3.2 mg, 13 µmol) and Bu3N (10 L, 42 µmol) in THF
(50 µL). The vial was heated from 67 C rising to 100 C while THF was evaporated off under
N2 gas purge (200 mL min-1; 3 min) through a vent needle.
The dry radioactive residue was re-dissolved in acetonitrile (700 L) and diluted
with water (2.5 mL). The aqueous mixture was injected onto a Luna C18 column (10 µm, 10
 250 mm; Phenomenex; Torrance, CA) eluted at 6 mL min-1 using a gradient of freshly
prepared 0.2% aqueous NH4OH (A, pH = 8.5) and acetonitrile (B) in the mobile phase. The
gradient began at 10% B for 1 min, increased to 45% B over 25 min, and finally reached 90%
B over 1 min. Absorbance was monitored at 254 nm (System Gold 166; Beckman Coulter
Inc., Pasadena, CA) while radioactivity was monitored with a pin-diode detector (Bioscan
Inc., Washington DC). The fraction eluting between 16.5 to 17.5 min was collected and then
evacuated at 80 C for 1 min to remove solvents. The radioactive residue was reconstituted
in saline (9 mL) plus ethanol (1 mL) and sterilized by filtration through a 0.22 µm sterile filter
(Millipore-MP; Waters Corp.; Milford, MA). An aliquot (~ 60 µL) of the formulated product
was analyzed with radio-HPLC (high-performance liquid chromatography) on a Luna C18 column
(10 µm, 4.6  250 mm; Phenomenex) to measure radiochemical purity and molar activity.
The column was eluted at 2 mL min-1 with a gradient of 0.2% aq. NH4OH (A) and acetonitrile
(B) as mobile phase. The gradient began at 40% B and was linearly increased to 70% B over
10 min. The eluate was monitored for absorbance at 254 nm (System Gold 166 detector)
and for radioactivity with a PMT detector (Bioscan Inc.). The retention time of [11C]4 was 6.7
min. The absorbance response had been pre-calibrated with respect to mass of 4 in the
injectate, to permit calculation of molar activity.
Synthesis of [11C]5: [11C]Methane was prepared through the 14N(p,α)11C
reaction by bombarding a mixture of nitrogen and hydrogen (10%) gas with a beam of 16.7
11

ACCEPTED MANUSCRIPT
MeV protons produced from a biomedical cyclotron (GE PETtrace; Uppsala, Sweden).
[11C]Methane was converted into [11C]methyl iodide by high temperature iodination in a gas-
phase recirculation system and subsequently transformed into [11C]methyl triflate, as
described previously.[18, 19] [11C]5 was obtained by trapping [11C]methyl triflate in a
solution of N-desmethyl precursor 6 (0.1−0.3 mg) and sodium hydroxide (0.1 M, 6 µL) in
acetone (400 µL) at room temperature. The sealed reaction vial was left for 2 min. The
reaction mixture was injected onto a reversed phased HPLC column (ACE C-18; 250 × 10 mm;
ACE) eluted with acetonitrile: HCO2NH4 (0.1 M) (35:65 v/v) at 5 mL min-1. Eluate was
monitored in series for absorbance at (254 nm) and radioactivity (Geiger–Müller tube). The
fraction containing [11C]5 (tR = 12−13.5 min) was collected, formulated in saline (pH = 6; 4−6
mL), and finally filtered through a sterile filter (0.22 µm; Millipore Millex®GV) into a sterile
vial.
The radiochemical purity of the product was determined with reversed phase
HPLC on ACE 5 C18-HL column (C18, 3.9 × 300 mm, 10 µm particle size; Advanced
chromatography technologies Ltd), with eluate monitored in series for absorbance at 254
nm and radioactivity (β-flow detector; Beckman, Fullerton, CA). For the analysis, the column
was eluted at 3 mL min-1 with a gradient of 10−90% acetonitrile in HCO2NH4 (0.1 M) for 10
min. [11C]5 (tR = 4.4 min) was identified by its co-mobility with reference sample of 5.
The molar activity of the final product was measured with HPLC under the
conditions described above for radiochemical analysis. The absorbance (λ = 254 nm)
response was calibrated for mass of ligand and calculated as the radioactivity of the
radioligand (GBq) divided by the amount of the associated carrier substance (µmol). Each
sample was analyzed three times and compared to a reference standard also analyzed three
times.
12

ACCEPTED MANUSCRIPT
Apparent pKa determination
The radiochemical purity of the [11C]4 preparation was 99.1% as determined by radio-HPLC
on an X-Terra C18 column (10 µm, 7.8 x 300 mm; Waters Corp.) and elution with a mobile
phase of MeOH:H2O:Et3N (82.5:17.5:0.1 by vol.) at 4.0 mL min-1.
Phosphate buffers (0.15 M) with pH 3.0 to 10.5 at half pH intervals were
prepared from NaH2PO4 (0.15M) and Na2HPO4 (0.15M) and distributed 1 mL per tube. [11C]4
(2.51 ± 0.06 MBq) formulated in saline/10% ethanol containing 0.7 mM ascorbic acid (5 µL, n
= 48) was added to each test tube. Triple sets of such tubes were prepared simultaneously.
Cyclohexane (1.0 mL) was added to each of the tubes containing the buffer and [11C]4. The
tubes were vigorously vortexed for 1 min, and then centrifuged for one minute. Each phase
was then separated, sampled (200 µL) and counted in a γ-counter. The distribution
coefficient (D) was then calculated by dividing the counts detected in the organic phase by
the total activity present in both phases.
The apparent pKa value was determined as the pH value where the
concentration of ionized and non-ionized [11C]4 were equal by plotting D versus pH and
applying nonlinear regression analysis (GraphPad Prism version 4.03 for windows (GraphPad
Software; San Diego, CA, USA) with “One site competition” curve-fitting).
LogD determination
The radioligand was placed in sodium phosphate buffer (0.15 M; pH 7.4) for the
duration of the logD dtermination followed by the measurement of its radiochemical purity
to obtain information on its stability in buffer. Furthermore, the radiochemical composition
of the aqueous buffer remaining after their extraction with 1-octanol was determined with
HPLC on an X-Terra C18 column (10 µm, 7.8 x 300 mm; Waters Corp.) and eluted with a
mobile phase of MeOH:H2O:Et3N (82.5:17.5:0.1 by vol.) at 3.5 mL min-1. The HPLC system
13

ACCEPTED MANUSCRIPT
consisted of Beckman Gold analytic pumps (Beckman Coulter, Inc.), an in-line photodiode-
array detector and a flow-through NaI(Tl) scintillation detector-rate meter (Bioscan). Data
from the radioanalysis were collected and stored with Bio-Chrome Lite software (Bioscan)
and analyzed after decay correction of the radiochromatograms.
Sodium phosphate buffer (0.15 M) was prepared by mixing NaH2PO4 and
Na2HPO4 solutions to give a buffer of 0.15 M concentration and pH of 7.4. 1-Octanol was
pre-equilibrated overnight with phosphate buffer.
[11C]4 (~ 363 MBq in 2.3 mL) was added to sodium phosphate buffer (0.15 M;
6.5 mL) and mixed well. Aliquots (1 mL) were distributed to eight borosilicate disposable
culture tubes (13 x 100 mm). Six tubes, each containing 1-octanol (1.0 mL), and two tubes
without 1-octanol, served for determining the stability of [11C]4 and adsorption to the walls
of the tube by incubation for the duration of the study. Radioactivity from six tubes were
extracted with 1-octanol by vortexing the tubes for 1 min. The tubes were then centrifuged
for 1 min. The organic and the aqueous phases were separated. Each phase was then
sampled (50 µL and 200 µL, respectively) and counted separately in a well-type -counter
(model 1480 Wizard; Perkin-Elmer) with an electronic window set at between 360 and 1,800
keV. The remaining two tubes contained the radioactivity in buffer and served as stability
measures for the radioligand. Radioactivity in the aqueous phases (200 µL) resulted in
counting errors of 0.7 ± 0.03% (n = 6) at one standard deviation. The aqueous phases were
injected onto the HPLC. The resulting radiochromatograms provided correction factors for
the -counter counts to determine parent radioactivity only (corrected radioactivity).
14

ACCEPTED MANUSCRIPT
Monkey PET imaging using [11C]4
PET imaging experiments in monkey at NIMH were performed in accordance
with the Guide for Care and Use of Laboratory Animals and were approved by the National
Institute of Mental Health Animal Care and Use Committee.[20]
A single male rhesus monkey (Macaca mulatta; 11.7 kg) was used for four PET
scan sessions in which the monkey was initially anesthetized with ketamine (10 mg kg-1, i.m.)
and then maintained in anesthesia with 1.5% isoflurane. Thus, at baseline [11C]4 (306 MBq;
69 GBq µmol-1) was injected intravenously as a bolus (session 1). On the same day, at more
than 3 h later, AR-A000002 (2 mg kg-1, i.v.) was administered intravenously at 15 min before
[11C]4 (343 MBq; 71 GBq mol-1) to pre-block 5-HT1B receptors (session 2). For P-gp-
preblocked conditions six weeks later, tariquidar (8 mg kg-1, i.v.) was administered at 25 min
before injection of [11C]4 (292 MBq; 18 GBq µmol-1) (session 3). Finally, 4 weeks later, the
monkey was imaged after administration of [11C]4 (240 MBq; 15 GBq mol-1) under both 5-
HT1B-receptor and P-gp preblocked conditions achieved by giving tariquidar (8 mg kg-1, i.v.)
at 22 min before [11C]4 and AR-A000002 (2 mg kg-1, i.v.) at 15 min before [11C]4 injection
(session 4). Tariquidar treatments had to be at least four weeks apart for the same monkey
for drug safety reasons. In each experiment, PET images of brain were acquired with a
microPET Focus 220 scanner (Siemens Medical Solution; Knoxville, TN) for 90 min with scan
durations ranging from 30 s to 5 min. The position of the head was fixed using a stereotaxic
frame. Electrocardiogram, body temperature, heart, and respiration rates were monitored
throughout the experiment.
Monkey plasma radiometabolite analysis
During monkey PET imaging in session 1 and 2, arterial blood samples were
drawn for analysis of radiometabolites in plasma and for determination of a
15

ACCEPTED MANUSCRIPT
radiometabolite-corrected arterial input function. Samples (0.5−1 mL each) were drawn at
15 s intervals until 120 s, followed by 0.5- to 4-mL samples at 3, 5, 10, 30, 60, 90, and 120
min. The concentration of parent radioligand was measured with HPLC on an X-terra column
(10 µm, 7.8 × 300 mm; Waters Corp.) eluted at 5.0 mL min-1 with
methanol−water−triethylamine (85:15:0.1, by vol.), after separating plasma from whole
blood and processing, as previously described.[21]
Compartmental modeling of PET data
Images were reconstructed using Fourier rebinning plus two-dimensional
filtered back-projection. PET images were co-registered to a standardized monkey MRI
template using SPM5 (Wellcome Trust Centre, London, UK). A set of 34 predefined brain
regions of interest from the template were then applied to the co-registered PET image to
obtain regional decay-corrected time-activity curves. All PET images were corrected for
attenuation and scatter. Uptake of radioactivity in each region of interest was expressed as
SUV. Total distribution volumes (VT) were estimated for different regions by two tissue
compartmental modeling with PET brain time-activity curves and the measured
radiometabolite-corrected arterial input function. The PET data analysis was performed
using PMOD (PMOD Technologies Ltd.; Zurich, Switzerland).
Plasma free fraction determination
The plasma free fraction (fp) of [11C]4 was measured with ultra-filtration
through membrane filters (Centrifree; Millipore), as previously described.[22] Briefly,
formulated [11C]4 (170 kBq at the time of counting in ~23 µL) was added to plasma (650 µL).
The mixture was incubated at room temperature for 10 min and then processed as
described previously. Radioactive ultra-filtration components were allowed time to decay
16

ACCEPTED MANUSCRIPT
within the optimal range of the -counter before being counted. Quantification of the ultra-
filtrates was done gravimetrically.
Monkey PET imaging using [11C]5
The PET study in nonhuman primate at Karolinska Institutet was approved by
the Animal Ethics Committee of the Swedish Animal Welfare Agency (Dnr N185/14) and was
performed according to the relevant guidelines of the Karolinska Institutet (“Guidelines for
Planning, Conducting, and Documenting Experimental Research” (Dnr 4820/06-600) and
“Guide for the Care and Use of Laboratory Animals”.
A female cynomolgus monkey (5.5 kg) was supplied by Astrid Fagraeus
Laboratory of the Swedish Institute for Infectious Disease Control (Solna, Sweden).
Anesthesia was initiated via intramuscular injection of ketamine hydrochloride (ca. 10 mg kg-
1) and maintained after endotracheal intubation by administration of a mixture of
sevoflurane, oxygen and medical air. The monkey was observed continuously during the
days of PET measurements. Body temperature was maintained by Bair Hugger-Model 505
(Arizant Healthcare Inc., MN) and monitored with an esophageal thermometer. Heart and
respiration rates were continuously monitored by using PC-VetGard ™ system. The monkey
head was fixed in position throughout PET scanning, as described previously.[23] For
baseline PET scan [11C]5 (146 MBq; 324 GBq mol-1) was injected intravenously during 5 s with
simultaneous start of PET data acquisition on the Siemens HRRT scanner. Radioactivity in
brain was measured continuously for 123 min according to a preprogrammed series of 34
frames.[24] For the measurement of metabolites and plasma free fraction (fp) manual blood
samples (2 mL) were obtained at 4, 10, and then every 10 min. The fp was determined by
HPLC with radio-detection as previously described.[25]
17

ACCEPTED MANUSCRIPT
Rat PET imaging
In one experiment, three rats (Sprague Dawley; 213.3 ± 6.4 g) that had been
anesthetized with ketamine (10 mg kg-1) and then maintained in anesthesia with 1.5%
isoflurane were used for PET scanning sessions after bolus intravenous administration of
[11C]4 (45 ± 1 MBq; 76 GBq µmol-1). One rat was studied at baseline. One rat was studied
under a P-gp-preblocked condition, achieved by administering tariquidar (16 mg kg-1, i.v.) at
25 min before [11C]4 injection. In another rat, tariquidar (16 mg kg-1) was administered 20
min before [11C]4 injection and AR-A000002 (2 mg kg-1) at 15 min after [11C]4 injection.
In a separate experiment, three rats (Sprague Dawley; 346.0 ± 15.5 g) were
anesthetized with ketamine (10 mg kg-1) and maintained in anesthesia with 1.5% isoflurane.
They were used for PET scanning sessions with non-radioactive 4, AR-A000002 and
fenfluramine as treatments. Thus, after treatment of all rats with tariquidar (8 mg kg-1, i.v.),
one rat was injected intravenously with 4 (2 mg kg-1, i.v.), another one with AR-A000002 (2
mg kg-1, i.v.); and the third with fenfluramine (2 mg kg-1, i.v.), all 15 min after injection of a
bolus of [11C]4 (39.3 ± 0.6 MBq; 36 GBq µmol-1).
In each experiment, PET images of brain were acquired with a microPET Focus
220 scanner (Siemens Medical Solution; Knoxville, TN) for 90 min. The PET data analysis was
performed using PMOD (PMOD Technologies Ltd.; Zurich, Switzerland).
Ex vivo measurement of radiometabolites of [11C]4 in rat brain
Tariquidar (8 mg kg-1) was injected intravenously through the penile vein into
three anesthetized rats (269 ± 53 g) at 6.7 ± 1.2 min before likewise injection of [11C]4. The
rats were anesthetized with 1.5% isoflurane in oxygen and sacrificed via thoracotomy at 30
min after injection. Anticoagulated blood was drawn from the myocardium and the rats
18

ACCEPTED MANUSCRIPT
were decapitated to harvest forebrain and cerebellum. Each tissue was removed, weighed,
and immediately subjected to radioanalysis. All tissue assays were corrected for physical
decay with a half-life of 20.385 min. The whole blood samples were centrifuged to separate
the plasma. The plasma samples (~450 µL) were then de-proteinated with acetonitrile (720
µL). Whole blood (50 µL each) and plasma samples (50 µL each) were quantified for
radioactivity as previously described[21] in an automatic -counter (model 1480 Wizard;
Perkin-Elmer) with an electronic window set at 360 – 1800 keV (counting efficiency =
51.84%). The radioactivity of brain tissue was also counted in the -counter before
homogenization in acetonitrile (3.0 mL) with a hand-held tissue Tearor (model 985-370;
BioSpec Products Inc.) and after a second homogenization in the presence of added water
(1.0 mL). These homogenates were counted in the -counter to calculate the percentage
recovery of radioactivity into the acetonitrile extracts. The homogenates were then
centrifuged at 10,000g for 1 min. The clear supernatant was injected onto the HPLC column
through nylon filters. Radioactivity in the resulting precipitate was used to calculate the
percent recovery of activity in the acetonitrile supernatant.
Radio-HPLC analysis was performed on X-terra® C18 (10 µm, 7.8 mm x 300 mm,
Waters Corp., Milford, MA), eluted with methanol:water:Et3N (80:20:0.1; by vol.) at 4.0 mL
min-1. Eluates were monitored with an in-line flow-through NaI(Tl) scintillation detector
(Bioscan). Data were stored and analyzed on a PC using the software “Bio-ChromeLite”.
Data in each radiochromatogram were decay-corrected to the respective HPLC injection
time.
Ex vivo measurement of radiometabolites of [11C]4 in P-gp/BCRP knock-out mice brain
P-gp/BCRP knockout (KO) mice (Taconic Farm, Germantown, NY) were
anesthetized with 1.5% isoflurane and oxygen and sacrificed via thoracotomy at 30 min after
19

ACCEPTED MANUSCRIPT
radioligand injection. Heparinized (anticoagulated) blood was drawn from the myocardium,
and the forebrains and cerebella were subsequently harvested. The tissues were weighed
and immediately subjected to radioanalysis. Whole blood samples were centrifuged to
separate the plasma. Plasma samples (300−350 µL) were then deproteinated with
acetonitrile (720 µL). Whole blood (50 µL each) and plasma samples (50 µL each) were
quantified for radioactivity, as previously described above in an automatic -counter. The
radioactivities of brain tissues were also counted in the -counter before homogenization in
acetonitrile (1.0−1.5 mL) using a hand-held tissue Tearor (model 985-370; BioSpec Products
Inc.) and again after homogenization in the presence of water (0.5 mL). Anticoagulated
blood (~1 mL; STD-A), plasma (STD-B) and fresh brain (1.3 g) were harvested from a control
rat (409 g) and a control mouse (48.2 g). The brains of the control animals were harvested
and immediately weighed and homogenized with twice their weights of ice cold saline.
Plasma was separated from whole blood by centrifugation. A radioligand dose was then
added to these three tissue samples from each animal before incubation in a shaking water
bath at 37 C for 1 h (plasma, blood and brain homogenate; 30 µCi mL-1). Aliquots (50 µL)
were removed for radio-analysis and processed as detailed above.
Quantum chemical calculations
The geometry and energetics of compounds 1-5 in the presence of a single
water molecule were obtained with density functional theory at the level of B3LYP/6-31G*.
These quantum chemical calculations were all carried out in the reaction field of water with
the polarizable continuum model as implemented in Gaussian 09 software.[26]
20

ACCEPTED MANUSCRIPT
Results
Chemistry
The structure of 4 lends itself to be labeled with carbon-11 in at least two
positions by different methods, namely 11C-carbonylation and 11C-N-mesylation.[27, 28]
Both options were considered, and 11C-carbonylation was selected mostly because of ready
access to [11C]carbon monoxide as labeling agent.[29] Two precursors were synthesized for
use in a reaction of [11C]carbon monoxide that couples a 4-iodophenyl methanesulfonate
(14) with a requisite homochiral amine [(R)-13]. The iodide, rather than the bromide, was
selected as a coupling partner to minimize the risk of Pd insertion into the aryl-sulfonate
bond. 14 was prepared simply by mesylation of 4-iodophenol with methanesulfonyl
chloride. (R)-13 required a multistep synthesis with careful consideration on how to
separate and analyze diastereomers (Scheme 1).
Starting from commercially available 8-methoxytetralone (7), reductive
amination with benzylamine, using NaCNBH3 as reducing agent, gave racemic 8 in 85% yield.
A homochiral protecting group was then incorporated through treatment of the amine with
(1R)-menthylchloroformate to afford a diastereomeric mixture (9) in 92% aggregate yield.
Only limited separation of the diastereomers could be achieved on silica gel. Therefore, the
diastereomeric mixture was used in the next step. Demethylation of 9 with BBr3 in
dichloromethane (DCM) gave the desired pair of diastereomeric phenols (10) in 83% yield.
These phenols were then treated with triflic anhydride and pyridine in DCM to give the
corresponding pair of diastereomeric aryl triflates (11) in 68% yield. A Buchwald-Hartwig
amination reaction between 11 and 1-methylpiperazine gave a mixture of diastereomers
(12) from which (R)-12 and (S)-12 were separated by silica gel flash chromatography. The
21

ACCEPTED MANUSCRIPT
diastereomers were deprotected separately in two steps to give (R)-13 and (S)-13. First, the
(1R)-menthyl group was removed by heating the carbamate with HBr in formic acid and
acetic acid in a sealed vial. The benzyl group was then removed with H2 and 10% Pd on
carbon. The desired homochiral amines [(R)-13 and (S)-13)] were each obtained in 90%
[ ]22
yield. The specific optical rotations ( 훼 ) of the amines were compared with that of an
퐷
authentic reference sample obtained from AstraZeneca. The analysis showed specific
[ ]22
rotations 훼 for enantiomer (R)-13 and the reference sample of +3.5 and +3.7
퐷
[ ]22
respectively, whereas enantiomer (S)-13 had a specific rotation 훼 of −2.8.
퐷
Radiochemistry
Radiosynthesis of [11C]4. [11C]Carbon monoxide was produced from cyclotron-
produced [11C]carbon dioxide by passage over heated molybdenum.[27] 4-Iodophenyl
methanesulfonate (14) was sonicated with Pd2(dba)3 and Xantphos in THF and injected
under pressure into an autoclave along with the [11C]carbon monoxide (Scheme 2). The
mixture was heated for 5 min before being released into a vial containing the chiral amine
(R)-13 and tributylamine (TBA). The vial was heated under helium flow to evaporate residual
solvent and generate the desired amide. This two-pot procedure gave fewer unidentified
radioactive side-products than the one-pot procedure.[30] [11C]4 was produced starting
from [11C]CO2 in yields of 6.4 ± 1.6% (1240 ± 347 MBq; n = 7) within 50 min. [11C]4 was then
separated with reversed phase HPLC and formulated in saline (with 10% v/v EtOH) for
intravenous injection. HPLC showed that the radioligand was radiochemically pure and free
of any appreciable chemical impurities (see Supporting Information for radiochromatogram).
Molar activities were 83 ± 51 GBq µmol-1 (n = 7) at end of synthesis.
22

ACCEPTED MANUSCRIPT
Radiosynthesis of [11C]5. Several experiments were performed to arrive at
these conditions to produce this radioligand for single PET measurement in NHP. Cyclotron
produced [11C]methane was converted into [11C]methyl iodide by high temperature
iodination in a gas-phase recirculation system and subsequently transformed into
[11C]methyl triflate. [11C]5 was obtained by trapping [11C]methyl triflate in a solution of 6 in
acetone (Scheme 3), isolated with reversed phase preparative HPLC (see Supporting
Information for radiochromatogram), and formulated in saline in a sterile vial. The
radioligand represented 70% of the radioactivity in the HPLC injectate. Formulated [11C]5
(1025 MBq) was obtained in 98% radiochemical purity and with a molar activity of 654 GBq
µmol-1 (n = 1) after 30 min total synthesis time after end of radionuclide production.
PET experiments with [11C]4 in monkey
PET images (Fig. 2A/B) and time-activity curves (Fig. 3A) obtained after
administration of [11C]4 showed very low radioactivity uptake in monkey brain (< 0.2 SUV)
before and after blockade of 5-HT1B receptors with AR-A000002 (AZD3783, 2 mg kg-1, i.v.).
AR-A000002, at an intravenously administered dose of 2 mg kg-1, has previously been shown
to block monkey brain 5-HT1B receptors in PET experiments with other radioligands.[31]
Tariquidar(TQD) is a P-glycoprotein (P-gp) inhibitor that may be administered intravenously
to inhibit P-gp in vivo.[32] Tariquidar (8 mg kg-1) increased brain radioactivity uptake to
about 0.5 SUV. No blocking effect was observed after combining tariquidar and AR-A000002
administration, indicating absence of any specific 5-HT1B receptor binding.
Very high radioactivity uptake in pituitary gland (> 6 SUV) was observed at
baseline and under all other conditions. Administration of tariquidar or tariquidar plus AR-
A000002 did not alter [11C]4 uptake in pituitary gland (Fig. 2C/D) but uptake was markedly
increased under 5-HT1B receptor blockade (Fig. 3B).
23

ACCEPTED MANUSCRIPT
Analysis of plasma in monkey PET experiments with [11C]4
Following injection of [11C]4 into monkey at baseline, radioactivity quickly
decreased in plasma. Analysis of plasma showed five radiometabolites, [11C]A−[11C]E (see
Fig. S5). The ability of these radiometabolites to enter brain is unknown. [11C]E, which was
the most lipophilic radiometabolite and the most likely able to enter brain, reached only
2.6% of radioactivity in plasma by 90 min after [11C]4 injection. The time-courses of
unchanged [11C]4, in terms of contribution to percentage radioactivity in plasma and to
unchanged [11C]4 concentration (SUV), were similar at baseline and under 5-HT1B receptor
blocked conditions (Fig. 4A). The curves in Fig. 4B constitute the radiometabolite-corrected
arterial input functions for use in compartmental modeling.
Compartmental modeling of monkey PET data from [11C]4
One-tissue (1TC) and two-tissue (2TC) compartmental models were used with
PET data and arterial input functions to derive whole brain total volumes of distribution
(VT’s) under four types of monkey condition (Fig. 5). By either model, VT values at baseline
were very low and did not increase appreciably under 5-HT1B receptor blocking condition.
Pretreatment with tariquidar increased VT by either model, which indicates [11C]4 is possibly
a substrate for the efflux transporter P-gp. However, the radiometabolites are also possibly
P-gp substrates. Therefore, the increased radioactivity in brain could have significant
contributions from radiometabolites.
Blockade of 5-HT1B receptors under P-gp preblocked conditions did not reduce
VT in either model (Fig.5), showing again absence of any clear receptor-specific signal. Both
models provided essentially the same VT data
24

ACCEPTED MANUSCRIPT
PET experiments with [11C]4 in rat
After injection of [11C]4 into Sprague Dawley rats, PET showed low uptake of
radioactivity into brain. Brain uptake was higher in rats pretreated with tariquidar, but this
uptake was not displaced by administration of 5-HT1B receptors ligands. No 5-HT1B receptor
specific binding could be detected in rats treated with tariquidar and AR-A000002 (Fig. 6A).
These results parallel those in monkey. Furthermore, treatment of rats with 4 or AR-
A000002 at 15 min after radioligand injection had no effect on brain time-activity curves
(Fig. 6B). Administration of fenfluramine at 15 min after radioligand injection to release
serotonin into the synaptic cleft showed no clearance of activity from brain. The tariquidar
dose was initially 16 mg kg-1 but was reduced to 8 mg kg-1 in the second set of PET imaging
experiment due to safety concerns for the rats.
Ex vivo experiments in rats and mice with [11C]4
In tariquidar-treated rats, plasma radioactivity was 43.7 ± 3.2% (n = 3)
unchanged parent radioligand at 30 min after injection. Radioactivity in ex vivo forebrain
and cerebellum of these rats was respectively 99.7 ± 0.4% (n = 3) and 98.8 ± 0.5% (n = 3)
unchanged parent radioligand at 30 min after injection. The forebrain uptake was 0.8 ± 0.1
SUV (n = 3) and the cerebellum uptake was 1.0 ± 0.1 SUV (n = 3).
In P-gp/BCRP knockout (KO) mice, the ex vivo plasma composition was 43.2 ±
0.8% (n = 3) parent radioligand at 30 min after injection. At this time, radioactivity in ex vivo
forebrain and cerebellum of KO mice was 97.0 ± 1.0% (n = 3) and 99.0 ± 0.4% (n = 3) parent
radioligand, respectively. There were no regional differences in the distribution of [11C]4 in
mouse brain. The forebrain uptake was 0.6 ± 0.2 SUV (n = 3) and the cerebellum uptake was
0.7 ± 0.1 SUV (n = 3).
25

ACCEPTED MANUSCRIPT
Apparent pKa, lipophilicity, and plasma free fraction of [11C]4
To probe for the underlying reasons behind the poor brain entry of [11C]4,
apparent pKa, lipophilicity, and plasma free fraction were measured for [11C]4. [11C]4 was
99.0% ± 0.1% (n = 2) stable in sodium phosphate buffer (pH 7.4) for at least 2.5 h. There was
no adsorption to the walls of glass test tubes from buffer (< 0.8 ± 0.6%; n = 2) or human
plasma (< 1.4%).
Apparent pKa. The apparent pKa of [11C]4 was 8.00 ± 0.03 (n = 3), based on a method
distributing [11C]4 between cyclohexane and buffer.[33] This shows that, according to the
Henderson-Hasselbalch equation, the radioligand would be protonated to a large extent
(79.9%) at physiological pH (7.4).
Lipophilicity (logD7.4). The measured logD7.4 of [11C]4 at room temperature in 1-octanol was
2.01 ± 0.01 (n = 6).
Plasma free fraction (fP). The plasma free fraction (fp) of [11C]4 in human plasma was 18.0 ±
0.8% (n = 3) and reflects the moderate lipophilicity of the radioligand.[34]
PET experiments with [11C]5 in monkey
[11C]5 was studied at baseline in one female cynomolgus monkey. Upon
administration of radioligand, brain uptake reached only 0.6 SUV and slowly decreased to
0.25 SUV over the 123-min duration of scanning. No differences in brain region uptakes
were detected (Fig. 7). High radioactivity uptake in the pituitary was also observed.
Analysis of plasma in monkey PET experiments with [11C]5
Analysis of plasma following injection of [11C]5 showed one major
radiometabolite and two minor radiometabolites(see Fig. S6). Parent compound decreased
26

ACCEPTED MANUSCRIPT
to 30% after 60 min. All radiometabolites were less lipophilic than parent compound and it
is unlikely they would enter the brain to any greater extent. The plasma free fraction was
greater than for [11C]4 and was measured to be 50.30 ± 0.03% (n = 6). Recovery was
Structure, P-gp substrate behavior and ability to cross the BBB.
The measured apparent pKa of 4 (8.00) indicates the N-methyl-piperazinyl ring
will be mostly (79.9%) protonated at physiological pH. This should be true for an N-methyl-
piperazinyl ring in similar structures, such as [11C]1−[11C]5. After intravenous injection, the
non-protonated forms of [11C]1−[11C]5 should be well represented in plasma (20%) and have
ability to diffuse across the BBB (blood-brain barrier), in the absence of efflux transporter
action.
Generally, charged species have very low ability to cross the BBB, except when
the formal charge is spread across the molecule. We considered that ligands 1−3 might form
an internal hydrogen bond (H-bond) with the amide nitrogen, potentially to delocalize
charge, and so confer ability to cross the BBB. To investigate such a possibility, we
performed quantum chemical calculations at the level of B3LYP/6-31G* to obtain the
structures and energetics for compounds 1−5 (Table 1 and Fig. S7). To single out the effect
of the potential hydrogen bond between the amide and piperazine nitrogen, these
calculations were all carried out with the protonated N-methyl-piperazine plus a water
molecule (Fig. 8) in the reaction field of water.
Conformer 1b was obtained by varying the dihedral angle (φ1) centered on the
C−N bond of 1a with the N-methyl proton facing the bicyclic ring (Fig. 8). Conformer 1c was
generated by further varying the dihedral angle (φ2) centered on the C−C bond of 1b making
its amide oxygen form an H-bond with the water molecule (Fig. 8). The energy minimized
27

ACCEPTED MANUSCRIPT
structure of 1b is less stable than 1a by ΔG of 2.3 kcal mol-1. Of the three conformers, 1c is
most stable due to the water molecule forming H-bonds with both the N-methyl proton and
the amide oxygen. The conformer 1b lacking such a water-bridge is less stable than 1a that
H-bonds with the water molecule in an open space. It is noted that in the absence of a water
molecule, 1c is the least energetically favorable due to the steric repulsion caused by the
amide oxygen to the neighboring ring oxygen atom.
Discussion
Absence of receptor-specific signal might be due to one or more of several
reasons. For example: (1) radioligand is not entering brain even under P-gp blocked
condition; (2) the radioligand will only bind to 5-HT1B receptors that are in the G-protein
coupled state, which may exist in relatively much lower density than the total population of
5-HT1B receptors seen by antagonists; or (3) the off-rate of the radioligand from the G-
protein coupled receptor is too fast.
The apparent pKa is attributable to the piperazinyl ring, most likely at the N-
methyl position.[35] The N-methyl-piperazinlyl group features in many effective PET
radioligands, such as the 5-HT1B receptor ligands shown in Fig. 1. Also many effective 5-HT1A
receptor PET radioligands have an aryl piperazine sub-structures, notably [11C]WAY-100635
([11C]WAY)[36] and [11C]DWAY.[37] Therefore, the pKa of 4 does not appear to be an
impediment to brain entry by passive diffusion.[38]
Because 4 has formal properties consistent with ability to penetrate the
BBB,[17] and because 4 is structurally related to other BBB-penetrant radioligands,[39] we
suspected that [11C]4 is prevented from entering brain by efflux transporters.[40] The
28

ACCEPTED MANUSCRIPT
lipophilicities, pKa and fP values of [11C]4 and [11C]5 appeared all to be conducive to
conferring good ability to penetrate the BBB by passive diffusion. In the case of [11C]4 at
least one efflux transporter, namely P-gp, was found to limit BBB penetration. This may also
be true for [11C]5, although this was not tested. The efflux transporter liability of [11C]4 is
somewhat surprising given that its structure resembles radioligands that do enter brain quite
well, such as [11C]1−3. A comparison of the structures of [11C]4 and [11C]5 with previous BBB-
penetrant 5-HT1B receptor radioligands [11C]1 and [11C]3 reveals notable differences, such as
the lack of planarity and heteroatoms in the bicyclic ring systems of [11C]4 and [11C]5, and
the reversed amide bond (Fig. 1). Nonetheless, these structural features are also present in
AR-A000002, which because it can be used as a brain 5-HT1B blocking agent, clearly has some
ability to enter brain after intravenous administration of a pharmacological dose. In the
study by Andersson et al,[39] a radiolabeled 5-HT1B receptor ligand (15, Fig. 9), having the
same bicyclic ring system and amide direction as [11C]3, failed to enter monkey brain to any
measurable extent after intravenous injection of a microdose.
The comparison of both energetics and structures of 1−5 suggests that the
protonated N-methyl-piperazinyl group makes compounds 1−3 more amenable to BBB
penetration by internal H-bonding mediated by a water molecule. This internal H-bonding
may well confer thermochemical stability to their closed conformation while attenuating the
charge on the protonated N-methyl-piperazinyl. The BBB-penetrant compounds 1−3 can
form energetically favorable water-bridged conformations in terms of both enthalpy (ΔH)
and Gibbs free energy (ΔG) whereas for the low-BBB-penetrating compounds 4−5 the open
conformations are more stable in terms of ΔG. This suggests that under physiological
conditions the piperazinyl nitrogen is fully charged and would hinder BBB penetration by
passive diffusion. Our quantum chemical calculations further indicate that the water-
29