Evaluation of [18F]2FP3 in pigs and non-human primates

Article abstract of DOI:10.1002/jlcr.3692

So far, no suitable 5-HT₇R radioligand exists for clinical positron emission tomography (PET) imaging. [¹⁸F]2FP3 was first tested in vivo in cats, and the results were promising for further evaluations. Here, we evaluate the radiolig

Full text of DOI:10.1002/jlcr.3692

Evaluation of [18F]2FP3 in pigs and non-human primates
Hanne D. Hansen1, Cristian C. Constantinescu2, Olivier Barret2, Matthias M. Herth1,3, Janus
H. Magnussen1, Szabolcs Lehel3, Agnete Dyssegaard1, Julie Colomb4, Thierry Billard4, Luc
Zimmer5, Gilles Tamagnan2, Gitte M. Knudsen1
1. Neurobiology Research Unit and Center for Integrated Molecular Imaging, Copenhagen University
Hospital, Rigshospitalet, Copenhagen, Denmark.
2. Invicro LLC, New Haven, CT, USA
3. PET and Cyclotron Unit, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark
4. Institute of Molecular Chemistry and Biochemistry, University of Lyon, CNRS, Lyon, France.
5. Lyon Neuroscience Research Centre, University of Lyon, CNRS, INSERM, Lyon Hospitals, Lyon,
France.
First author and Corresponding author:
Hanne D. Hansen, Ph.D., postdoctoral fellow
Neurobiology Research Unit and Center for Integrated Molecular Brain Imaging
Copenhagen University Hospital, Rigshospitalet
Blegdamsvej 9
DK-2100 Copenhagen, Denmark
Phone: (+45) 35 45 88 21
Fax: (+45) 35 45 67 13
Email:
hanne.d.hansen@nru.dk
Running title: Evaluation of [18F]2FP3 in pigs and NHPs
Keywords: [18F]2FP3, [3H]SB-269970, 5-HT7 receptor, PET, autoradiography
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/jlcr.3692
This article is protected by copyright. All rights reserved.

Abstract
So far, no suitable 5-HT7R radioligand exists for clinical PET imaging. [18F]2FP3 was first
tested in vivo in cats and the results were promising for further evaluations. Here, we evaluate
the radioligand in pigs and non-human primates (NHP). Furthermore, we investigate species
differences in 5-HT7R binding with [3H]SB-269970 autoradiography in post mortem pig, NHP,
and human brain tissue.
Specific binding of [18F]2FP3 was investigated by intravenous administration of the 5-HT7R
specific antagonist SB-269970. [3H]SB-269970 autoradiography was performed as previously
described.
[18F]2FP3 was synthesized in an overall yield of 35-45%. High brain uptake of the tracer was
found in both pigs and NHPs, however pretreatment with SB-269970 only resulted in decreased
binding of 20% in the thalamus, a 5-HT7R rich region. Autoradiography on post mortem pig,
NHP and human tissues revealed that specific binding of [3H]SB-269970 was comparable in
the thalamus of pig and NHP.
Despite the high uptake of [18F]2FP3 in both species, the binding could only be blocked to a
limited degree with the 5-HT7R antagonists. We speculate that the affinity of the radioligand
is too low for imaging the 5-HT7Rs in vivo and that part of the PET signal arises from targets
other than the 5-HT7R.
This article is protected by copyright. All rights reserved.

1. Introduction
Serotonin (5-hydroxytryptamine, or 5-HT) exerts its many psychophysiological functions
through its fourteen different receptors and the latest addition to the family is the serotonin 7
receptor (5-HT7R). 5-HT7R is associated with various physiological effects, including
serotonin-induced phase-shifting of the circadian rhythm1. The 5-HT7R may also be involved
in the pathophysiology of various psychiatric disorders such as anxiety and depression, as has
been reviewed by Matthys et al. 2. The accumulated evidence suggests that the 5-HT7R can be
a relevant drug target for the treatment of these psychiatric disorders. Although in vivo studies
of the cerebral 5-HT7R binding with positron emission tomography (PET) would improve the
understanding of its role in central nervous system disorders, no suitable PET radioligands for
imaging the 5-HT7R in humans exist.
Several selective 5-HT7R compounds have been developed3 and some have been
radiolabeled for use in radioligand binding assays and in vitro autoradiography studies. For
example, [3H]SB-269970 has been used for in vitro experiments in rodent, pig, marmoset, and
human brain tissue4–7. In recent years, there has been a large effort to identify a radioligand for
positron emission tomography (PET) imaging. PET is a very sensitive imaging tool that can be
applied to determine receptor availability and occupancy and can be used to inform critically
about drug discovery and development decisions8.
We and others have investigated a number of ligands from different compound classes
for the purpose of developing a 5-HT7R PET radioligand5,9–15. Of these, [11C]Cimbi-717
displayed the most promising results for selective 5-HT7R PET imaging in pigs5. A series of
18F-labeled compounds have been developed using SB-269970 as lead compound16–18 because
of its high affinity (KD = 1.3 nM) and high selectivity for the 5-HT7R19. One promising
analogue of SB-269970 is [18F]2FP3 (KD = 8.4 nM, human recombinant receptors in CHO
cells), which displayed promising results when evaluated in cats16. The aim of this project was
This article is protected by copyright. All rights reserved.

to evaluate [18F]2FP3 in pigs and non-human primates (NHP) for its potential as PET
radioligand for the 5-HT7R. Furthermore, we investigated species differences in 5-HT7R
binding in post mortem, pig, NHP and human brain tissue.
2. Methods
2.1 2FP3: Precursor and reference compound syntheses
Precursor (2NP3) and reference compound (2FP3) were synthesized as previously
described16.
2.2 Radiosynthesis of [18F]2FP3
2.2.1 Copenhagen University Hospital: To an aqueous [18F]fluoride solution, Kryptofix®222
(ABX GmbH, Germany), (15 mg), 15 µL potassium carbonate (1 N) and 1-mL acetonitrile
were added. The mixture was dried in a stream of nitrogen at 80 °C. The drying procedure was
repeated three times until the reaction mixture was absolutely dry. 4 mg (of nitro precursor
((R)-4-methyl-1-(2-(1-((2-nitrophenyl)sulfonyl)pyrrolidin-2-yl)ethyl)piperidine
(2NP3)
dissolved in 0.3 mL DMSO was added, and the reaction mixture heated at 150 °C for 10 min.
After dilution with 150 mL of water, the reaction mixture was purified by semi-preparative
high-performance liquid chromatography (HPLC) separation (LunaC18(2), 5µm, 250×10 mm;
Phenomenex;0.01 borax buffer (pH=9.21)/MeCN (30/70), flowrate: 6 mL; retention time
([18F]2FP3): ~ 600 sec). After diluting the collected HPLC fraction containing [18F]2FP3 with
water, the product was trapped on an activated Sep-PakW-C18-cartrige and subsequently
eluted with 3 mL of ethanol into 12 mL of phosphate buffered saline (100 mM, pH 7). Quality
control was performed to determine radiochemical purity and molar activity (LunaC18(2),
5µm, 150×4.6 mm; Phenomenex;0.01 borax buffer (pH=9.21)/MeCN (30/70), flowrate: 2 mL;
retention time [18F]2FP3: 3.01 min).
This article is protected by copyright. All rights reserved.

2.2.2 Molecular NeuroImaging LLC: [18F]fluoride solution (Cardinal Health, CT, USA) was
trapped onto an ion exchange column (Waters QMA light) before being eluted in the reactor
with a solution of Kryptofix®222 (ABX GmbH, Germany), (10 mg) and potassium carbonate
(1.2 mg) in aqueous acetonitrile (1 mL, acetonitrile/water 4/1 v/v). The mixture was dried in a
stream of helium and under vacuum at 95 °C for 4 min. Acetonitrile (1 mL) was added to the
reactor and the drying continued for 3 min. A second portion of acetonitrile (1 mL) was added,
and the drying was continued for 3 min before the reactor was cooled to 70 °C. A solution of
the nitro precursor (2NP3, 2 mg) dissolved in DMSO (1 mL) was added, and the reaction
mixture heated at 150 °C for 10 min. After dilution with water (2 mL) and HPLC eluent (2
mL), the reaction mixture was purified by semi-preparative HPLC (Agilent Eclipse XDB,
250x10 mm) using a mixture of methanol/water/triethylamine (70/30/0.1 v/v/v) at a flow rate
of 4 mL/min. The retention time of [18F]2FP3 was 750 sec. After diluting the collected HPLC
fraction containing [18F]2FP3 with water (15 mL), the product was trapped on a Waters tC18
light Sep-Pak cartridge and eluted with ethanol (1 mL) into saline (10 mL). The cartridge was
rinsed with saline (4 mL), adding the rinse to the product. Quality control was performed to
determine radiochemical purity and molar activity (Waters XBridge C18, 5µm, 150×4.6 mm,
eluting with methanol/water/triethylamine (70/30/0.1 v/v/v) at a flow rate of 1 mL/min). The
retention time of [18F]2FP3 was 10 min.
2.3 In vivo imaging in pigs
Three female domestic pigs (crossbreed of Landrace x Yorkshire x Duroc, mean weight ± S.D.,
23 ± 0.6 kg) were used for in vivo PET imaging. All animal procedures were approved by the
Danish Council for Animal Ethics (journal no. 2012-15-2934-00156).
[18F]2FP3 was synthesized according to the method described in section 2.2.1. The tracer was
given as intravenous (i.v.) bolus and the injected dose was 365 ± 88 MBq (mean ± SD, n = 3).
This article is protected by copyright. All rights reserved.

The pigs were scanned for 150 min in list mode. Ninety minutes after the injection of the
radioligand, a challenge of 1 mg/kg SB-269970 (Tocris Bioscience) was given i.v. In one pig,
a constant infusion of 1 mg/kg/h was started 30 min prior to the injection of [18F]2FP3 resulting
in a total infusion time of 3 hours.
See supplementary information for more details on animal procedure, blood sampling,
metabolite analysis, reconstruction and pre-processing of PET data.
2.4 In vivo imaging in NHP
Two male rhesus monkeys (Macaca mulatta), one male and one female, were used as research
subjects. These animals were housed and imaged following identical procedures at the Yale
University School of Medicine in New Haven, Connecticut. PET imaging was carried out under
institutional animal care protocols complying with Federal regulations. Animal care approval
and oversight for this study was provided by the Yale University Institutional Animal Care and
Use Committee.
[18F]2FP3 was synthesized according to section 2.2.2 and was administered via iv bolus over
3 minutes and the injected doses were 182.7 ± 65 MBq (mean ± SD, n = 3). Following
radiotracer administration, list mode emission data was collected over 180 min on the
microPET FOCUS 220 camera (Siemens Medical Solutions, Inc.). A transmission scan with a
Ge-68 source prior to injection of radiotracer was used for attenuation correction. In one animal
SB-269970 (3 mg/kg) was administrated i.v. over 15 min starting 70 min before the injection
of the radioligand. In a different experiment, WAY100635 (1 mg/kg) was administrated i.v.
over 5 min starting 35 min before the injection of the radioligand.
See supplementary information for more details on animal procedure, MR imaging,
reconstruction and pre-processing of PET data.
This article is protected by copyright. All rights reserved.

2.5 Kinetic modelling
The PET imaging data for both pig and NHP were analysed with the Logan graphical analysis
(LGA), using the metabolite corrected arterial plasma concentration to calculate the total
distribution volume (VT).
2.6 In vitro autoradiography
For autoradiography, smaller pieces of one non-human primate brain were sliced on a
HM500OM cryostat (Microm Intl GmbH) in 20 µm coronal sections except for cerebellum
which was sliced in the sagittal plane. Human temporal cortical tissue was obtained from
anterior temporal lobectomies in patients with medically intractable temporal lobe epilepsy
with hippocampal onset, aged 19 and 55 years. Written informed consent was obtained from
both patients before surgery. The study was approved by the Ethical Committee in the Capital
Region of Denmark (H-2-2011-104) and performed in accordance with the Declaration of
Helsinki. The tissue was dissected and immediately frozen on dry ice and stored at -80°C until
use. Human brain tissue was sliced in 12 µm coronal sections. Autoradiography was performed
on 3 to 6 adjacent sections for both NHP and human brain tissue. See supplementary
information for details of the autoradiographic protocol.
3. Results
3.1 Radiosynthesis of [18F]2FP3
Injection-ready [18F]2FP3 for pig studies was prepared according to section 2.2.1 in 40-65 min
from the end of bombardment with a radiochemical yield of 0.4-2 GBq (Fig. 1). The
radiochemical yield was 35.2% (n = 3). The molar activity at the time of injection (TOI) was
24 ± 23 GBq/µmol (n = 3) resulting in injected masses of 12.5 ± 12.4 μg. For NHP studies the
radiosynthesis of [18F]2FP3 (see section 2.2.2) was automated using a commercial
This article is protected by copyright. All rights reserved.

radiofluorination device (GE Tracerlab FXFN) to afford high radiochemical yields (43.8%, n
= 3) and specific activities of 270.4 ± 44.2 GBq/µmol (n = 3) resulting injected masses of 0.116
± 0.110 μg. Both radiochemical labs confirmed the identity of the product by co-injection with
a sample of reference compound, which under the same elution conditions showed an identical
retention time. The compound was stable over 4 h and the radiochemical purity > 98 %.
3.2 In vivo studies of [18F]2FP3 in pigs
[18F]2FP3 readily entered the pig brain and displayed reversible tracer kinetics (Fig. 2A and
2D) but with relatively slow wash-out of the tracer. The distribution of the radioligand was to
some degree in accordance with the 5-HT7R distribution within pigs5 with high uptake in the
thalamus and lowest uptake in the cerebellum. However, an unexpected higher uptake was
found in the striatum, a region that has a low density of 5-HT7Rs. In pigs the binding of
[18F]2FP3 was challenged by administrating the 5-HT7R specific antagonist SB-269970 (1
mg/kg) 90 minutes after the injection of the radioligand (Fig. 2A). Administration of SB-
269970 resulted in an increase in plasma radioactivity, suggesting that SB-269970 displaces
[18F]2FP3 from peripheral receptor sites but no significant change was observed in any of the
brain TACs. The specificity of radioligand binding was also investigated by starting a constant
infusion of SB-269970 (1 mg/kg/h) 30 min prior to the injection of [18F]2FP3 (Fig. 2B) and
lasting throughout the scan time (total infusion time 3h). This resulted in a slightly higher
uptake of radioligand across all brain regions, but with visually similar kinetics compared to
the baseline experiment. Although a higher uptake was seen after pretreatment with SB-
269970, changes in radioactivity in the blood and metabolism of the radioligand should also be
considered. Therefore, we quantified the radioligand uptake in the brain using the Logan
graphical model (LGA). Kinetic modelling revealed a decrease in binding in all investigated
regions of the brain (see Supplementary Table 1) and the occupancy of SB-269970 in the pig
This article is protected by copyright. All rights reserved.

brain was determined to 35 % by the occupancy plot (Fig 2C). Additionally, the occupancy
plot determined the non-displaceable distribution volume to 6.29 ml/cm3
.
3.3 In vivo studies of [18F]2FP3 in NHP
After intravenous injection, [18F]2FP3 rapidly entered the NHP brain (Fig. 3A). The
radioligand accumulation was fairly uniform in the NHP brain but higher initial uptake of
radioligand was found in cerebellum and occipital cortex and lower accumulation was found
in the thalamus (Fig. 3B). The tracer kinetics of [18F]2FP3 appeared fast with peak uptake after
approximately 10 min followed by fast wash-out from all investigated brain regions. The
specificity of binding of [18F]2FP3 was also investigated in NHPs by administrating SB-
269970 (3 mg/kg) and the 5-HT1AR selective WAY100635 (1 mg/kg). In both blocking
experiments, the TACs had slightly faster washout rates consistent with a modest blocking
effect of the drugs (Fig. 3C and 3D). Quantification of binding after administration of SB-
269970 revealed a decrease in VT of 13 % in the thalamus and of 20 % in the temporal cortex
(Supplementary Table 2). Administration of WAY100635 resulted in similar blocking effects,
with decreases in VT values: 20 % in the thalamus and 19 % in the temporal cortex. It was not
possible to reliably estimate occupancy of the two drugs and non-displaceable binding of
[18F]2FP3 in the NHP due to the low blocking effects of drugs.
3.4 Assessment of metabolic profile
Finally, the systemic metabolism of [18F]2FP3 was determined in vivo. The metabolism of
[18F]2FP3 in pigs was fast, with only 50 % of parent compound remaining after 10 min (Fig.
4A). When a challenge of 1 mg/kg SB-269970 was given after 90 minutes of scan time, the
plasma parent compound fraction as well as activity in whole blood and plasma increased,
suggesting that [18F]2FP3 was displaced by SB-269970 at peripheral receptor sites.
Pretreatment with SB-269970 did not influenced the rate of metabolism of the parent
compound significantly.
This article is protected by copyright. All rights reserved.

In NHPs, the rate of metabolism of [18F]2FP3 was slower than in pigs, with 50 % parent
compound remaining after 60 min (Fig. 4B). Contrary to what was observed in pigs, the
pretreatment with SB-269970 lead to an increased metabolic breakdown of [18F]2FP3 in NHPs.
This effect was not observed with pre-administration of WAY100635.
In both species, the metabolites detected were more polar than the parent compound and thus
unlikely to cross the blood-brain barrier (Supplementary Figure 1).
3.6 In vitro autoradiography
The specific binding (SB) was determined with [3H]SB-269970 autoradiography in the human,
NHP, and pig temporal cortex: 18.9 ± 4.99 (human), 21.5 ± 4.32 (NHP), and 41.0 ± 3.03 (pig)
fmol/mg tissue equivalent (TE). SB was also determined in the thalamus and cerebellum of
NHP and pig tissue: 32.3 ± 3.61 (NHP, thalamus), and 45.6 ± 4.99 0.19 (pig, thalamus), 14.3
± 0.93 (NHP, cerebellum) and 13.4 ± 2.22 (pig, cerebellum) fmol/mg TE. The results from pig
brain tissue are taken from Hansen et al. (2014)5 Comparing the specific binding of [3H]SB-
269970 and the VT of [18F]2FP3 revealed no apparent relationship between the in vitro and in
vivo measures for the pig (Fig. 5C). No relationship was found for the NHP data either,
although only three different regions could be compared.
4. Discussion
In pigs, we observed a high uptake of the radioligand and a relative slow wash-out from the
brain, which could indicate specific binding to the target. However, pretreatment with SB-
269970 (1 mg/kg/h for 3h) resulted in a [18F]2FP3 occupancy of 35% whereas the same dose
of SB-269970 resulted in 60% occupancy of the 5-HT7R with [11C]Cimbi-717 in pigs5,
suggesting that [18F]2FP3 has either off-target binding or high non-specific binding. The NHPs
also had high brain uptake of [18F]2FP3 but with faster tracer kinetics than in the pig. Further,
This article is protected by copyright. All rights reserved.

the regional separation between the TACs was minimal and pretreatment with SB-269970 and
with WAY100635 resulted in 13-20% decrease in VT values, respectively. This supports that
in the NHP, [18F]2FP3 has limited but measurable specific binding to the 5-HT7R and 5-HT1AR
in vivo. The binding to the 5-HT1AR was unexpected, given that 2FP3 does not have any
appreciable affinity for the 5-HT1AR in vitro16. Although WAY100635 is considered highly
selective for the 5-HT1AR, it does retain some affinity for the 5-HT7R (248 nM)20. With the
administered dose (1 mg/kg), it is therefore possible that WAY100635 also occupied a fraction
of the 5-HT7Rs. The mismatch between regional distribution volumes of [18F]2FP3 and the in
vitro autoradiography outcomes support that [18F]2FP3 has off-target binding. We can only
speculate in the nature of the potential in vivo off-target binding as affinity was only tested for
the 5-HT7, 5-HT1A and 5-HT6 receptors and with apparent affinity for the two latter targets16.
Similarly, low affinity for the D2 (790 nM), D3 (631 nM) and 5-HT6 (3.21 µM) receptors was
identified of the closely related compound 12 in the publication by Lovell et al.19. The blocking
effect obtained in these studies contrasts the results obtained in cats, where the TACs strongly
indicated a blocking effect after pre-administration of SB-269970 (5 mg/kg). However, since
the arterial input function was not determined in these experiments, the binding before and
after pretreatment with SB-269970 was not quantified16.
There could be species differences to the amino acid sequence in the 5-HT7R, however
comparing the amino acid sequence using NCBI BLAST of the predicted (partial) sequence of
the cat 5-HT7R (XP_006938099) and the pig 5-HT7R (XP_005671333) resulted in 99 %
sequence identity. Comparing the pig and the NHP (XP_001143930) sequences, however,
resulted in 96 % sequence identity. The exact function of these amino acids in relation to loops,
helices, or more importantly the binding pocket is not known. If these amino acids interact with
ligands, it could potentially result in a large difference in binding between species.
This article is protected by copyright. All rights reserved.

The binding potential is the product of receptor density (Bmax) and the in vivo affinity of
the radioligand (1/KD). In humans, the 5-HT7R is, along with the 5-HT4R, one of the least
abundant 5-HT receptors7,21. Thus, to determine the amount of 5-HT7Rs in the NHP and
compare this directly to amounts in human brain, in vitro autoradiography was performed on
post mortem NHP and human tissue and directly compared to that of post mortem pig.
Generally, the level of binding of the 5-HT7R radioligand, [3H]SB-269970, was higher in the
thalamus and temporal cortex of the pig compared to NHP and human. In the cerebellum, the
levels of specific binding were comparable between pig and NHP. The specific binding in the
human temporal cortex were similar to the NHP temporal cortex. However, the amount of
specific binding found in all regions in the present study was 10-fold higher than those reported
by Varnäs et al.7 using the same radioligand, but with small differences in the experimental
protocol. Varnäs et al. report 5-7-fold difference in binding between the temporal cortex and
thalamus, and if we extrapolate this ratio to our own data, the predicted specific binding in the
human thalamus would be in the range of 116-162 fmol/mg TE. Unfortunately, we did not have
access to thalamic tissue from a post mortem human brain to confirm this hypothesis. PET
imaging with [11C]Cimbi-717 in pigs demonstrated that specific binding in vivo could be
detected at 5-HT7R levels as low as 20 fmol/mg5. To obtain a high binding potential, a high
affinity of the radioligand for the target is necessary, especially if the target in question has low
receptor density (Bmax). The affinity of [18F]2FP3 (8.4 nM) was determined by competition
binding experiments in over-expressing h5-HT7R CHO cells16. Unfortunately, the KD has not
been determined in rats and cats16,18 and neither did we determine the affinity in the current
study. We speculate that the affinity of [18F]2FP3 may be too low for the 5-HT7R and this could
be one reason why we saw relatively low occupancies with SB-269970, both in pigs and in
NHP. Lastly, selectivity of the radioligand is also an important factor, especially in the case of
developing ligands for the 5-HT7R, which has a relatively low Bmax in the brain compared to
This article is protected by copyright. All rights reserved.

other targets. As described by Herth et al., a 100-200-fold difference in KD is needed to ensure
less than 10 % of the signal from e.g. the α1 adrenergic receptor in the thalamus13. The full
selectivity profile of 2FP3 has not been established, and thus we cannot rule out that some of
the PET signal arises from other targets in the brain. The high uptake in the striatum (which
has low abundance in 5-HT7s and 5-HT1ARs) of both pigs and NHP may be another indication
of off-target binding.
A limitation to the studies with [18F]2FP3 in pigs was varying specific activities resulting
in injected masses ranging from 2.6 - 26 µg. However, we found that baseline VT values in two
different pigs were very similar despite the large difference in injected mass, it is unlikely that
the injected radioligand has caused a self-blockade. Nevertheless, the occupancy plot (Fig 2C)
was made with the baseline experiment with the highest molar activity. Another limitation of
the studies in pigs is that the baseline and blocking experiments was not performed in the same
animal, due to the long half-life of the radioligand. The baseline and blocking experiment were,
however, carried out in the same animal.
Recently, 2FP3 served as a lead compound for further development of radioligands for
the 5-HT7R with higher affinity for the receptor17. Unfortunately, these chemical modifications
rendered the radioligands substrates for the P-glycoprotein and thus the brain uptake in rats of
the two of the radiolabelled compounds was rather low. Due to species differences of the P-
glycoprotein it is possible that the evaluation of the most promising of these new compounds,
2F3P3, will have brain uptake in the pig and NHP brain and that the increased affinity will
increase the chances of imaging the 5-HT7Rs in vivo.
5. Conclusion
Although initial [18F]2FP3 experiments in cats showed some promise for the radioligand being
a good 5-HT7R PET radioligand we here show in pigs and in NHP that the brain distribution
This article is protected by copyright. All rights reserved.

of the radioligand does not match the 5-HT7R distribution determined in vitro. Furthermore,
we could only detect limited effects of blocking the 5-HT7R in these two species. We speculate
that the affinity of the radioligand is too low for imaging the 5-HT7Rs in vivo and that part of
the PET signal arises from targets other than the 5-HT7R and 5-HT1AR.
6. Acknowledgements
The authors wish to thank Charlotte K. Fink and Karsten P. Hammelev at the Department of
Experimental Medicine (University of Copenhagen) for preparing the animals for the
experiments. Bente Dall is acknowledged for technical expertise with the HRRT and PET
experiments. The John & Birthe Meyer Foundation’s financial support in acquisition of the
HRRT and Cyclotron system is greatly appreciated. Jacob Hooker is thanked for providing the
non-human primate tissue for autoradiography.
7. Compliance with ethical standards
7.1 Conflict of interest
GMK has been an invited lecturer at Pfizer A/S, worked as a consultant and received grants
from H. Lundbeck A/S and is a stock holder of Novo Nordisk/Novozymes.
7.2 Author contribution
Study conception and design: HDH, MMH, GMK, CCC, OB, GT. Acquisition of data: HDH,
MMH, AD, JHM, OB, CCC, SL. Analysis and interpretation of data: HDH, MMH, OB,
CCC, GMK, GT. Writing the manuscript: HDH, MMH, GMK, CCC, OB. All authors revised
the manuscript.
7.3 Funding
The Lundbeck Foundation has funded Center for Integrated Molecular Brain Imaging (R90-
A7722). The work was also supported by University of Copenhagen.
This article is protected by copyright. All rights reserved.

8. References
1.
Lovenberg TW, Baron BM, de Lecea L, et al. A novel adenylyl cyclase-activating
serotonin receptor (5-HT7) implicated in the regulation of mammalian circadian
rhythms. Neuron. 1993;11(3):449-458. doi:10.1016/0896-6273(93)90149-L.
Matthys A, Haegeman G, Van CK, et al. Role of the 5-HT7 Receptor in the Central
Nervous System: from Current Status to Future Perspectives. Mol Neurobiol.
2011;43(Copyright (C) 2011 American Chemical Society (ACS). All Rights
Reserved.):228-253. doi:10.1007/s12035-011-8175-3.
2.
3.
4.
Pittala V, Salerno L, Modica M, Siracusa MA, Romeo G. 5-HT7 receptor ligands:
recent developments and potential therapeutic applications. Mini Rev Med Chem.
2007;7(9):945-960.
Horisawa T, Ishiyama T, Ono M, Ishibashi T, Taiji M. Binding of lurasidone, a novel
antipsychotic, to rat 5-HT7 receptor: Analysis by [3H]SB-269970 autoradiography.
Prog Neuropsychopharmacol Biol Psychiatry. 2013;40(0):132-137.
doi:http://dx.doi.org/10.1016/j.pnpbp.2012.08.005.
5.
6.
7.
8.
Hansen HD, Herth MM, Ettrup A, et al. Radiosynthesis and In Vivo Evaluation of
Novel Radioligands for PET Imaging of Cerebral 5-HT7 Receptors. J Nucl Med.
2014;55(4):640-646. doi:10.2967/jnumed.113.128983.
Thomas DR, Atkinson PJ, Hastie PG, Roberts JC, Middlemiss DN, Price GW. [3H]-
SB-269970 radiolabels 5-HT7 receptors in rodent, pig and primate brain tissues.
Neuropharmacology. 2002;42(1):74-81. doi:10.1016/S0028-3908(01)00151-4.
Varnas K, Thomas DR, Tupala E, et al. Distribution of 5-HT7 receptors in the human
brain: A preliminary autoradiographic study using [3H]SB-269970. Neurosci Lett.
2004;367(3):313-316. doi:10.1016/j.neulet.2004.06.025.
Piel M, Vernaleken I, Rösch F. Positron Emission Tomography in CNS Drug
This article is protected by copyright. All rights reserved.

Discovery and Drug Monitoring. J Med Chem. 2014;57(22):9232-9258.
doi:10.1021/jm5001858.
9.
Zhang M-R, Haradahira T, Maeda J, et al. Synthesis and preliminary PET study of the
5-HT7 receptor antagonist [11C]DR4446. J Label Compd Radiopharm.
2002;45(10):857-866. doi:10.1002/jlcr.606.
10. Herth MM, Hansen HD, Ettrup A, et al. Synthesis and evaluation of [(11)C]Cimbi-806
as a potential PET ligand for 5-HT(7) receptor imaging. Bioorg Med Chem.
2012;20(14):4574-4581. doi:10.1016/j.bmc.2012.05.005.
11. Andersson JD, Leopoldo M, Finnema S, et al. Development of a PET radioligand for
the 5-HT7 receptor. Preliminary characterization using autoradiography. In: Vol ;
2010:224.
12. Lacivita E, Niso M, Hansen HD, et al. Design, synthesis, radiolabeling and in vivo
evaluation of potential positron emission tomography (PET) radioligands for brain
imaging of the 5-HT7 receptor. Bioorg Med Chem. 2014;22(5):1736-1750.
doi:10.1016/j.bmc.2014.01.016.
13. Herth MM, Volk B, Pallagi K, et al. Synthesis and In Vitro Evaluation of Oxindole
Derivatives as Potential Radioligands for 5-HT 7 Receptor Imaging with PET. ACS
Chem Neurosci. 2012;3(12):1002-1007. doi:10.1021/cn3001137.
14. Herth MM, Andersen VL, Hansen HD, et al. Evaluation of 3-Ethyl-3-
(phenylpiperazinylbutyl)oxindoles as PET Ligands for the Serotonin 5-HT 7 Receptor:
Synthesis, Pharmacology, Radiolabeling, and in Vivo Brain Imaging in Pigs. J Med
Chem. 2015;58(8):3631-3636. doi:10.1021/acs.jmedchem.5b00095.
15. Hansen HD, Andersen VL, Lehel S, et al. Labeling and preliminary in vivo evaluation
of the 5-HT7 receptor selective agonist [11C]E-55888. Bioorg Med Chem Lett.
2015;25(9):1901-1904. doi:10.1016/j.bmcl.2015.03.039.
This article is protected by copyright. All rights reserved.

16. Lemoine L, Andries J, Le Bars D, Billard T, Zimmer L. Comparison of 4 Radiolabeled
Antagonists for Serotonin 5-HT7 Receptor Neuroimaging: Toward the First PET
Radiotracer. J Nucl Med. 2011;52(11):1811-1818. doi:10.2967/jnumed.111.089185.
17. Colomb J, Becker G, Forcellini E, et al. Synthesis and pharmacological evaluation of a
new series of radiolabeled ligands for 5-HT7 receptor PET neuroimaging. Nucl Med
Biol. 2014;41(4):330-337. doi:10.1016/j.nucmedbio.2014.01.008.
18. Andries J, Lemoine L, Le Bars D, Zimmer L, Billard T. Synthesis and biological
evaluation of potential 5-HT7 receptor PET radiotracers. Eur J Med Chem.
2011;46(8):3455-3461. doi:S0223-5234(11)00372-2
[pii]10.1016/j.ejmech.2011.05.010.
19. Lovell PJ, Bromidge SM, Dabbs S, et al. A novel, potent, and selective 5-HT7
Antagonist: (R)-3-(2-(2-(4-Methylpiperidin-1-yl)-ethyl)pyrrolidine-1-sulfonyl)phenol
(SB-269970). J Med Chem. 2000;43(Figure 1):342-345. doi:10.1002/bit.260190912.
20. Louie AS, Vasdev N, Valliant JF. Preparation, Characterization, and Screening of a
High Affinity Organometallic Probe for α-Adrenergic Receptors. J Med Chem.
2011;54(9):3360-3367. doi:10.1021/jm2001162.
21. Varnäs K, Halldin C, Hall H. Autoradiographic distribution of serotonin transporters
and receptor subtypes in human brain. Hum Brain Mapp. 2004;22(3):246-260.
doi:10.1002/hbm.20035.
This article is protected by copyright. All rights reserved.

Fig. 1 Radiosynthesis of [18F]2FP3.
This article is protected by copyright. All rights reserved.

Fig. 2 A) Representative TACs from an experiment with a within-scan challenge of SB-269970
at 90 min. B) Representative TACs from an experiment with pre-administration of SB-269970
(1 mg/kg/h for 3h). C) Occupancy plot for [18F]2FP3 determined by two scans in two different
pigs: one pig with a baseline scan (VT (baseline)) and another pig scanned after pre-
administration of SB-269970 (1 mg/kg/h, VT (drug)). D) Summed PET image (0-90 min) of
the baseline biodistribution of [18F]2FP3 in the pig brain.
This article is protected by copyright. All rights reserved.

Fig. 3 A) Summed PET image of the baseline biodistribution of [18F]2FP3 in the NHP brain.
B) Representative baseline TACs for [18F]2FP3. C) Representative TACs after pre-
administration of SB-266970 (3 mg/kg). D) TACs after pre-administration of WAY100635 (1
mg/kg).
This article is protected by copyright. All rights reserved.

Fig. 4 HPLC analysis of parent compound ([18F]2FP3) in plasma in pigs (A) and NHP (B) as
a function of time after intravenous injection. Filled circles represent baseline (bsl)
conditions, open circles represent an experiment where the animal has been pretreated with
SB-269970, and open triangles represent an animal pretreated with WAY-100635. In pigs,
open squares represent the parent compound in plasma after a challenge (cha) with SB-
269970 ninety min after the injection of the radioligand. In pigs, experiments with 0-90 min
baseline and 90-120 min challenge phase, were repeated and values are presented for each
experiment, as indicated with different colors. All other experiments were performed once.
This article is protected by copyright. All rights reserved.

Fig. 5 A) Representative sections of 5-HT7R autoradiography in pigs (a+b), non-human
primates (c+d), and human (e). Total binding in left panel, non-specific binding determined
with 10 µM SB-258719 in middle panel, and indication of ROI in right panel. B) Bar graph
showing the amount of specific binding found in human, non-human primate (NHP) and pig
post mortem brain tissue. C) Comparison of the binding of [3H]SB-269970 and [18F]2FP3 in
pig and NHP tissue. All [3H]SB-269970 data for pig tissue is taken from Hansen et al.5 Cb:
cerebellum, Ctx: cortex, Hip: Hippocampus, Str: striatum, Tha: Thalamus, Tcx: temporal
cortex.
This article is protected by copyright. All rights reserved.