Selective binding to monoamine oxidase A: In vitro and in vivo evaluation of 18F-labeled β-carboline derivatives

Article abstract of DOI:10.1016/j.bmc.2014.11.040

In this study we synthesized four different ¹⁸F-labeling precursors for the visualization of the monoamino oxidase A using harmol derivatives. Whereas two are for prosthetic group labeling using [¹⁸F]fluoro-d₂-methyl tosylate and 2-[¹⁸F]fluoroethyl-tosylate, the other three precursors are for direct nucleophilic ¹⁸F-labeling. Additionally the corresponding reference compounds were synthesized. The syntheses of [¹⁸F]fluoro-d₂-methyl-harmol and 2-[¹⁸F]fluoroethyl-harmol were carried out using harmol as starting material. For direct nucleophilic ¹⁸F-labeling of the tracers carrying oligoethyled spacers (PEG), a toluenesulfonyl leaving group was employed. The radiolabeling, purification and formulation for each tracer was optimized and evaluated in vitro and in vivo. Stability tests in human serum showed that all tracers were stable over the observation period of 60 min. μPET studies using of the synthesized tracers revealed that the tracers carrying PEG spacers showed no sufficient brain uptake. Consequently, the ¹⁸F-fuoro alkylated tracers [¹⁸F]fluoro-d₂-methyl-harmol and 2-[¹⁸F]fluoroethyl-harmol were further evaluated showing SUVs in the brain of 1.0 ± 0.2 g/mL and 3.4 ± 0.5 g/mL after 45 min, respectively. In blockade studies the selectivity and specificity of both tracers were demonstrated. However, for [¹⁸F]fluoro-d₂-methyl-harmol a rapid washout from the brain was also observed. In vitro binding assays revealed that 2-[¹⁸F]fluoroethyl-harmol (IC₅₀ = 0.54 ± 0.06 nM) has a higher affinity than the ¹⁸F-fluoro-d₂-methylated ligand (IC₅₀ = 12.2 ± 0.6 nM), making 2-[¹⁸F]fluoroethyl-harmol superior to the other evaluated compounds and a promising tracer for PET imaging of the MAO A.

Full text of DOI:10.1016/j.bmc.2014.11.040

Bioorganic & Medicinal Chemistry 23 (2015) 612–623
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Selective binding to monoamine oxidase A: In vitro and in vivo
evaluation of 18F-labeled b-carboline derivatives
a
c
b
Hanno Schieferstein a, Markus Piel a, , Friderike Beyerlein , Hartmut Lüddens , Nicole Bausbacher ,
⇑
Hans-Georg Buchholz b, Tobias L. Ross a, Frank Rösch a
a Institute of Nuclear Chemistry, Johannes Gutenberg-University, 55128 Mainz, Germany
b Department of Nuclear Medicine, University Medical Center Mainz, 55131 Mainz, Germany
c Department of Psychiatry and Psychotherapy, University Medical Center Mainz, 55131 Mainz, Germany
a r t i c l e i n f o
a b s t r a c t
In this study we synthesized four different 18F-labeling precursors for the visualization of the monoamino
oxidase A using harmol derivatives. Whereas two are for prosthetic group labeling using [18F]fluoro-d2-
methyl tosylate and 2-[18F]fluoroethyl-tosylate, the other three precursors are for direct nucleophilic 18F-
labeling. Additionally the corresponding reference compounds were synthesized. The syntheses of
Article history:
Received 26 July 2014
Revised 4 November 2014
Accepted 27 November 2014
Available online 8 December 2014
[
18F]fluoro-d2-methyl-harmol and 2-[18F]fluoroethyl-harmol were carried out using harmol as starting
material. For direct nucleophilic 18F-labeling of the tracers carrying oligoethyled spacers (PEG), a toluene-
sulfonyl leaving group was employed. The radiolabeling, purification and formulation for each tracer was
optimized and evaluated in vitro and in vivo. Stability tests in human serum showed that all tracers were
Keywords:
b-Carboline
Monoamine oxidase-A
Fluorine-18
stable over the observation period of 60 min. lPET studies using of the synthesized tracers revealed that
the tracers carrying PEG spacers showed no sufficient brain uptake. Consequently, the 18F-fuoro alkylated
tracers [18F]fluoro-d2-methyl-harmol and 2-[18F]fluoroethyl-harmol were further evaluated showing
SUVs in the brain of 1.0 0.2 g/mL and 3.4 0.5 g/mL after 45 min, respectively. In blockade studies
the selectivity and specificity of both tracers were demonstrated. However, for [18F]fluoro-d2-methyl-
harmol a rapid washout from the brain was also observed. In vitro binding assays revealed that 2-[18F]
fluoroethyl-harmol (IC50 = 0.54 0.06 nM) has a higher affinity than the 18F-fluoro-d2-methylated ligand
(IC50 = 12.2 0.6 nM), making 2-[18F]fluoroethyl-harmol superior to the other evaluated compounds and
a promising tracer for PET imaging of the MAO A.
lPET-study
Positron emission tomography
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
amine neurotransmitters and regulation of intracellular amine
concentration. MAO is related to some psychiatric and neurode-
In 1928, Mary Hare isolated an enzyme1 called monoamine
oxidase (MAO), which is responsible for the deamination of mono-
amines.2,3 This flavin-containing enzyme, located in the outer
mitochondrial membrane,4 can be divided into two subtypes,
MAO A and MAO B with 70% of amino acid sequence homology.5,6
The distribution of MAO varies throughout the human body and in
animal models. Both isoforms occur in neuronal and non-neuronal
cells7, but differ in their substrate specifity.8–11 MAO A degradates
noradrenaline and serotonin, while MAO B is responsible for the
deamination of phenethylamine and benzylamine (cf. Fig. 1).12
Dopamine and tyramine are metabolized by both isoforms.4
Thus, the monoamine oxidases are involved in the protection of
neurons from exogenous amines, termination of the actions of
generative disorders, such as depression. Its role in Parkinson’s
⇑
Corresponding author. Tel.: +49 6131 3925701; fax: +49 6131 3926606.
E-mail address: piel@uni-mainz.de (M. Piel).
Figure 1. Proposed pathways of the degradation reactions of the MAO is involved
in.
http://dx.doi.org/10.1016/j.bmc.2014.11.040
0968-0896/Ó 2014 Elsevier Ltd. All rights reserved.

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
613
chromatography was performed using Acros silica gel (0.040–
0.063 nm). For thin-layer chromatography (TLC) Merck silica gel
60 F254 plates were used for detection at 254 nm. The micro-
wave-supported syntheses were carried out using a commercially
available CEM Discover. Semipreparative HPLC was performed on
a Dionex P680A pump, a Raytest NaI scintillation counter (Gabi)
and a Dionex UVD 170U (254 nm) absorbance detector using a
Figure 2. Commonly used 11C-labeled MAO A tracers.
set of Phenomenex Synergi C12 Max-RP columns (4 lm,
disease is well established.13–16 Depression is still under specific
circumstances treated with MAO inhibitors, though this therapy
approach leads to serious side effects like the ‘cheese reaction’.17–19
Positron emission tomography (PET), as a non-invasive method
to visualize enzyme concentrations in vivo, is a very efficient way
to determine MAO levels in the brain. Hence, several 11C-labeled
tracers were developed to map MAO levels, like [11C]clorgyline,
250 ꢀ 10 mm). Dionex Chromeleon software was used for UV-data
analysis and Raytest Gina star software for radioactivity detection.
For the purification all tracers a stepwise gradient elution system
composed of 50 mM ammonium formate buffer as solvent A and
acetonitrile as solvent B (initial A/B = 70:30, 7 min 70:30, 15 min
60:40, 22 min 50:50, 28 min 30:70), flow rate of 5 mL/min, was
used. Radioanalytical HPLC was performed using an isocratic sys-
tem of 65% acetonitrile and 35% 50 mM ammonium formate buffer
with a flow rate of 1 mL/min.
[
11C]befloxatone or [11C]harmine (Fig. 2).20–22
All of them have their drawbacks, like short biological half-lives,
sophisticated syntheses or species differences in binding.23,24
Although harmine, a reversible MAO A inhibitor with a Ki value
of 2 nM, shows an extensive metabolization in plasma, it is the
most often used tracer for PET-studies.25 Thus it would be desir-
able to obtain a 18F-fluorinated analogue of [11C]harmine, due to
the better isotopic properties of 18F. In a previous study, Blom
et al. synthesized fluoroalkylated and fluoropegylated harmine
derivatives by varying the fluorinated moieties at the phenolic
hydroxyl function.21,23,26 Initial autoradiographic experiments of
rat brain slices have been performed showing promising results,
like decreased non-specific binding in the brain, by using the 18F-
labeled harmine derivatives.27
2.1. Synthesis of reference compounds and precursors
2.1.1. Synthesis of harmol (2)
The demethylation of harmine was performed in accordance to
Bergström et al.26 In brief, to a solution of 2 g (9.5 mmol) harmine
(1) in 60 mL of glacial acid 12.5 mL of hydrobromic acid were
added slowly and refluxed for 24 h. The crude reaction mixture
was evaporated to dryness and suspended in water. After addition
of 1 M potassium hydroxide solution the precipitate dissolved
completely. The aqueous phase was washed two times with
dichloromethane and two times with ether and the final product
precipitated after addition of 15 mL of a saturated ammonium
chloride-solution. The precipitate was isolated by filtration and
subsequently freeze-dried to give 1.55 g (7.8 mmol, 82%) of a pale
green solid. 1H NMR (300 MHz, CDCl3) d [ppm] = 11.419 (s, 1H),
8.143 (d, 1H), 8.047 (d, 1H), 7.793 (d, 1H), 7.003–6.806 (dd, 1H),
2.788 (s, 3H). ESI-MS (ESI+): m/z: 199.09 ([M+1]+).
For the synthesis of 18F-fluorinated analogues of 11C-methyl-
ated tracers different strategies can be followed. One of these,
the 18F-fluoroethylation, is a convenient and very often used
approach, although it may result in different biological properties
of the 18F-fluoroethylated tracer due to steric or lipophilic
differences compared to a 11C-methylated tracer.28,29 Recently,
the 18F-fluoromethylation has become an interesting alternative
to the 18F-fluoroethylation, since one major drawback of this
method, the usually occurring fast metabolism of the 18F-fluorom-
ethylated compounds, can be often avoided by an alternative
approach. By using [18F]fluoro-d2-methyl derivatives as labeling
agents, the metabolism of the tracer is reduced, since the
carbon-deuterium bond is six to ten times stronger than the car-
bon-hydrogen bond.30,31 Furthermore, in a previous study in our
group a reliable and automated radiosynthesis of [18F]fluoro-d2-
methyl tosylate ([18F]F-d2-MTos) was successfully developed,
resulting in short reaction times, high radiochemical yields and a
high radiochemical purity for this prosthetic group.32
2.1.2. Synthesis of fluoro-d2-methyl-harmol (3)
200 mg (1 mmol) of harmol (2) were dissolved in dry acetoni-
trile under Schlenck conditions. Then potassium tert-butoxide
(112.2 mg, 1 mmol) was added in one portion and the solution stir-
red for 15 min at RT. 137 lL (1 mmol) diiodo-d2-methane were
added and stirred for another 2 h at RT. After addition of 4 mL of
a 1 M TBAF solution in THF the reaction turned brownish and
was allowed to react for 18 h at RT. The solvent was removed
under vacuo and the residue purified by column chromatography
(DCM/MeOH 15:1 + 2% TEA). The product fractions were combined,
the solvent evaporated, redissolved in 10 mL dichloromethane and
washed with 10 mL of a 1 M NaHCO3 solution. The organic layer
was separated, dried over Na2SO4 and eluted over a SepPak Silica
Plus cartridge, to trap the product. Then the cartridge was eluted
with 20 mL ethyl acetate and the solvent evaporated in vacuo, to
obtain 30 mg (0.13 mmol, 13%) of fluoro-d2-methyl-harmol. 1H
NMR (300 MHz, CDCl3) d [ppm] = 8.172 (d, 1H), 7.884 (d, 1H),
7.736 (d, 1H), 7.070 (d, 1H), 6.862–6.826 (dd, 1H), 2.909 (s, 3H).
13C NMR (75 MHz, CDCl3) d [ppm] = 162.441, 146.267, 136.622,
134.151, 132.683, 127.616, 123.352, 113.887, 113.797, 113.242,
95.418, 84.204, 70.769, 70.634, 70.485, 70.230, 69.495, 67.950,
45.953, 16.329. HR-ESI-MS (ESI+): m/z: 377.1861 [M+1]+;
378.2017 [M+2]+.
Hence, the aim of this study was to synthesize [18F]fluoro-d2-
methyl-harmol
([18F]F-d2MH),
2-[18F]fluoroethyl-harmol
([18F]FEH) and the 18F-fluorinated PEG-derivatives developed by
Blom et al. and to evaluate their suitability to image the MAO A sta-
tus in vitro and in vivo.
2. Materials and methods
All reagents and solvents were of analytical-grade quality und
purchased from Sigma–Aldrich, Acros, Fluka or Merck. Unless
otherwise noted, all chemicals were used without further purifica-
tion. 2-Fluoroethyl-tosylate was prepared according to a published
procedure.33 1H and 13C spectra were recorded on a Bruker AC-300-
Spectrometer. 400 MHz 1H spectra were recorded on a DIX-400-
Spectrometer. 1H shifts were given in ppm with reference to the
internal tetramethylsilane (TMS) standard. Electro-spray mass
spectra (ESI) were recorded on a ThermoQuest Navigator Instru-
ment (Thermo Electron). High resolution mass spectra (HRMS)
were obtained on a Q-TOF-Ultima 3-Instrument (Waters). Column
2.1.3. Synthesis 2-fluoroethyl-harmol (4)
To a solution of 100 mg (0.5 mmol) harmol (2) in 10 mL dry eth-
anol, 97 mg (0.7 mmol) K2CO3 were added in one portion and stir-
red for 15 min at RT. Then 87.2 mg (0.5 mmol) of 2-fluorethyl
tosylate in 5 mL dry ethanol were added and the reaction heated
to 70 °C for 18 h before the ethanol was evaporated. The crude

614
H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
reaction mixture was purified by column chromatography (DCM/
MeOH 10:1 + 2% TEA) yielding 56 mg (0.23 mmol, 46%) of a slightly
yellow solid. 1H NMR (300 MHz, CDCl3) d [ppm] = 11.434 (s 1H),
8.152 (d, 1H), 8.077 (d, 1H), 7.813 (d, 1H), 7.026 (d, 1H), 6.88
(dd, 1H), 4.885 (t, 1H), 4.726 (t, 1H), 4.393 (t, 1H), 4.292 (t, 1H),
2.712 (s, 3H). ESI-MS (ESI+): m/z: 245.10 ([M+1]+).
67.83, 21.60, 19.08. (ESI+): m/z: 485.19 ([M+1]+). HR-ESI-MS
(ESI+): m/z: 485.1755 [M+1]+; 486.1790 [M+2]+; 487.1834 [M+3]+.
2.1.5.3.
7-(2-{2-[2-(2-Tosylethoxy)ethoxy]ethoxy}ethoxy)-1-
1H NMR (300 MHz, CDCl3) d
methyl-9H-b-carboline (7c).
[ppm] = 8.277 (d, 1H), 7.912 (d, 2H), 7.700 (q, 3H), 7.252 (d, 1H),
7.000 (d, 1H), 6.865–6.829 (dd, 1H), 4.173 (t, 2H), 4.139 (t, 2H),
3.865 (t, 2H), 3.675–3.568 (m, 10H), 2.735 (s, 3H), 2.366 (s, 1H).
13C NMR (75 MHz, CDCl3) d [ppm] = 159.82, 144.86, 142.26,
141.06, 137.55, 134.83, 132.71, 129.79, 127.83, 125.97, 122.35,
115.54, 112.04, 109.99, 95.80, 70.72, 70.67, 70.59, 70.46, 69.67,
69.29, 68.58, 67.72, 21.56, 19.97. ESI-MS (ESI+): m/z: 529.26
([M+1]+). HR-ESI-MS (ESI+): m/z: 529.2018 [M+1]+; 530.2124
[M+2]+; 531.2052 [M+3]+.
2.1.4. General procedure for the synthesis of the oligoethylated
compounds (PEG2, PEG3; 6a, 6b)
The syntheses of the ditosylated spacers were performed using
a modified protocol of Mohler et al.34 Hence, to a solution of
20 mmol of the respective oligoethylene glycol 5a or 5b in 10 mL
of dry dichloromethane 50 mmol of toluenesulfonyl chloride were
added in one portion and the mixture cooled to 0 °C. Then 44 mmol
of DABCO, dissolved in 8 mL dichloro methane, were added slowly
through a dropping funnel. The reaction mixture was allowed to
stir for additional 3 h at RT and finally filtered. The filtrate was
washed three times with water, the organic phase dried over
sodium sulfate and the solvent removed under vacuo to give a pale
yellow oil. To the crude reaction mixture 10 mL of dichlorometh-
ane were added to dissolve the oil. Addition of diethyl ether led
to precipitation of the ditosylated oligoethylene glycols 6a and
6b as colorless crystals (yield: 70–85%).
2.1.6. General procedure for the synthesis of the oligoethylated
reference compounds (8a, 8b, 8c)
The syntheses of the reference compounds were performed in
analogy to Blom et al.27 0.02 mmol of 7a, 7b or 7c were dissolved
in 6 mL dry THF before 0.02 mmol of a 1 M TBAF solution in THF
were added and heated at 55 °C. After 3 h the solvent was
removed to afford the crude product. Chromatography of the res-
idue (CHCl3/MeOH, 8:1 + 2% TEA) gave the pure product with a
yield of 40–50%.
2.1.4.1. Diethylenglycol ditosylate (6a).
1H NMR (300 MHz,
CDCl3) d [ppm] = 7.580 (d, 4H), 7.039 (d, 4H), 3.448 (t, 4H), 3.567
(t, 4H), 2.200 (s, 6H).
2.1.6.1. 7-[2-(2-Fluoroethoxy)ethoxy]-1-methyl-9H-b-carboline
(8a).
1H NMR (300 MHz, CDCl3) d [ppm] = 8.173 (d, 1H),
7.897 (d, 1H), 7.693 (d, 1H), 6.979 (d, 1H), 6.860 (dd, 1H), 4.204
(t, 1H), 4.169 (t, 1H), 4.120 (m, 2H), 3.949 (m, 2H), 3.935 (m, 2H),
2.812 (s, 3H).
2.1.4.2. Triethylenglycol ditosylate (6b).
CDCl3) d [ppm] = 7.77 (d, 4H), 7.33 (d, 4H), 4.11 (t, 4H), 3.64 (t,
4H), 3.50 (s, 4H), 2.42 (s, 6H).
1H NMR (300 MHz,
2.1.6.2. 7-{2-[2-(2-Fluoroethoxy)ethoxy]ethoxy}-1-methyl-9H-
b-carboline (8b).
(s, 1H), 8.276 (d, 1H), 7.911 (d, 1H), 7.682 (d, 1H), 6.943 (d, 1H),
6.863 (dd, 1H), 4.611 (t, 1H), 4.452 (t, 1H), 3.828 (m, 4H), 3.717
(m, 6H), 2.757 (s, 3H).
2.1.5. General procedure for the synthesis of the oligoethylated
harmol precursors (7a, 7b, 7c)
1H NMR (300 MHz, CDCl3) d [ppm] = 9.321
To a solution of 0.5 mmol harmol (2) in 15 mL dry acetone and
2 mL DMSO 0.75 mmol of sodium carbonate were added and stir-
red at RT. After 15 min 2 mmol of the respective ditosylate 6a, 6b
or 6c, dissolved in 10 mL dry acetone, were added. The reaction
mixture was refluxed overnight, quenched with water and three
times extracted with 20 mL of dichloromethane. The organic phase
was dried over sodium sulfate and evaporated to dryness. The
crude reaction mixture was purified by column chromatography
(EA + 2% TEA). The combined fractions were evaporated, dissolved
in dichloromethane and washed with 1 M sodium bicarbonate
solution. The organic layer was dried over sodium sulfate and
evaporated giving 55–75% yield for each particular compound.
2.1.6.3. 7-(2-{2-[2-(2-Fluoroethoxy)ethoxy]ethoxy}ethoxy)-1-
methyl-9H-b-carboline (8c).
1H NMR (300 MHz, CDCl3)
d [ppm] = 8.172 (d, 1H), 7.884 (d, 1H), 7.736 (d, 1H), 7.070 (d,
1H), 6.862–6.826 (dd, 1H), 4.579 (t, 1H), 4.406 (t, 1H), 4.155 (t,
2H), 4.124 (t, 2H), 3.755–3.551 (m, 10H), 2.909 (s, 3H). 13C NMR
(75 MHz, CDCl3) d [ppm] = 162.441, 146.267, 136.622, 134.151,
132.683, 127.616, 123.352, 113.887, 113.797, 113.242, 95.418,
84.204, 70.769, 70.634, 70.485, 70.230, 69.495, 67.950, 45.953,
16.329. HR-ESI-MS (ESI+): m/z: 377.1861 [M+1]+; 378.2017 [M+2]+.
2.1.5.1. 7-[2-(2-Tosylethoxy)ethoxy]-1-methyl-9H-b-carboline
1H NMR (300 MHz, CDCl3) d [ppm] = 8.23 (d, 1H), 7.87
2.2. Radiosyntheses
(7a).
(d, 1H), 7.73 (d, 2H), 7.65 (d, 1H), 7.22 (d, 2H), 6.99 (s, 1H), 6.79
(dd, 1H), 4.13 (t, 2H), 4.02 (t, 2H), 3.72–3.66 (m, 4H), 2.74 (s, 3H),
2.31 (s, 3H). 13C NMR (75 MHz, CDCl3) d [ppm] = 159.66, 145.02,
142.13, 141.37, 137.99, 134.96, 132.57, 129.86, 128.96, 128.34,
127.85, 125.98, 122.37, 115.72, 112.13, 110.11, 95.96, 69.92,
69.41, 68.73, 67.82, 21.55, 20.08. (ESI+): m/z: 441.18 ([M+1]+).
HR-ESI-MS (ESI+): m/z: 441.1470 [M+1]+; 442.1551 [M+2]+;
443.1558 [M+3]+.
2.2.1. Radiosynthesis of [18F]fluoro-d2-methyl-harmol (9)
[
18F]F-d2-MT was synthesized from n.c.a. [18F]fluoride (6–
8 GBq) in an automated synthesis module as described in the liter-
ature and finally trapped on Strata™ X SPE cartridge (Phenomenex)
and dried with helium.32 The [18F]F-d2-MT was eluted into a reac-
tion vial using 0.7 mL DMSO, tempered at 120 °C and added to a
solution of 3 mg (15 lmol) harmol and 5.6 lL of a 5 N NaOH solu-
tion in 0.3 mL DMSO. The mixture was heated for 20 min at 110 °C.
After cooling, the solution was diluted with 1 mL of the HPLC sol-
vent mixture and the [18F]F-d2-MT purified by semipreparative
HPLC. The product fraction was diluted with water (25 mL),
trapped on a Strata™ X SPE cartridge (Phenomenex), washed with
10 mL water, dried in a helium stream and eluted with 1 mL etha-
nol. The solvent was evaporated in vacuo and the residue dissolved
in 0.4 mL isotonic saline solution to obtain the [18F]fluoro-d2-
methyl-harmol (0.8–1.2 GBq) after a synthesis time of 115 min in
2.1.5.2. 7-{2-[2-(2-Tosylethoxy)ethoxy]ethoxy}-1-methyl-9H-b-
carboline (7b).
1H NMR (300 MHz, CDCl3) d [ppm] = 8.924
(s, 1H), 8.281 (d, 1H), 7.916 (d, 1H), 7.765 (d, 2H), 7.675 (d, 1H),
7.282 (d, 2H), 7.009 (d, 1H), 6.871 (dd, 1H), 4.166–4.120 (m, 4H),
3.823 (t, 2H), 3.690–3.519 (m, 6H), 2.740 (s, 3H), 2.378 (s, 3H).
13C NMR (75 MHz, CDCl3) d [ppm] = 160.32, 144.94, 142.92,
140.07, 134.45, 132.74, 129.84, 128.93, 127.89, 125.95, 122.55,
115.19, 112.24, 110.94, 98.09, 95.85, 70.78, 69.66, 69.38, 68.68,

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
615
an overall radiochemical yield of about 35 5% with a high radio-
chemical purity of >97%.
dance with the European Communities Council Directive of 24.
November 1986 (86/609/EEC) and the German law for animal wel-
fare and were approved by the local ethical committee for animal
experiments.
2.2.2. Radiosynthesis of 2-[18F]fluoroethyl-harmol (10)
2-[18F]fluorethyl-tosylate ([18F]FETos) was synthesized from
n.c.a. [18F]fluoride (3–5 GBq) using a homemade automated syn-
thesis module as described elsewhere and finally trapped on a
Strata™ X SPE cartridge (Phenomenex) and dried with helium.35
The [18F]FETos was eluted into a reaction vial with 0.8 mL DMSO,
2.5. In vivo imaging studies in Sprague Dawley rats
MicroPET imaging was performed with a Siemens/Concorde
Microsystems microPET Focus 120 small-animal PET camera (Sie-
mens/Concorde, Knoxville, TN, USA). Sprague Dawley rats were
anaesthetized with 2.5% isoflurane in medical oxygen, placed in
supine position in the small-animal PET scanner and warmed with
a heating lamp. After cannulation of the tail vein, a 10 min trans-
mission scan was performed using a 57Co point source. Following
an intravenous bolus injection of [18F]F-d2-MH (26.2 4.3 MBq)
or [18F]FEH (28.5 3.7 MBq) to wild type (n = 5), clorgyline treated
rats (n = 3, 10 mg/kg clorgyline iv 10 min before injection of the
tempered at 130 °C and added to a solution of 3 mg (15
lmol) har-
mol and 5.6 L of a 5 N NaOH solution in 0.2 mL DMSO. The mix-
l
ture was heated for 10 min at 130 °C. After cooling, the solution
was diluted with 1 mL water and the 2-[18F]fluoroethyl-harmol
purified by semipreparative HPLC. The product fraction was
diluted with water (25 mL), trapped on a Strata™ X SPE cartridge
(Phenomenex), washed with 10 mL water, dried in a helium stream
and eluted with 2 mL ethanol. The solvent was evaporated in vacuo
and the residue dissolved in 1.0 mL isotonic saline solution to
obtain the tracer (1.2 GBq) after a synthesis time of 100 min in
an overall radiochemical yield of about 47 2% with a high radio-
chemical purity of >97%.
tracer) and L-deprenyl treated rats (n = 3, 1 mg/kg L-deprenyl iv
10 min before injection of the tracer), a dynamic emission scan
was accomplished for 1 h. Furthermore, for [18F]FEH challenge
experiments in rats (n = 3) were performed using moclobemide, a
reversible MAO A inhibitor with a Ki-value of 57.1 5.2 l
M.36
2.2.3. Radiosyntheses of the PEGylated compounds (PEG2–
PEG4)(11a–c)
Hence, following a published procedure 20 mg/kg moclobemide
were injected in a bolus 20 min p.i. followed by a second injection
of 10 mg/kg at 45 min p.i.37 The 18F-labeled PEGylated harmol
derivatives were examined dynamically for 60 min in untreated
animals (n = 2) after intravenous bolus injection (31.6 2.4 MBq)
of the tracer.
The PEGylated harmol derivatives were synthesized starting
from n.c.a. [18F]fluoride (3–5 GBq) by a 18F-direct fluorination in
a micro wave oven. The aqueous [18F]fluoride solution was trapped
on a Sep-Pak Accell Plus QMA light cartridge and eluted with a
solution of 15
l
L 1 M K2CO3 and 15 mg KryptofixÒ222 in 0.8 mL
The obtained list-mode files were rebinned to sinograms with
19 frames (3 ꢀ 20 s, 3 ꢀ 60 s, 3 ꢀ 120 s, and 10 ꢀ 300 s) and
reconstructed by four iterations of 2-dimensional ordered-subset
expectation maximization (OSEM2D) with scatter and attenuation
correction. This finally results in a 128 ꢀ 128 ꢀ 95 matrix, with
0.87 ꢀ 0.87 ꢀ 0.80 mm3 voxels. Reconstructed images were ana-
lyzed with the modeling computer software PMOD (PBAS & PFUS,
PMOD Technologies Ltd, Zurich, Switzerland). Using the summed
PET images (0–60 p.i.) the brain was co-registered to a MRI tem-
plate (93 ꢀ 93 ꢀ 120 matrix) via rigid translations and rotations
for each axis, till the PET image matches the MRI template. The
obtained transformation matrix then was applied to each frame
of the dynamic PET images, resulting in a PET study with a
93 ꢀ 93 ꢀ 120 matrix. Then a volume of interest (VOI) for the brain
was drawn on the MRI template and applied to the PET images, to
determine the time activity curve (TAC) of the brain uptake for
each animal. Afterwards, the TAC of this VOI was decay corrected
to the injection time and normalized concerning the injected dose
and the weight of the animal to obtain the SUV [SUV = (radioactiv-
ity in Bq per mL/injected radioactivity in Bq) ꢀ body weight in g]
for each time point.
MeCN. The water was removed using a stream of N2 at 80 °C and
coevaporated to dryness with MeCN (3 ꢀ 1 mL). Afterwards, the
dried K[18F]F-K222 complex was dissolved in 400
lL acetonitrile
and added to a solution containing 2.8
lmol of the respective pre-
cursor (7a,7b or 7c) in 200 L acetonitrile. The reaction vial was
l
sealed and placed in a microwave, heated for 5 min at 125 °C and
quenched with water. The crude product was purified via semipre-
parative HPLC using the conditions described above. The isolated
tracer was diluted with water (25 mL), trapped on a Strata™ X
SPE cartridge (Phenomenex), washed with 10 mL water and eluted
with 2 mL of ethanol. The solvent was evaporated in vacuo and the
residue dissolved in 1.0 mL isotonic saline solution to obtain the
tracer (1 GBq) after a synthesis time of 85 min in an overall radio-
chemical yield of about 20 10% with a high radiochemical purity
of >97%.
2.3. Plasma stability
3–5 MBq of [18F]F-d2-MH (9), [18F]FEH (10), [18F]PEG2-harmol
(11a), [18F]PEG3-harmol (11b) and [18F]PEG4-harmol (11c) in
200 lL saline solution were incubated in triplicates in 400 lL of
human serum at 37 °C. At 5, 20, 30 and 60 min the incubation
2.6. In vivo metabolism studies
was stopped by putting the Eppendorf vial on ice. Subsequently,
400
and the vial was centrifuged at 10,000 rpm for 8 min. A sample
of 20 L of the supernatant was taken and analyzed in the radioan-
l
L of MeCN were added to precipitate the plasma proteins
The metabolism of [18F]F-d2-MH (9) and [18F]FEH (10) was
determined in analogy to Herth et al.38 Briefly, blood samples were
taken 5, 10, 20, 30, 60 min p.i. Erythrocytes and plasma proteins
were precipitated and plasma water was measured using a Perkin-
Elmer Wizard2 auto-gamma counter.
l
alytical HPLC system described above.
2.4. Animals
2.7. Monoamine oxidase assays
For all experiments Sprague Dawley rats (285 50 g) were used,
which were obtained from the animal husbandry of the Johannes
Gutenberg University of Mainz, Germany. Animals had access to
food and water ad libitum. Temperature and humidity were kept
at 24 2 °C and 60%, respectively. All animals were maintained
on a 12-h light/12-h dark cycle. Handling occurred only during
the light cycle. All animal experiments were performed in accor-
Rat brain cortical membranes were obtained as previously
described.39 After thawing the membranes were centrifuged at
23,000g for 20 min and resuspended in 50 mM Tris/HCl, pH 7.5
(Tris/HCl).
The MAO assay was performed according to Krajl et al. with
some modifications.40 In brief: cortex membranes were

616
H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
pre-incubated in Tris/HCl with 37.5 nM clorgyline and 37.5 nM
-deprenyl for 5 min at 37 °C to selectively and irreversibly inhibit
MAO A and MAO B, respectively (Fig. 8). The range of compounds
1–4 and dopamine is given in the text. Total MAO activity was
obtained in the absence of any additive and total nonspecific val-
replacing the methyl group of harmine by an oligoethylene glycol
spacer (PEG2, PEG3) resulted in a decrease of unspecific tracer-
binding, Consequently, we decided to synthesize and evaluate an
additional PEG4 derivative of harmol to investigate the influence of
all PEG spacers on the brain uptake and specific binding.
L
ues were determined in the presence of 8.33
l
M pargyline. The
All reference compounds and the precursors of this lead struc-
ture were synthesized and reaction conditions for their labeling
screened and optimized. Furthermore, the tracers were evaluated
volume during incubation amounted to 200
performed at least three times in duplicates.
lL. All assays were
Final protein concentrations were kept in the range from 50
to 115 g in each tube in order to obtain little variation in the
MAO-subtype irreversible inhibition by clorgyline and -deprenyl,
respectively.41 The reaction was started by the addition of 25
l
g
in plasma stability tests and lPET imaging studies, as well as their
affinity tested in binding assays.
l
L
lM
3.1. Synthesis of labeling precursors and reference compounds
kynuramine hydro bromide, slightly below the KM value in order
to avoid substrate inhibition.42 After 30 min at 37 °C the reaction
was terminated by 267 mM HClO4. Samples were centrifuged at
13,000g for 2 min and an aliquot of the supernatant was added
to a four-fold excess in volume of 1 M NaOH. Data were normalized
The precursors and reference compounds were synthesized
according to the routes shown in Scheme 1. The organic reactions
were screened and optimized with regard to solvents, base, reac-
tion temperature and side reactions, considering that the reactions
follow a SN2 mechanism.
to the value in the presence of clorgyline (MAO-B) and L-deprenyl
(MAO A), set to 100%. The sum of the thus determined values for
MAO A activity and MAO B activity were around 100% of the total
MAO activity for all assays. Fluorescence of the developed 4-OH-
quinoline was measured in a Fluoromax 2 fluorescence spectro-
photometer (Horiba Scientific, Unterhaching, Germany) at
313 nm absorption and 410 nm emission with slit widths of
2 nm. The IC50 values were fitted to a sigmoidal dose–response
curve using GraphPad (version 5).
Harmol, the key intermediate of every synthesis, was prepared
via demethylation of harmine (1) using HBr/glacial acid according
to a published procedure26, giving the harmol in high yields of
about 80%. The synthesis of 2-fluoroethyl-harmol was performed
in a similar manner by reacting harmol (2) with 2-fluoroethyl-tos-
ylate in the presence of potassium carbonate as base and potas-
sium iodide as catalyst, resulting in moderate yields of 46% for
this reference compound (Scheme 2).
Possible synthetic routes for the preparation of the first refer-
ence compound, fluoro-d2-methyl-harmol, were limited by the
commercial availability of suitable deuterated methylating agents.
Hence, a two-step procedure was examined, which consists of the
alkylation of harmol with diiodo-d2-methane, followed by a fluori-
nation of the obtained iodo-d2-methyl-harmol with tetra-n-butyl-
ammonium fluoride (TBAF). First attempts to purify the crude
product of the alkylation step via column chromatography were
unsuccessful, due to the decomposition of the iodo-d2-methyl-har-
mol on the column. Thus, the reaction route was performed as a
two-step one-pot synthesis, giving the fluoro-d2-methyl-harmol
in a high chemical purity and a yield of 13%.
3. Results and discussion
The main purpose of this study was to synthesize and evaluate
18F-fluorinated harmine derivatives for the in vivo visualization of
the monoamine oxidase A with PET. A common approach for the syn-
thesis of 18F-fluorinated derivatives is the18F-fluoroalkylationof11C-
labeling precursors, for example, giving 2-[18F]fluoroethyl-harmol.
Moreover recent developments in 18F-fluoro-d2-methylation tech-
niques, enabled us to synthesize [18F]fluoro-d2-methyl-harmol,
which seems to be a promising candidate because of the similar ste-
ric demands of a fluoromethyl and a methyl moiety. New PEGylated
harmine derivatives were also synthesized, because they seem to be
interesting, due to the promising results of autoradiographic studies
performed by Blom et al.27 Particulary, it has been shown, that
Although the syntheses of the reference compounds and label-
ing precursors for the PEGylated compounds have already been
described in the literature, the use of toluenesulfonyl moieties
Scheme 1. Synthesis routes of precursors and reference compounds. Reagents and conditions: (i) HBr/HAc; ii) KOtBu, CD2I2, TBAF; (iii) K2CO3, EtOH, FETos, 70 °C; (iv) DABCO,
DCM, 0 °C; (v) Cs2CO3, DMSO, acetone, reflux; (vi) THF, TBAF, 55 °C.

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
617
Table 2
Optimization of solvents for the coupling reaction of harmol and the PEG-spacers
Solvent
DMF
15
DMSO/acetone
71
DCM
22
CHCl3
0
DMSO
65
Yield (%)
The focus of the whole optimization process was set on the
reactions using PEG4 (6c), because this synthesis gave the best
yields during the base screening process. Unfortunately, this
screening showed that a mixture of DMSO and acetone was supe-
rior to solvents like DMF, pure DMSO and dichloromethane. Thus,
it was only possible to reduce the amount of high boiling solvents,
to simplify the purification of the products.
In summary, the optimization showed that potassium carbon-
ate and a mixture of acetone/DMSO as base and solvent were the
system of choice, and hence were used for the general synthesis
procedure for the preparation of precursors 7a, 7b and 7c. Further-
more, compared to a previous study, the yields of the precursors
were increased significantly, by using the ditosylated PEG-spacers
and the reaction conditions described above. The fluorination steps
for the synthesis of the reference compounds were performed fol-
lowing a general procedure using a 1 M TBAF solution in THF,
whereas no optimization was carried out.27 These reactions pro-
ceeded without any side reactions, requiring no special workup
procedure, and gave the reference compounds in good yields.
Scheme 2. Radiosynthesis of harmol derivatives. Reagents and conditions: (i)
18F]F-d2-MTos, 2 equiv 5 N NaOH; (ii) [18F]FETos, 2 equiv 5 N NaOH; (iii) K18F,
K2.2.2/MeCN, microwave.
[
Table 1
Optimization of base systems for the coupling reaction of harmol and the PEG-spacers
Base
6a yield (%)
6b yield (%)
6c yield (%)
TMA
DEA
TEA
K2CO3
Cs2CO3
NaH
<1
7
18
36
15
17
nd
5
15
42
20
20
nd
10
20
65
28
10
3.2. Radiochemistry
for the oligoethylene spacer carrying tracers as leaving groups
instead of halides was preferred for this study, to increase the
radiochemical yields.27 A further advantage of using toluenesulfo-
nyl groups is the prolonged shelf-life of the precursors, which facil-
itates the storage of the precursors at 4 °C instead of –78 °C. In
contrast to the tetraethylene glycol ditosylate (6c), which is com-
mercially available, the ditosylated oligoethylene spacers diethyl-
ene glycol ditosylate (6a) and triethylene glycol ditosylate (6b)
had to be synthesized using a modified protocol of Mohler et al.34
Hence, the oligoethylene glycol spacers were ditosylated by using
toluenesulfonyl chloride and DABCO instead of triethylamine, fol-
lowed by precipitation of the products due to the addition of a
huge excess of cold diethyl ether. For the following coupling
reaction the different reaction parameters were screened and opti-
mized (Table 1) because of the low yields achieved in previous
studies.
The radiolabeling of the different precursors was performed as
shown in Scheme 2.
The synthesis of [18F]fluoro-d2-methyl-harmol was performed
in a two-step procedure. In the first step, the labeling agent
[
18F]fluoro-d2-methyl tosylate was produced, which was then used
to 18F-fluoro methylate harmol. Therefore, the [18F]fluoro-d2-
methyl-harmol was produced in an automated synthesis module.
According to a recently published study the 18F-labeled prosthetic
group could be obtained after a synthesis time of 50 min in a radio-
chemical yield of about 45%.32 For the next step the [18F]F-d2-MTos,
which was finally trapped on a Strata-X 33 lM SPE cartridge and
eluted with DMSO, was tempered to the reaction temperature.
Then it was reacted with a solution of harmol and 5 N NaOH in
DMSO to obtain the final [18F]F-d2-MH. For the labeling reaction
different reaction parameters, like precursor concentration,
amount of base, reaction time and temperature were optimized,
The use of non-nucleophilic bases like trialkylamines or carbon-
ates for this step was necessary to avoid side reactions. Whilst di-
or trialkylamines (DEA, TEA) seemed to be less suitable, indicated
by the low yields of up to 20%, the use of sodium hydride led
almost exclusively to decomposition of the harmol (2). In contrast,
the inorganic bases, like carbonates, showed the most promising
results, especially potassium carbonate, which gave the highest
yields in all three coupling reactions and a minimum of side reac-
tions. However, the use of potassium or cesium carbonate led to
by-products in all cases showing similar Rf-values compared to
the desired products. Since these by-products were predominantly
observed when strong bases were used for this reaction, the by-
products of a reaction of 6c and harmol in the presence of cesium
carbonate were examined. The characterization of the side product
via NMR- and mass-spectroscopy showed, that an N-alkylation of
harmol took place. To verify that hypothesis, that N-alkylation is
possible, a N-Boc-protected harmol derivative was synthesized
and reacted with PEG4-ditosylate under the same reaction condi-
tions, indicating no formation of the by-product. Beside the base
screening, different solvent systems were tested to facilitate the
whole work-up procedure and thus avoiding high boiling solvents
(Table 2).
resulting in the following reaction conditions: 15
5.6
lmol harmol,
l
L 5 N NaOH, 110 °C and 20 min reaction time. This gave radio-
chemical yields of about 80% for the 18F-fluoromethylation step.
The purification of the reaction mixture was performed by HPLC,
trapping the product on SPE cartridge, elution with ethanol, evap-
oration of the ethanol and subsequent formulation in isotonic
sodium chloride solution, yielding the [18F]F-d2-MH in an overall
radiochemical yield of about 35% after
a
synthesis time of
115 min and specific activities of 40–90 GBq/
lmol.
In a first approach to synthesize [18F]fluoroethyl harmol via a
18F-direct fluorination, a precursor from a previous study,27 2-(1-
methyl-9H-pyrido[3,4-b]indol-7-yloxy)ethyl
4-methylbenzene
sulfonate, was prepared. Attempts to label this precursor resulted
in decomposition of the precursor and no labeling was obtained.
An explanation of this behavior is may be the high reactivity of
the toluenesulfonyl leaving group used. As a consequence, the
radiosynthesis of [18F]fluoroethyl harmol was performed in a
two-step procedure, comparable to [18F]F-d2-MH. The labeling
agent used in this route, 2-[18F]FETos, was prepared on a self-made
automated synthesis module, yielding a radiochemical yield of
about 60% after 55 min. The [18F]FETos, trapped on a SPE cartridge

618
H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
during the final step of the automated synthesis, was eluted with
DMSO and tempered to the reaction temperature. Then it was
reacted with a solution of harmol and 5 N NaOH in DMSO, to obtain
the [18F]FEH. The reaction parameters of this 18F-fluorethylation
heating. Longer reaction times led to decomposition of all precur-
sors. After the optimization, the purification of the tracers was
performed in an analogue manner to
[
18F]F-d2-MH, yielding
[
18F]11a–c in an overall radiochemical yield of 20 10% after a syn-
were optimized, resulting in the following conditions: 15
lmol
thesis time of 85 min.
precursor, 5.6 L NaOH, 110 °C and 20 min reaction time. Using
l
these conditions radiochemical yield of up to 95% could be
achieved. Purification of the reaction mixture was performed in
an analogue manner to [18F]F-d2-MH, yielding the [18F]FEH in an
overall radiochemical yield of 47 2% after a synthesis time of
3.3. Plasma stability tests
Plasma stability is an important issue for PET-tracers, since a
fast metabolism in the plasma reduces drastically the tracer-
uptake target regions. For brain tracers this circumstance is even
more crucial if non-polar metabolites are formed, increasing
unspecific binding and complicating the kinetic modeling of the
tracer. Thus, all tracers were incubated in human serum, the
plasma proteins were precipitated by addition of MeCN and sepa-
rated by centrifugation. The obtained supernatant was analyzed by
HPLC, to determine the stability of the respective tracer in vitro,
showing that all tracers had a sufficient stability over an observa-
tion period of 60 min. The results of this study are shown in
Table 4.
105 min and specific activities of 40–90 GBq/lmol. Although, this
two-step procedure results in a longer synthesis time, it finally
produces slightly higher yields of [18F]FEH and seems to be robust,
compared to the method published by Blom et al.27
In contrast to [18F]F-d2-MH and [18F]FEH, the labeling of the
PEGylated harmol derivatives was achieved via direct 18F-fluorina-
tion of the tosylated precursors. The tosylated precursors gave
higher yields than the halogenated precursors used in the study
of Blom et al.27 To obtain the basic conditions, which are needed
for 18F-direct fluorination,
a
potassium carbonate/KryptofixÒ
2.2.2. system was used. However, first attempts to label the
precursor using different solvents and temperatures resulted in
low radiochemical yields, which may be explained by the acidic
indole hydrogen hampering the labeling reaction. Hence, in the
next experiments the ratio of precursor to potassium carbonate/
KryptofixÒ 2.2.2. was raised from 1/1 to 1/4.3, resulting in higher
radiochemical yields of up to 20% in DMF. These high base concen-
trations did not only increase the conversion, but also reduced the
reliability of the labeling, resulting in more by-products and a more
complicated purification of the tracers. Therefore, as an alternative
route to the conventional heating, a microwave-assisted synthesis
was tested. During the first experiments, using potassium carbon-
ate/KryptofixÒ 2.2.2., it was observed, that higher radiochemical
yields were obtained compared to the conventional heating and
side products could be minimized. Hence, this method was
screened regarding to temperature, which had the biggest impact
on conversion rates, cf. Table 3. Reaction times and amount of pre-
cursor seemed not to be major rate limiting parameters (data not
shown).
3.4. In vivo imaging studies in Sprague Dawley rats
To determine brain uptake kinetics of the 18F-fluorinated
harmol derivatives, lPET studies in Sprague Dawley rats were per-
formed. Furthermore, for [18F]F-d2-MH and [18F]FEH the selectivity
of the tracers was investigated by a selective blocking of MAO B
and MAO A. In a first imaging study the brain uptake of [18F]F-
d2-MH in untreated, in
animals was measured (Fig. 3).
18F]F-d2-MH showed a high initial uptake and rapid clearance
L-deprenyl treated and in clorgyline treated
[
out of the brain. The initial SUV of 2.1 0.4 g/mL after 10 s
decreased to 1.2 0.2 g/mL within 30 min, which can be assigned
to blood flow. The highest uptake levels of 3.7 0.5 g/mL were
reached 90 s p.i., followed by a fast washout to 1.5 0.2 g/mL after
20 min and 1.0 0.2 g/mL after 45 min. Blocking of MAO B was
performed using L-deprenyl, a selective MAO B inhibitor, following
a procedure described in the literature.26 The uptake kinetics of
treated animals showed a behavior very similar to the untreated
animals, hence a high initial uptake and fast clearance out of the
brain. The highest uptake values of 3.8 0.8 g/mL were reached
at 50 s p.i., followed by washout to 1.4 0.5 g/mL after 20 min
and 0.9 0.2 g/mL after 45 min. At later time points a slightly
lower uptake, compared to the untreated animals, was observed,
but under consideration of the errors of both groups there is no sig-
nificant difference. Finally, blocking of the MAO A was done using
clorgyline a selective MAO A inhibitor.20 These animals showed
very similar initial brain uptake of 3.8 0.6 g/mL, but in contrast
to the other two groups a faster washout was observed, giving
SUVs of 1.0 0.2 g/mL after 20 min. The highest differences in
brain uptake between the untreated and the clorgyline-treated
animals, representing the specific binding, were observed from 3
to 20 min p.i., giving SUV differences of >0.5 g/mL. In summary,
both blocking studies indicate, that [18F]F-d2-MH binds selective
to MAO A albeit it possesses very fast pharmacokinetics in vivo.
The optimization of this direct 18F-fluorination showed, that
DMF or DMSO resulted in no or low radiochemical yields, whilst
MeCN gave radiochemical yields of up to 50% and thus was chosen
for the next steps. Subsequently, the reaction temperature was
screened giving the highest RCY between 100–135 °C. When the
temperature was raised beyond that range decomposition of the
precursor increased drastically, while lowering it beyond that
range resulted in very low radiochemical yields. Additionally, the
precursor concentration was screened showing the highest RCY
with a precursor concentration of at least 12.5 lmol/mL. Finally,
the influence of the reaction time on the RCY was investigated
showing the highest RCY in a time range of 1–6 min of microwave
Table 3
Optimization of 18F-labeling of 2 and 7a–c
Temp (°C)
Tracer
11b
11a
11c
Table 4
115a
120b
125a
135a
140b
3%
5%
32 6%
28%
nd
10%
47%
39 11%
15%
nd
2%
Stability of the 18F-fluorinated harmol derivatives in human plasma as a function of
the incubation time at 37 °C
10%
46 2%
38%
nd
Incubation time
5 min
15 min
30 min
60 min
[
[
[
[
[
18F]F-d2-MH
95.5 1.9
98.6 2.9
96.0 1.8
95.4 0.2
97.8 0.6
95.8 0.1
98.4 2.8
97.2 1.4
97.0 0.4
97.1 1.1
95.8 0.1
97.9 2.8
96.7 1.6
95.4 2.3
98.0 0.9
94.4 0.1
97.5 2.5
94.7 1.9
95.7 1.4
97.0 0.7
18F]FEH
For optimization of the reaction temperature the optimized precursor and base
concentrations were used.
18F]PEG2-harmol
18F]PEG3-harmol
18F]PEG4-harmol
a
Microwave assisted heating.
Conventional heating.
b

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
619
Figure 3. Time activity curves from microPET-studies of [18F]F-d2-MH. Curves were drawn from activity measurements in the brain for untreated (n = 5), clorgyline-treated
(n = 3) and deprenyl-treated (n = 3) Sprague Dawley rats. Each curve gives the mean SUV SEM.
Representative images for the early distribution pattern of
[
18F]F-d2-MH in the rat brain are shown in Figure 4. Besides the
high uptake in the brain, a high uptake in the harderian glands
(upper left and right corner of each image) was also observed. Fur-
thermore, a high accumulation in the skull was noticed in all three
groups of animals at later time points, although the plasma stabil-
ity test of the tracer showed a high stability in human serum. This
indicates tracer metabolism under release of [18F]fluoride, a highly
sensitive bone-seeking radionuclide, which accumulates in skull.43
For [18F]FEH the
lPET studies were done using a similar setup
to [18F]F-d2-MH, resulting in SUV curves shown in Figure 5. For
the untreated animals a high initial uptake, comparable to [18F]F-
d2-MH, was observed followed by a slower washout from the brain.
Highest SUV of [18F]FEH of 3.7 0.4 g/mL was reached at 7 min p.i.,
followed by a slow clearance out of the brain resulting in
3.6 0.4 g/mL after 20 min and 3.4 0.4 g/mL after 45 min,
respectively.
Figure 4. Representative horizontal brain slices of the brain uptake of [18F]F-d2-MH
in Sprague Dawley rats. This series of OSEM3d reconstructed images (dorsal to
ventral, slice interval 1.0 mm) shows a sum image from 0–12.5 min p.i. (color code:
80–170 kBq/g).
Additionally, the selectivity of 10 was investigated in blocking
studies using L-deprenyl and clorgyline. L-Deprenyl-blocked rats
Figure 5. Time activity curves from microPET-studies of [18F]FEH. Curves were drawn from activity measurements in the whole brain in untreated and clorgyline- and
deprenyl-treated Sprague Dawley rats. Each curve gives the mean SUV SEM.

620
H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
in the blood and consequently an increased radioactivity signal in
the brain. Following this slight increase, a significant washout,
compared to the untreated rats was observed, demonstrating the
reversibility of the [18F]FEH binding. In summary, these experi-
ments demonstrate promising properties of [18F]FEH for the visu-
alization of the MAO A status in vivo, showing high affinity, high
selectivity and the reversibility of the tracer to MAO A.
Although the MAO A is ubiquitous expressed in the brain, for
some regions a higher uptake of [18F]FEH was observed. These
regions include cortical regions, the thalamus, the colliculus infe-
rior and the midbrain. Interestingly, for the cerebellum, a brain
region with a low concentration of MAO A, a moderate uptake
was determined. An explanation for this behavior could be that
the locus coeruleus, a region with one of the highest MAO A
expressions in brain, is directly located next to the grey matter of
the cerebellum, which results, via spill-over effects, in an increased
apparent uptake in the cerebellum. Representative images for the
distribution pattern of [18F]FEH in the rat brain, are shown in Fig-
ure 6. The relatively high MAO A concentrations in most brain
regions aggravate an exact determination of the SUVs, due to par-
tial volume and spill-over effects from neighboring regions. Thus, a
detailed determination of the SUVs for the different brain regions
was not performed.
Figure 6. Representative horizontal brain slices of the brain uptake of [18F]FEH in
Sprague Dawley rats. This series of OSEM3d reconstructed images (dorsal to ventral,
slice interval 1.0 mm) shows a sum image from 30–60 min p.i. (color code: 200–
650 kBq/g).
showed nearly identical SUV-normalized uptake curves as the
untreated animals. Thus, a high initial uptake was followed by a
slow washout from the brain. Highest uptake of 3.6 0.5 g/mL
was achieved at 7 min p.i., followed by a slow clearance giving
SUVs of 3.6 0.5 g/mL after 20 min and 3.4 0.5 g/mL after
45 min. These results indicate that [18F]FEH does not bind to
MAO B. In contrast, blocking with clorgyline showed a significantly
different uptake curve compared to the untreated rats. Following a
moderate initial uptake of the tracer in the brain, a rapid clearance
was observed, showing a maximum SUV of 2.6 0.5 g/mL after 50 s
and SUVs of 1.1 0.2 and 1.0 0.03 g/mL after 20 and 45 min.
Hence, the binding of [18F]FEH can be drastically reduced by an
MAO A inhibitor. This clearly shows the high selectivity of the tra-
cer for MAO A.
Finally, challenge experiments using moclobemide were per-
formed, to determine the reversibility of the [18F]FEH binding to
MAO A. Following a published procedure, moclobemide was
applied in two bolus injections at 20 min and 45 min p.i. of the tra-
cer. As shown in Figure 5, the first application of moclobemide
directly results in a slight increase of the radioactivity in the brain.
This effect can be explained by the clearance of [18F]FEH from the
peripheral organs with high MAO A concentrations, like heart, liver
or pancreas. This clearance results in a higher tracer concentration
The evaluation of the brain uptake of the 18F-PEGylated harmol
derivatives, [18F]PEG2-harmol, [18F]PEG3-harmol and [18F]PEG4-
harmol, lPET studies in untreated Sprague Dawley rats were per-
formed. Surprisingly, none of these tracers showed a significant
brain uptake. In the previous work from Blom et al., however, it
was demonstrated in autoradiographic studies, that two of these
derivatives,
[ [
18F]PEG2-harmol and 18F]PEG3-harmol, bind to
MAO A and also show a high selectivity for this target. Hence,
the most likely explanation for this behavior is that the PEGylation
results in a reduction of the lipophilicity of these compounds,
which hinders their passage of the blood–brain barrier. To investi-
gate the differences in lipophilicity we determined the capacity
factors (k0) for all the derivatives ending up with the following
order: [18F]FEH: k0 = 4.03 > [18F]F-d2-MH: k0 = 4.01 > [18F]PEG2-har-
mol: k0 = 3.97 > [18F]PEG3-harmol: k0 = 3.77 > [18F]PEG4-harmol:
k0 = 3.74. Thus, [18F]FEH is the most lipophilic compound of the
derivatives, whereas the lipophilicity decreases due to PEGylation.
This order is in accordance with the experimental observation that
the brain uptake of the PEGylated compounds, although very low,
decreases further from [18F]PEG2-harmol to [18F]PEG4-harmol.
Besides the lower lipophilicity of the PEGylated compounds, the
Figure 7. Metabolism of [18F]F-d2-MH and [18F]FEH in Sprague Dawley rats as a function of time post injection

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
621
100
80
60
40
20
0
clorgyline
deprenyl
100
80
60
+ 37.5 nM clorgyline
40 + 37.5 nM deprenyl
20
0
0
1
1
0.
1
0
0
0
0
1
00
0
.
001
1
0
1
0.
µM Dopamin
0.001
0
0.0001
0.01
0.1
1
10
100
µM
Figure 8. Dose–response curves of clorgyline and
Clorgyline and -deprenyl were co-incubated at concentrations ranging from 0.1 nM up to 100
3.0 M (lg IC50: 0.47 0.31 (lg
M)) for clorgyline (squares) at the high and low affinity sites, respectively. The IC50 values obtained were 3.2 nM (lg IC50: ꢁ2.50 0.23 (lg
and 1.0 M)) for deprenyl (triangles) at the high and low affinity sites, respectively. n = 4. The fractions of the high affinity sites amounted to
M (lg IC50: ꢁ0.019 0.375 (lg
46 5% and 52 4% for clorgyline and
L
-deprenyl against MAO activity in vitro and determination of the IC50 values of dopamine against MAO A and MAO B.
M. The IC50 values obtained were 3.1 nM (lg IC50: ꢁ2.5 0.3 (lg M)) and
M))
L
l
l
l
l
l
l
l
L
-deprenyl, respectively.
FMH
FflEuoHroethyl-harmol
Hhaarrmmiinne
fluoro-d2-methyl-harmol
100
80
60
40
20
0
Hhaarrmmooll
0
0.1
10
1000
nM
Figure 9. MAO A activity of harmol, harmine, F-d2-MH and FEH. The ranges of concentrations were between 0.1 nM and 100 nM for harmine and FEH, between 1 nM and
M for F-d2-MH, and between 10 nM and 10 M for harmol. The corresponding IC50 values are given in the text.
1
l
l
capability of the PEG-chains to form H-bonds may be responsible
for the low brain uptake. Since each pair of hydrogen bonds formed
with the solvent results in a 10-fold decrease in membrane perme-
ation, this limits the ability of the PEGylated compounds to pass
penetrate the BBB with increasing length of the PEG-chain.44,45
Due to these results no further in vivo experiments were per-
formed with these compounds.
ride. Hence, the metabolic profile of [18F]F-d2-MH was examined
in four animals, resulting in the data depicted in Figure 7. In
contrast to the in vitro experiments, the tracer undergoes a fast
in vivo metabolism showing only 50% of intact tracer after
5 min, which gets further metabolized, resulting in only 30% of
intact tracer after 20 min p.i.. HPLC analysis of the plasma water
showed exclusively polar metabolites having retention times
comparable to the dead time of the HPLC. This underlines the
hypothesis that this tracer is metabolized under release of
3.5. In vivo metabolism studies
[
18F]fluoride.
In the in vivo imaging studies of [18F]F-d2-MH a high accumu-
lation of activity in the skull was observed at later time points,
indicating a metabolism of the tracer under release of [18F]fluo-
In a further evaluation the metabolic profile of the [18F]FEH was
examined in an analogous manner to [18F]F-d2-MH, cf Figure 7. In
comparison to the 18F-fluoromethylated compound, a fast in vivo

622
H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
100
80
60
40
20
0
FflMuoHro-d2-methyl-harmol
FflEuoHroethyl-harmol
hhaarrmmiinnee
harmol
harmol
0
1
10
100
1000
nM
Figure 10. MAO B activity of harmol, harmine, F-d2-MH and FEH. The ranges of concentrations were between 1 and 100
1 mM for harmol and F-d2-MH.
lM for harmine and FEH and between 1 lM and
metabolism was observed as well, showing only 15% of intact tra-
cer after 5 min, which gets further metabolized, resulting in only
5% of intact tracer after 60 min p.i.. HPLC analysis of the plasma
water showed two different polar metabolites having retention
times slightly higher than the dead time of the HPLC, but no de-
fluorination could be observed.
presence of clorgyline and L-deprenyl (Figs. 3 and 5) may be
reflected in the in vitro selectivities (Figs. 9 and 10). The IC50 values
of FEH were the highest we observed (0.54 0.06 nM). This high
affinity may account for the slow washout compared to that of F-
d2-MH.
4. Conclusion
3.6. Monoamine oxidase assays
In the present work five 18F-labeled harmine derivatives were
evaluated in lPET studies to determine their suitability to visualize
Based on the results of the in vivo imaging studies, potential
ligands, namely [18F]F-d2-MH and [18F]FEH, were evaluated in
enzyme assays to determine their affinity towards MAO-A and
MAO-B. To verify the validity of the MAO assay, dopamine was
assayed in the presence of clorgyline and L-deprenyl (MAO B) in
order to obtain the specific MAO A activity, respectively (Fig. 8,
insert).
the MAO A status in vivo. The organic syntheses of the precursors
and the reference compounds were established and optimized. For
radiolabeling of [18F]F-d2-MH and [18F]FEH a two-step procedure
was chosen. Both compounds were obtained after synthesis times
from 85 to 115 min in high RCY of about 30–47%. Using toluenesul-
fonyl leaving groups and a microwave-supported radiosynthesis it
was possible to increase the RCY of the direct 18F-fluorinations of
the PEGylated harmol compounds up to 45%. Plasma stability tests
in human serum showed that all tracers are stable over an incuba-
tion time of 60 min, showing 95% or higher for the unbound intact
tracer. Surprisingly, the 18F-labeled PEGylated compounds showed
The calculated IC50 were 34
tively, matching with the published KM value of 285
7
l
M and 160 23
l
M, respec-
M using
l
[3H]dopamine as substrate and PC12 cells as source of total MAO
activity.46 The amount to block MAO A activity decreased in the
order FEH, harmine, F-d2-MH and harmol Fig. 9) with a range from
0.54 0.06 nM for FEH to 208 16 nM for harmol. Still, the IC50
value for harmine (2.5 0.3 nM) and F-d2-MH (12.2 0.6 nM) were
in the low nanomolar range.
no significant brain uptake in first lPET studies. A possible reason
for this behavior might be the reduced lipophilicity and an
increased formation of H-bonds caused by the PEGylation, thus
preventing the passage of the blood–brain barrier. In contrast,
Clorgyline and L-deprenyl (37.5 nM each) were used to specifi-
cally inhibit MAO A and MAO-B activity. These concentrations
were the mid points between the lg IC50 values of the two respec-
tive dose–response curves (Fig. 8), allowing to monitor effectively
and precisely the MAO B and MAO A activities as the IC50 values
differ by at least 300-fold.
[
18F]F-d2-MH showed a high brain uptake and a good specificity,
albeit this was also accompanied by a fast pharmacokinetic profile.
l
PET studies performed for
[
18F]FEH showed that 18F]FEH
[
revealed very promising properties for the in vivo visualization of
the MAO A, like high brain uptake, high specificity to MAO A and
reversibility of binding. In MAO binding assays a high affinity for
the MAO A was observed for F-d2-MH and FEH, showing IC50 values
of 12.2 0.6 nM and 0.54 0.06 nM, respectively. Surprisingly, F-
d2-MH also inhibited the MAO-B, though with a 1000-fold lower
The MAO-B activity of harmol, harmine and FEH was negligible
in the solubility range (Fig. 10). F-d2-MH displayed an IC50 of
11
4 lM, nearly 1000-fold lower than for MAO A. Harmol, har-
mine and FEH are highly MAO A selective compounds, whereas
F-d2-MH, surprisingly, substantially inhibited MAO-B though with
potency (IC50(MAO-B): 11
4 lM). It remains to be evaluated
a 1000-fold lower potency (11
4 lM and 12.2 0.6 nM for MAO-B
whether the increase in affinity from harmol over F-d2-MH to
FEH can be extended to fluoro-butyl-harmine and fluoro-t-butyl-
harmine. Although [18F]F-d2-MH showed very fast pharmacokinet-
ics in in vivo imaging studies, it is nevertheless a potential
candidate for MAO-A imaging, since it has been shown be that
and MAO A, respectively). Moreover, the F-d2-MH efficacy was only
78% to inhibit MAO B in contrast to the action against MAO A,
which amounted to nominally 100% for all compounds tested.
The differences in the washout between F-d2-MH and FEH in the

H. Schieferstein et al. / Bioorg. Med. Chem. 23 (2015) 612–623
623
21. Ametamey, S. M.; Beer, H.; Antonini, A.; Leenders, K. L.; Schubiger, P. A. Nucl.
Med. Biol. 1996, 23, 229.
22. Kersemans, K.; Van Laeken, N.; De Vos, F. J. Labelled Compd. Radiopharm. 2013,
56, 78.
23. Dolle, F.; Valette, H.; Bramoulle, Y.; Guenther, I.; Fuseau, C.; Coulon, C.;
Lartizien, C.; Jegham, S.; George, P.; Curet, O.; Pinquier, J.-L.; Bottlaender, M.
Bioorg. Med. Chem. Lett. 2003, 13, 1771.
24. Fowler, J. S.; Ding, Y. S.; Logan, J.; MacGregor, R. R.; Shea, C.; Garza, V.; Gimi, R.;
Volkow, N. D.; Wang, G. J.; Schlyer, D.; Ferrieri, R.; Gatley, S. J.; Alexoff, D.;
Carter, P.; King, P.; Pappas, N.; Arnett, C. D. Nucl. Med. Biol. 2001, 28, 779.
25. Kim, H.; Sablin, S. O.; Ramsay, R. R. Arch. Biochem. Biophys. 1997, 337, 137.
26. Bergström, M.; Westerberg, G.; Kihlberg, T.; Långström, B. Nucl. Med. Biol. 1997,
24, 381.
species differences between animals and humans can drastically
influence the pharmacokinetics of a tracer.24 Hence, for both com-
pounds,
indicated.
[ [ a further evaluation is
18F]F-d2-MH and 18F]FEH,
Acknowledgments
This work was supported by the research cluster SAMT of the
Johannes Gutenberg-University Mainz. We thank Kerstin Lüd-
dens-Dämgen for technical assistance.
27. Blom, E.; Karimi, F.; Eriksson, O.; Hall, H.; Långström, B. J. Labelled Compd.
Radiopharm. 2008, 51, 277.
28. Smart, B. E. J. Fluorine Chem. 2001, 109, 3.
References and notes
29. Müller, K.; Faeh, C.; Diederich, F. Science 2007, 317, 1881.
30. Bell, R. P. Chem. Soc. Rev. 1973, 3, 513.
31. Zhang, M.-R.; Maeda, J.; Ito, T.; Okauchi, T.; Ogawa, M.; Noguchi, J.; Suhara, T.;
Halldin, C.; Suzuki, K. Bioorg. Med. Chem. 2005, 13, 1811.
32. Beyerlein, F.; Piel, M.; Höhnemann, S.; Rösch, F. J. Labelled Compd. Radiopharm.
2013, 56, 360.
33. Block, D.; Knochel, A.; Hamburg, U.; Beckmann, R. J. Labelled Compd.
Radiopharm. 1986, 23, 467.
34. Mohler, D. L.; Shen, G. Org. Biomol. Chem. 2006, 4, 2082.
35. Riss, P. J.; Hoehnemann, S.; Piel, M.; Roesch, F. J. Labelled Compd. Radiopharm.
2013, 56, 356.
36. Ucar, G. Hacettepe Univ. Eczacilik Fak. Derg. 2004, 24, 25.
37. Bergström, M.; Westerberg, G.; Németh, G.; Traut, M.; Gross, G.; Greger, G.;
Müller-Peltzer, H.; Safer, A.; Eckernäs, S. A.; Grahnér, A.; Långström, B. Eur. J.
Clin. Pharmacol. 1997, 52, 121.
38. Herth, M. M.; Debus, F.; Piel, M.; Palner, M.; Knudsen, G. M.; Lüddens, H.;
Rösch, F. Bioorg. Med. Chem. Lett. 2008, 18, 1515.
39. Lüddens, H.; Lang, H. J.; Korpi, E. R. Eur. J. Pharmacol. 1998, 344, 269.
40. Krajl, M. Biochem. Pharmacol. 1965, 14, 1684.
41. Lueddens, H. W.; Newman, A. H.; Rice, K. C.; Skolnick, P. Mol. Pharmacol. 1986,
29, 540.
1. Hare, M. L. C. Biochem. J. 1928, 22, 879.
2. Zeller, E. A. Helv. Chim. Acta 1937, 21, 880.
3. Finberg, J. P. M.; Gillman, K. Int. Rev. Neurobiol. 2011, 100, 169.
4. Shih, J. C.; Chen, K.; Ridd, M. J. Annu. Rev. Neurosci. 1999, 22, 197.
5. Lan, N. C.; Chen, C. H.; Shih, J. C. J. Neurochem. 1989, 52, 1652.
6. Bach, A. W.; Lan, N. C.; Johnson, D. L.; Abell, C. W.; Bembenek, M. E.; Kwan, S.
W.; Seeburg, P. H.; Shih, J. C. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4934.
7. Tong, J.; Meyer, J. H.; Furukawa, Y.; Boileau, I.; Chang, L. J.; Wilson, A. A.; Houle,
S.; Kish, S. J. J. Cereb. Blood Flow Metab. 2013, 33, 863.
8. Strolin Benedetti, M.; Dostert, P.; Tipton, K. F. Dev. Pharmacol. Ther. 1992, 18,
191.
9. Tsang, D.; Ho, K. P.; Wen, H. L. Dev. Neurosci. 1986, 8, 243.
10. Nicotra, A.; Pierucci, F.; Parvez, H.; Senatori, O. Neurotoxicology 2004, 25, 155.
11. Kalgutkar, A. S.; Dalvie, D. K.; Castagnoli, N.; Taylor, T. J. Chem. Res. Toxicol.
2001, 14, 1139.
12. Ginovart, N.; Meyer, J. H.; Boovariwala, A.; Hussey, D.; Rabiner, E. A.; Houle, S.;
Wilson, A. A. J. Cereb. Blood Flow Metab. 2006, 26, 330.
13. Livingston, M. G.; Livingston, H. M. Drug Saf. 1996, 14, 219.
14. Cohen, G. J. Neural Transm. Suppl. 1990, 32, 229.
15. Youdim, M. B. H.; Edmondson, D.; Tipton, K. F. Nat. Rev. Neurosci. 2006, 7, 295.
16. Vaya, J.; Aluf, Y.; Finberg, J. P. M. In Oxidative Stress and Free Radical Damage in
Neurology; Gadoth, N., Gobel, H. H., Eds.; Springer: New York, 2012; pp 191–
224.
17. Da Prada, M.; Zürcher, G.; Wüthrich, I.; Haefely, W. E. J. Neural Transm. Suppl.
1988, 26, 31.
42. Gabay, S.; Achee, F. M.; Mentes, G. J. Neurochem. 1976, 27, 415.
43. Harmer, C. L.; Burns, J. E.; Sams, A.; Spittle, M. Clin. Radiol. 1969, 20, 204.
44. Pardridge, W. M. J. Cereb. Blood Flow Metab. 2012, 32, 1959.
45. Stein, W. D. The Molecular Basis of Diffusion Across Cell Membranes; Academic
Press: New York, NY, 1967; pp 66–125.
18. Blob, L. F.; Sharoky, M.; Camobell, B. J.; Kemper, E. M.; Gilmor, M. G.;
VanDenberg, C. M. CNS Spectr. 2007, 12, 25.
19. Stahl, S. M.; Felker, A. CNS Spectr. 2008, 13, 855.
46. Youdim, M. B.; Heldman, E.; Pollard, H. B.; Fleming, P.; McHugh, E. Neuroscience
1986, 19, 1311.
20. Bergström, M.; Westerberg, G.; Långström, B. Nucl. Med. Biol. 1997, 24, 287.