Synthesis and: In vitro evaluation of fluorine-18 benzimidazole sulfones as CB2 PET-radioligands

Article abstract of DOI:10.1039/c9ob00656g

Cannabinoid type 2 receptor (CB2) is up-regulated on activated microglial cells and can potentially be used as a biomarker for PET-imaging of neuroinflammation. In this study the synthesis and pharmacological evaluation of novel fluorinated pyridyl and et

Full text of DOI:10.1039/c9ob00656g

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Synthesis and in vitro evaluation of fluorine-18
benzimidazole sulfones as CB2 PET-radioligands†
Cite this: DOI: 10.1039/c9ob00656g
Annukka Kallinen,a Rochelle Boyd,b Samuel Lane,c Rajiv Bhalla,d Karine Mardon,d
Damion H. R. Stimson,d Eryn L. Werry,a Roger Fulton,c Mark Connor b and
a
Michael Kassiou
*
Cannabinoid type 2 receptor (CB2) is up-regulated on activated microglial cells and can potentially be
used as a biomarker for PET-imaging of neuroinflammation. In this study the synthesis and pharmacologi-
cal evaluation of novel fluorinated pyridyl and ethyl sulfone analogues of 2-(tert-butyl)-5-((2-fluoropyri-
din-4-yl)sulfonyl)-1-(2-methylpentyl)-1H-benzo[d]imidazole (rac-1a) are described. In general, the
ligands showed low nanomolar potency (CB2 EC50 < 10 nM) and excellent selectivity over the CB1
subtype (>10 000×). Selected ligands 1d, 1e, 1g and 3l showing high CB2 binding affinity (Ki < 10 nM) were
radiolabelled with fluorine-18 from chloropyridyl and alkyl tosylate precursors with good to high isolated
radioactive yields (25–44%, non-decay corrected, at the end of synthesis). CB2-specific binding of the
radioligand candidates [18F]-1d and [18F]-3l was assessed on rat spleen cryosections using in vitro auto-
radiography. The results warrant further in vivo evaluation of the tracer candidates as prospective CB2
PET-imaging agents.
Received 21st March 2019,
Accepted 1st May 2019
DOI: 10.1039/c9ob00656g
rsc.li/obc
neurodegenerative diseases.8–10 Microglial activation is
accompanied by changes in receptor expression patterns,
Introduction
Neuroinflammation is an innate immune system response which offers opportunities for detecting activated cells by
observed in a variety of central nervous system disorders specific PET-radioligands. The field has largely focused on the
including Alzheimer’s and Parkinson’s diseases (AD, PD), translocator protein 18 kDa (TSPO) for the detection of in vivo
demyelinating diseases such as multiple sclerosis and trau- neuroinflammation, however, TSPO imaging poses challenges,
matic brain injury.1–4 It is characterised by the activation of as the target is not specific for microglia but is also encoun-
microglial cells which can differentiate into either pro- or anti- tered in astrocytes.11 There is also a considerable amount of
inflammatory phenotypes, often classified as M1 and M2, TSPO present in the endothelium, which may complicate
respectively.5,6 However, recent evidence suggests that the dual image analysis, as the signal arising from the vascular TSPO
polarization is likely to be an over-simplification of a more binding has to be accounted for.12 More importantly, most of
dynamic system, where activated microglia represent a multi- the current TSPO-radioligands show variable affinity for the
tude of phenotypes, rather than two opposing functional target in the general population due to a common single
states.7
nucleotide polymorphism (SNP) rs6971 in TSPO.13,14 This
Positron Emission Tomography (PET) is a non-invasive leads to interindividual differences in radioligand binding,
molecular imaging technique that has shown promise in which requires prior genotyping for the SNP for valid quantifi-
detecting activated microglia and neuroinflammation in cation of TSPO imaging.15 New PET-radioligands targeting
other molecular biomarkers of activated microglia are
thus needed to understand the role of microglia in
neurodegeneration.
Cannabinoid receptor 2 is a G-protein coupled receptor
(GPCR) which predominantly signals through Gαi/o subunits.
It is present in low amounts in the healthy brain, whereas a
marked increase in the expression level is observed in micro-
glia under inflammatory conditions.16,17 Microglial CB2 acti-
vation is considered to relate to the early neuroprotective
phase of CNS inflammation, contrary to the TSPO up-regu-
lation that is thought to occur as part of the delayed microglial
aSchool of Chemistry, The University of Sydney, NSW 2006, Australia.
E-mail: michael.kassiou@sydney.edu.au; Tel: +612 9351 2745
bBiomedical Sciences, Faculty of Medicine and Health Sciences,
Macquarie University, NSW 2109, Australia
cFaculty of Health Sciences, The University of Sydney, NSW 2050, Australia
dThe Centre for Advanced Imaging, The University of Queensland, QLD 4072,
Australia
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c9ob00656g
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
activation in a pro-inflammatory fashion.18–20 Specific CB2 models of inflammation, and in models of AD and PD.31,32
radioligands could be used as radiopharmacological tools to The neuroprotective and anti-inflammatory effects of CB2-ago-
investigate pathophysiological mechanisms in the earliest nists may thus offer a new therapeutic approach for the treat-
phases of neuroinflammation, and to potentially image a ment of neurodegenerative diseases. A CB2-specific PET-radi-
different pool of microglia, thus complementing data gathered oligand would also be of great importance in the evaluation of
with TSPO radioligands.
the efficacy of novel CB2-based therapeutics.
There have been several CB2 PET-radioligands developed to
Sulfonyl benzimidazole-based CB2-agonists were originally
date (Fig. 1). Despite these efforts, no candidates have pro- reported by Pfizer and Janssen in the late 2000s.33,34 To the
gressed from the preclinical and clinical development phases best of our knowledge, benzimidazole-based ligands have not
into routine clinical use. 2-Oxoquinoline [11C]NE40 is the first been evaluated as CB2-radioligands thus far. In this study, we
CB2-radioligand with promising pharmacokinetic properties, describe the design and synthesis of new fluorinated benzimi-
and has been evaluated in clinical studies in healthy volun- dazole sulfones and their in vitro pharmacological characteris-
teers and AD patients (Fig. 1).21,22 However, [11C]NE40 was ation. Selected ligands were 18F-radiolabelled as candidate
found not to accumulate in the β-amyloid deposits, and no co- radioligands and evaluated for the specific binding on rat
localisation of CB2 and amyloid deposits could be observed.22 spleen cryosections using in vitro autoradiography.
4-Oxoquinoline [11C]RS-016 was observed to have CB2-specific
binding in LPS-induced neuroinflammation model in mice,
however, the compound was suspected to suffer from poor
Results and discussion
Synthesis and in vitro evaluation of the fluorinated
blood–brain barrier permeability.23 Pyridine [11C]RSR-056
showed increased brain uptake under inflammatory con-
ditions in LPS mouse model, however, only part of the binding
was found to be CB2-specific.24 In the initial studies, thiazole
[11C]A-836339 was observed to have CB2-specific binding in AD
mouse model, but failed to show increased uptake in the
inflammation area in animal models of neuroinflammation
and in cerebral ischemia.25 Triazine based ligand [18F]-19, on
the other hand, was found to suffer from rapid metabolism
and elimination, hindering its use as a CB2-radioligand.26
Oxadiazole [18F]MA3 was recently evaluated in non-human pri-
mates and was found to have moderate uptake and washout
from the brain, but CB2-specific binding could not be
observed.27 CB2-radioligands developed so far have been
recently reviewed, providing a thorough update on the topic.28
CB2 receptors also represent a largely unexploited drug
target that has attracted recent interest.29,30 Selective CB2-ago-
nists attenuate pro-inflammatory cytokine release and
migratory activity of immune system cells in several animal
benzimidazole 5-sulfones
The lead compound rac-1a (Scheme 1) was selected from the
patent literature, as it has been reported to have high potency
at hCB2 (EC50 0.26 nM) and excellent selectivity over hCB1
subtype (EC50 hCB1/hCB2 >38 900) in an assay measuring the
inhibition of forskolin induced cAMP production (CHO-K1 cell
line).34 It was noted that the lead structure contains an elec-
tron deficient pyridine ring with fluoro-substituent in 2-posi-
tion that was proposed to readily undergo SNAr, making the
incorporation of 18F facile in the structure. As ethyl sulfone
benzimidazoles have shown potential for good brain uptake,
we also included aliphatic 2-fluoroethyl sulfones (2) in the
screening.35 N1-2-Fluoroethoxyethyl chain was proposed as an
alternative position for the incorporation of 18F-radiolabel in
the ethyl sulfone structures (derivative 3l). This was surmised
to result in potentially improved stability, as primary aliphatic
18F-substituents β to a heteroatom have been reported to
Fig. 1 Examples of recent CB2-radioligands: [11C]NE40, [11C]RS-016, [11C]RSR-056, [11C]A-836339, [18F]-19 and [18F]MA3.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 1 Synthetic pathway A to make fluorinated benzimidazole sulfones 1 and 2. Reagents and conditions: (i) 4-Methoxy-α-toluenethiol, KOH,
ethanol, 80 °C (avg. >99%); (ii) pivaloyl chloride, pyridine, DMAP, DCM, RT (avg. 76%); (iii) Fe, NH4Cl, ethanol, 40 °C or sonication at RT (avg. 81%); (iv)
R1CHO, H+ cat., DCM; then NaBH4, methanol, 40 °C (55–82%); (v) p-TsOH, toluene or DMF, 130 °C (70–79%); (vi) TFA, 80 °C; (vii) 2-fluoro-4-bromo-
pyridine, Pd2(dba)3, Xantphos, Cs2CO3, toluene, 110 °C (62–66% over 2 steps); (viii) 2-fluoroethylbromide, Cs2CO3, dioxane, 50 °C (61–65% over 2
steps); (ix) m-CPBA, DCM, 0 °C → RT (avg. 74%).
undergo in vivo defluorination at a slower rate than isolated
Unmasking the PMB-protected sulfides in refluxing TFA
afforded free thiol intermediates that were immediately used
aliphatic fluorides.36,37
The lead compound rac-1a and the analogues were syn- for the next step. Pd-Catalysed C–S cross-coupling with
thesised following Phillip’s method based on the conden- 2-fluoro-4-bromopyridine or substitution with 2-fluoroethyl-
sation of ortho-dianilines with carboxylic acid derivatives.34,38 bromide gave the sulfide intermediates 8 and 9, respectively
The synthesis started from commercially available nitroben- (Schemes 1 and 2). In these reactions, formation of the di-
zenes 4 (route A, Scheme 1), 11a or 11b (route B, Scheme 2). sulfide by-products could not be avoided, which lowered the
Following the route A,
4 was reacted with 4-methoxy- isolated yields of the sulfides. In the final step, the sulfides
α-toluenethiol to give PMB-protected aryl sulfide in high yield. were oxidised with m-CPBA or potassium peroxymonosulfate
The aniline was acylated with pivaloyl chloride, followed by (Oxone) to afford the benzimidazole sulfone target compounds
reduction of the nitro-group to give the intermediate 5. (1a–f, 1h, 2a–e and 2h) (Schemes 1 and 2).
Reductive amination with pentanal, rac-2-methylpentanal or
Methyl esters sulfides 8h and 9h were readily converted
cyclobutane carbaldehyde was used to derivatise the into primary alcohols (8i, 9i) by DIBAL-H reduction, which was
N1-nitrogen, which afforded intermediates 6a, 6b and 6c. followed by oxidation to the corresponding sulfones 1i and 2i
Condensation of the intermediates 6a–c was carried out under (Scheme 2). The carboxylic acid derivatives 1k and 2k were
acidic conditions at elevated temperature to form benzimida- accessed directly from 1h and 2h. The ester hydrolysis was first
zole core (7a–c).
trialled with aqueous sodium hydroxide, however, this was
Alternatively, N1-alkyl side chains d, e, f and k were intro- found to result in the formation of pyridone by-product via
duced via SNAr reaction with the starting material 11a/11b and SNAr reaction at the 2-fluoropyridine moiety. The hydrolysis
the required amines in the first step (Scheme 2, route B). In was then performed with trimethyltinhydroxide in refluxing
the case of 4-aminobutanoic acid derivative (k), the SNAr 1,2-dichloroethane, which afforded the desired carboxylic acid
product was protected as methyl ester (h) for the subsequent derivatives (1k, 2k) in high yield. Primary amides (1j and 7j)
steps. Nitro-reduction and acylation of the resulting aniline could be readily accessed from the corresponding carboxylic
with pivaloyl chloride were conducted as described earlier to acids 1k and 7k using CDI as the coupling reagent in the pres-
give pivalamide intermediates 12d, 12e, 12f and 12h which ence of ammonium chloride in basic conditions. Intermediate
were then subjected to condensation, as above, to give the 7j was taken through the steps of deprotection and substi-
corresponding bromo-substituted benzimidazoles. The bro- tution to give the intermediate 9j. Primary amides 1j and 9j
mides could then be coupled with 4-methoxy-α-toluenethiol were then treated with trifluoroacetic anhydride and triethyl-
under Pd-catalysis to give the key PMB-sulfide intermediates amine to afford the nitriles 1g and 9g, after which the latter
7d, 7e, 7f and 7h.
was oxidised to the target compound 2g (Scheme 2).
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 2 Synthetic pathway B to make fluorinated benzimidazole sulfones 1 and 2. Reagents and conditions: (i) NHR1, DIPEA, ethanol, 100 °C
(38–89%); (ii) Fe, NH4Cl, ethanol, 40 °C (82–95%); (iii) pivaloyl chloride, pyridine, DMAP, DCM, RT (68–98%); (iv) p-TsOH, toluene or DMF, 130 °C
(58–90%); (v) 4-methoxy-α-toluenethiol, Pd2(dba)3, Xantphos, DIPEA, dioxane, 90 °C (44–91%); (vi) TFA, 80 °C; (vii) 2-fluoro-4-bromopyridine,
Pd2(dba)3, Xantphos, Cs2CO3, toluene, 110 °C (38–86% over 2 steps); (viii) 2-fluoroethylbromide, Cs2CO3, dioxane, 50 °C (42–80% over 2 steps); (ix)
m-CPBA, DCM, 0 °C → RT (avg. 72%) or KHSO5·1/2KHSO4·1/2K2SO4, methanol, H2O, 0 °C → RT (54%); (x) LiOH, THF/MeOH/H2O, RT (96%); (xi) CDI,
NH4Cl, DIPEA, DMF, RT (70–85%); (xii) DIBAL-H, THF, −78 °C → RT (avg. 97%); (xiii) TFAA, TEA, DCM, 0 °C → RT (82–91%); (xiv) trimethyltinhydroxide,
1,2-DCE, 80 °C (83–88%).
The synthesis of analogue 3l followed the same steps as onist radioligands may interact with and detect different recep-
described in Scheme 2 with some modifications. Here, pivala- tor pools in vivo.39 The lead compound rac-1a was confirmed
mide intermediate 12m was prepared in 61% yield in a four- to act as an agonist with similar high efficacy (Emax 89%) to
step telescoped process starting from 11b and glycine the standard CP-55940 (Table 1). rac-1a had a good, but
(Scheme 3). The carboxylic acid was protected as a bulky slightly lower potency at CB2 in this assay (EC50 = 3.5 nM)
isopropyl ester to slow down the intramolecular trans- when compared to earlier data (EC50 = 0.26 nM).34 The appar-
lactamisation with the aniline in the subsequent step. ent selectivity over CB1 subtype (240×) was significantly lower
p-TsOH catalysed condensation of the dianiline 12m was than the value reported earlier (>38 900×).34 It is to be noted
carried out in DMSO at 100 °C to give 13m which underwent though, that the lead compound was measured here as a race-
Pd-catalysed C–S coupling with 4-methoxy-α-toluenethiol to mate, and the information on the stererochemical purity was
afford 7m. The isopropyl ester was reduced by DIBAL-H to not disclosed in the previous reports.34 The apparent CB2/CB1
give 2-hydroxy ethyl chain (7n). The pendant hydroxyl group selectivity of compounds in functional assays also depends on
was then subjected to O-alkylation with 2-fluoroethyl bromide the relative amount of receptors expressed in each cell line,
to give an intermediate 7l. Deprotection, followed by and so direct comparisons between different experiments are
S-alkylation, gave the sulfide 10l that was lastly oxidised to difficult.
sulfone 3l.
The N1-alkyl chain variants (a–c) were included in the first
The ligands were evaluated for their functional activity in a screening to gauge whether the potency and selectivity is
fluorometric membrane potential assay that measures the acti- affected by the geometry of the N1-alkyl chain (Table 1). Both
vation of G-protein-coupled inwardly rectifying potassium the flexible, linear side chain and the conformationally
channels (GIRKs) in an AtT-20 cell line. This was done to study restricted cyclobutyl ring resulted in low nanomolar potency at
the intrinsic activity of the compounds, as agonist and antag- CB2 in the membrane potential assay. Typically, non-opti-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 3 Synthesis of fluorinated benzimidazole sulfone 3l. Reagents and conditions: (i) Glycine, DIPEA, ethanol, 100 °C; (ii) iPrOH, H2SO4, 60 °C; (iii)
Fe, NH4Cl, ethanol, 40 °C or sonication at RT; (iv) pivaloyl chloride, pyridine, DMAP, DCM, RT (61% over four steps); (v) p-TsOH, DMSO, 100 °C (74%); (vi)
4-methoxy-α-toluenethiol, Pd2(dba)3, Xantphos, DIPEA, dioxane, 90 °C (84%); (vii) TFA, 80 °C; (viii) ethylbromide, TBAI, NaOH (aq.), MeOH, 60 °C (99%
over two steps); (ix) DIBAL-H, THF, −78 °C → RT (avg. 88%); (x) 2-fluoroethylbromide, NaH, THF, 60 °C (avg. 76%); (xi) m-CPBA, DCM, 0 °C → RT (72%).
Table 1 Activity of fluorinated benzimidazole sulfones with chains a–c at CB1 and CB2 in fluorometric membrane potential assay
CB1
CB2
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
tPSAe
clog Pd (Å2)
Compd R1 R2
SIb
1a
1b
1c
2a
2b
2c
R1: rac-CH2CH(CH3)(CH2)2CH3 6.07 0.66 (843) 0.51 0.22 (48) 8.46 0.30 (3.5) 0.91 0.09 (89)
241 5.35
62.1
62.1
62.1
49.7
49.7
49.7
R2: 2-Fluoropyridyl
R1: (CH2)4CH3
R2: 2-Fluoropyridyl
R1: c-Butyl CH2
R2: 2-Fluoropyridyl
d.n.c.c
d.n.c.
d.n.c.
d.n.c.
d.n.c
8.05 0.14 (9.0) 0.98 0.05 (96) >10 000 4.95
8.49 0.12 (3.3) 0.97 0.04 (95) >10 000 4.36
8.50 0.15 (3.1) 0.94 0.05 (92) >10 000 4.73
R1: rac-CH2CH(CH3)(CH2)2CH3 d.n.c.
R2: 2-Fluoroethyl
R1: (CH2)4CH3
R2: 2-Fluoroethyl
R1: c-Butyl CH2
R2: 2-Fluoroethyl
6.48 0.27 (333) 0.59 0.10 (57) 8.65 0.15 (2.2) 0.89 0.05 (87)
d.n.c. (20 840) d.n.c. 8.37 0.11 (4.3) 0.99 0.04 (97)
149 4.33
>4900 3.75
a Ligand efficacy given as maximum response Emax
standard error of the mean, and as intrinsic activity compared to 1 μM CP-55940.
b Selectivity index given as a ratio EC50 hCB1/EC50 hCB2. c d.n.c. = data not converted, EC50 > 10 μM. d clog P values were predicted with
ChemDraw Professional version 15.0.0.107, BioByte fragment based method. e Topological polar surface area, calculated with ChemDraw
Professional version 15.0.0.107.
mised cannabinoid receptor agonists suffer from high lipophi- reducing the free fraction of radioligand available for crossing
licity. Accordingly, predicted lipophilicity of the lead rac-1a BBB and binding to the target.39 According to previously
(clog P 5.35, ChemDraw Professional version 15.0.0.107, reported studies, bulky and lipophilic tert-butyl group in the
BioByte fragment based method) exceeds the optimal range 2-position of the imidazole ring is optimal for the activity.34,41
(clog P 2–4) for CNS-targeted ligands.40 Too high lipophilicity To reduce lipophilicity, H-bond acceptor and donor groups
may result in high non-specific binding to plasma proteins, were introduced to the N1-substituent.
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Hydrogen bond acceptors. We initially tested linear methoxy
Hydrogen bond donors. According to Verbist et al., the N1-
propyl, linear N,N-dimethylamino propyl and bulky tetrahydro- substituent in benzimidazole 5-sulfones fits in a hydrophobic
4-pyranyl methyl side chains in the N1-area (Table 2). The binding site and only H-bond acceptors are tolerated.41 In
linear methoxy propyl and N,N-dimethylamino propyl side the previous studies on indole cannabinoids, terminal
chains resulted in excellent CB2 subtype selectivity (>10 000×), hydroxyl group at N1-alkyl chain has been observed to
whereas the bulky tetrahydro-4-pyranyl methyl derivatives (1e increase the CB2 potency and selectivity over CB1.42–44 Also
and 2e) showed micromolar activities at CB1 receptor, lowering in this series, 4-hydroxybutyl substituent resulted in nano-
the selectivity ratio to 100–200×. It is possible that increased molar potency at CB2 (EC50 = 8.6 nM and 37 nM for 1i and
steric bulk of the tetrahydropyranyl moiety may result in 2i, respectively) with excellent CB2-subtype selectivity
binding at both CB1 and 2, whereas linear chains may confer (>10 000×) (Table 3). However, conspicuous decrease in CB2
better CB2 selectivity. In general, all the ligands tested showed activity was associated with primary carboxamide (1j and 2j)
low nanomolar potency at CB2. Of the pyridyl derivatives, 1d and furthermore, in carboxylic acid H-bond donors (1k and
had the highest potency (EC50 = 3.7 nM) and apparent selecti- 2k). The same observation has been reported with indole
vity over CB1 (>10 000×). Accordingly, the ethyl sulfone ana- cannabinoids.42,43 It is to be noted though, that the binding
logue 3l with an isosteric ethoxy ethyl chain was found to have orientation of the ligands is likely to vary, and more than
high activity at CB2 (EC50 = 5.1 nM) and excellent selectivity one plausible binding site may exist, complicating the com-
over CB1 (>10 000×). Since the methoxy group is potentially parison of the SAR-data between different ligand classes.45
metabolically labile, analogues for methoxypropyl chain with The potency for CB2 was regained by removing the H-bond
similar bioactivity were sought from the literature database donor functionality, as seen in the methyl ester analogues
(SwissBioIsostere®). This prompted us to include the propyl (1h and 2h, Table 2). When comparing the matched pairs
nitriles (1g and 2g) in the screening. Gratifyingly, 1g also across the H-bond acceptor and donor series, the fluoroethyl
showed promise as a highly selective (ratio >10 000×) and sulfones showed, in general, lower CB2 potency than 2-fluor-
potent (EC50 = 3.0 nM) CB2-ligand.
opyridyl sulfones.
Table 2 Activity of fluorinated benzimidazole sulfones with chains d–h and l at CB1 and CB2 in fluorometric membrane potential assay
CB1
CB2
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
tPSAe
Compd R1 R2
SIb
clog Pd (Å2)
1d
1e
1f
R1: (CH2)3OCH3
R2: 2-fluoropyridyl
d.n.cc
d.n.c.
8.43 0.02 (3.7) 1.06 0.02 (105) >10 000 3.05
71.3
71.3
65.3
85.9
88.4
59.0
59.0
73.5
76.0
59.0
R1: Tetrahydro-4-pyranyl CH2 5.97 0.04 (1064) 1.01 0.03 (100) 8.28 0.02 (5.2) 1.06 0.01 (104)
R2: 2-Fluoropyridyl
R1: (CH2)3N(CH3)2
R2: 2-Fluoropyridyl
R1: (CH2)3CN
R2: 2-Fluoropyridyl
R1: (CH2)3COOCH3
R2: 2-Fluoropyridyl
R1: (CH2)3OCH3
R2: 2-Fluoroethyl
R1: Tetrahydro-4-pyranyl CH2 6.31 0.04 (486)
R2: 2-Fluoroethyl
204 3.08
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
8.07 0.07 (8.4) 1.00 0.04 (98)
>10 000 3.20
1g
1h
2d
2e
2g
2h
3l
8.52 0.04 (3.0) 1.06 0.02 (106) >10 000 2.63
8.17 0.04 (6.7) 1.01 0.02 (101) >10 000 3.23
7.77 0.04 (17)
1.04 0.03 (103) >10 000 2.44
112 2.47
1.06 0.03 (105) 8.36 0.05 (4.3) 1.07 0.03 (105)
R1: (CH2)3CN
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
8.05 0.04 (8.9) 1.03 0.02 (102) >10 000 2.01
8.07 0.09 (8.6) 1.06 0.05 (105) >10 000 2.61
8.29 0.11 (5.1) 1.02 0.05 (100) >10 000 2.53
R2: 2-Fluoroethyl
R1: (CH2)3COOCH3
R2: 2-Fluoroethyl
R1: CH2CH2OCH2CH2F
R2: Ethyl
a Ligand efficacy given as maximum response Emax
standard error of the mean, and as intrinsic activity compared to 1 μM CP-55940.
b Selectivity index given as a ratio EC50 hCB1/EC50 hCB2. c d.n.c. = data not converted, EC50 > 10 μM. d clog P values were predicted with
ChemDraw Professional version 15.0.0.107, BioByte fragment based method. e Topological polar surface area, calculated with ChemDraw
Professional version 15.0.0.107.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Table 3 Activity of fluorinated benzimidazole sulfones with chains i–k at CB1 and CB2 in fluorometric membrane potential assay
CB1
CB2
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
pEC50 SEM
(EC50 nM)
Emax SEMa
(% CP 55940)
tPSAe
Compd
R1 R2
SIb
clog Pd
(Å2)
1i
R1: (CH2)4OH
d.n.c.c
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
d.n.c.
8.06 0.06 (8.6)
6.98 0.08 (105)
5.58 0.08 (2630)
7.43 0.04 (37)
6.28 0.09 (520)
d.n.c.
1.06 0.04 (106)
1.06 0.04 (105)
1.04 0.07 (104)
1.05 0.02 (104)
0.96 0.06 (95)
d.n.c.
>10 000
2.43
82.3
R2: 2-Fluoropyridyl
R1: (CH2)3CONH2
R2: 2-Fluoropyridyl
R1: (CH2)3COOH
R2: 2-Fluoropyridyl
R1: (CH2)4OH
R2: 2-Fluoroethyl
R1: (CH2)3CONH2
R2: 2-Fluoroethyl
R1: (CH2)3COOH
R2: 2-Fluoroethyl
1j
>10 000
>10 000
>10 000
>10 000
N/A
2.01
2.80
1.82
1.40
2.18
105.2
99.4
70.0
92.8
87.0
1k
2i
2j
2k
a Ligand efficacy given as maximum response Emax
standard error of the mean, and as intrinsic activity compared to 1 μM CP-55940.
b Selectivity index given as a ratio EC50 hCB1/EC50 hCB2. c d.n.c. = data not converted, EC50 > 10 μM. d clog P values were predicted with
ChemDraw Professional version 15.0.0.107, BioByte fragment based method. e Topological polar surface area, calculated with ChemDraw
Professional version 15.0.0.107.
Overall, the control ligands rac-1a, 1c and 1e, showed lower
CB2-potency and subtype selectivity in the GIRK-based assay,
as compared to the values reported earlier in the assay measur-
ing cAMP inhibition.34 However, in general, functional activity
data is highly dependent on the assay conditions, cell line
used and the secondary messenger system used to measure
the receptor activation, and such data from different labora-
tories are difficult to compare.
Table 4 Apparent binding affinities (Ki) of the compounds 1d, 1g, 1e, 2e
and 3l. Ki values were measured after competitive displacement of 2 nM
(CB2) or 3 nM (CB1) of [3H]CP-55940
Based on the functional activity screening, a few of the
ligands were selected further for the radiolabelling studies.
Potency, selectivity, lipophilicity and topological polar surface
area of the compounds were considered. Of the pyridyl sul-
fones, 1d (EC50 = 3.7 nM, clog P = 3.05, tPSA = 71.3) and 1g
(EC50 = 3.0 nM, clog P = 2.63, tPSA = 85.9) were selected due to
their low nanomolar potency and excellent subtype selectivity.
Despite the lower subtype selectivity, the pyridyl sulfone ana-
logue 1e (EC50 = 5.2 nM, clog P = 3.08, tPSA = 71.3) was
included as a back-up ligand with potentially different phar-
macokinetic properties. 2-Fluoroethyl sulfone 2e (clog P = 2.47,
tPSA = 59.0) and ethyl sulfone 3l (clog P = 2.53, tPSA = 59.0)
Compd
R1 R2
Ki CB1 (nM)
Ki CB2 (nM)
1d
R1: (CH2)3OCH3
R2: 2-Fluoropyridyl
R1: (CH2)3CN
R2: 2-Fluoropyridyl
R1: Tetrahydro-4-pyranyl CH2
R2: 2-Fluoropyridyl
R1: Tetrahydro-4-pyranyl CH2
R2: 2-Fluoroethyl
R1: CH2CH2OCH2CH2F
R2: Ethyl
>10 000
2.6 1.0
1g
1e
2e
3l
>10 000
374 170
228 54
>10 000
3.5 1.7
0.3 0.1
1.1 0.5
18.1 7.2
were chosen from the alkyl sulfone series due to their high The measurements were conducted on membrane fractions
potency (EC50 = 4.3 nM and 5.1 nM, respectively). The clog P prepared from CHO cells expressing hCB1 or hCB2 receptors
and tPSA values of the candidate ligands are within the range and the inhibition constants (Ki) of the ligands were determined
of optimal lipophilicity (clog P 2–4) and topological polar using the Cheng–Prusoff equation.47 The binding affinity data
surface area (tPSA < 90 Å2) set for passive entry into CNS, were in good agreement with the functional activity measure-
further supporting the selection of these compounds.40
ments, as all the ligands were found to have low nanomolar
Next, the selected ligands were investigated for their binding affinities at CB2 (Table 4). In general, the binding
binding affinity towards CB2. Competitive binding of the affinities were comparable to those of the CB2 PET-radioligands
ligands was measured against a known radioligand standard reported so far, e.g. [11C]A-836339 (Ki = 0.6 nM), [18F]-MA3 (Ki =
[3H]CP-55940, a non-selective cannabinoid receptor agonist.46 0.8 nM) and [11C]NE40 (Ki = 9.6 nM).21,25,27
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Radiolabelling of the pyridyl sulfone ligands
(0.9%). Using the same method, the other tracer candidates
[18F]-1e and [18F]-1g were produced in 26 6% (n = 3) and 28%
(n = 1) isolated, radioactive yields (EOS, n.d.c.) with high radio-
chemical purity (>99%). The molar activities of [18F]-1d, [18F]-
1e and [18F]-1g were measured to be 263 56 GBq μmoL−1 (n =
8), 158 32 GBq μmoL−1 (n = 3) and 128 GBq μmoL−1 (n = 1),
respectively, at the end of synthesis. The formulated products
[18F]-1e and [18F]-1g were confirmed to be stable for up to
4 hours in a solution consisting of 10% ethanol in saline
(0.9%) with 0.5% sodium ascorbate (w/v).
18F-Fluorination of pyridyl derivatives was first trialled with tri-
methylammonium- and chloro-leaving groups (Scheme 4).
Precursor 14d was prepared in two steps: 1d underwent SNAr
substitution with dimethylamine and the product was sub-
sequently N-methylated to afford the trimethylammonium salt
14d. 2-Chloropyridine precursors (15d, 15e and 15g) were pre-
pared from the protected aryl sulfide intermediates 7d, 7e and
7j following the same chemical conversions as described in
the Scheme 2. Deprotection followed by Pd-catalysed C–S-coup-
ling with 2-chloro-4-bromopyridine gave the biaryl sulfides
that were then oxidised to the corresponding sulfones 15d, 15e
and 15g.
Low activity radiolabelling test reactions were carried out
with dried [18F]-[K(K2.2.2)]F complex in DMSO at 105 °C and at
130 °C for 15 min, using precursor 14d or 15d, respectively.
Lower reaction temperature was chosen for the trimethyl-
ammonium precursor, as ammonium salts have a rather poor
stability in higher temperatures.48 The incorporation of [18F]F−
was found to be very similar in these reactions – 65% and 70%
for 14d and 15d, respectively, which is rather surprising given
the poorer leaving group ability of the chloride in the pyridine
SNAr.49 However, the dimethylamino side product, formed by
reverse Menschutkin reaction48 from 14d, had a very similar
retention time to the product [18F]-1d, which complicated the
purification process. Based on the test reactions, we chose
chloro as the leaving group for the further radiolabelling.
The radiosynthesis of [18F]-1d was fully automated onto the
Synthra RN plus module. Radioactive yields of the reactions
were calculated based on the activity measurement at the start
of synthesis. The synthesis was tested with 20–41 GBq of start-
ing activity, which afforded [18F]-1d with high, isolated radio-
active yield (24 5% at EOS, non-decay corrected, n = 8). The
yield was sufficient for the preclinical production of [18F]-1d
and was not further optimised. Radiochemical purity of the
formulated product was analysed to be >99%, and the product
was confirmed to be stable for 4 hours in 10% ethanol/saline
Radiolabelling of the ethyl sulfone ligands
Precursor 16e was prepared from the protected sulfide 7e
(Scheme 5A) according to the steps described earlier. After de-
protection, the thiol intermediate was alkylated with 2-bromo-
ethanol, and the resulting sulfide oxidised to sulfone. The
pendant hydroxyl group was then reacted with tosyl chloride
giving the labelling precursor 16e. Precursor 18, on the other
hand, was prepared from the key alcohol intermediate 7n. The
chain was first extended by alkylation, followed by ester
reduction to give the alcohol intermediate 17. This was then
subjected to tosylation to give the precursor 18.
Attempts to radiolabel tosylate 16e failed to give the desired
product [18F]-2e (for details, see ESI†). Whilst our initial ana-
lysis of the 19F-standard 2e in 10% ethanol/saline (0.9%) indi-
cated that this compound was stable for greater than 3 hours
at room temperature, our subsequent studies confirmed that
2e does completely degrade when subjected to the labelling
conditions (1 mg mL−1 in DMF, 5 μmol of K2.2.2., 5 μmol of
K2CO, 150 °C, 5 min). It is possible that 2e undergoes E1cb
elimination in basic conditions at high temperature to form
vinyl sulfone by-products. The observed instability of the
target compound 2e and the precursor 16e sets limitations on
reaction conditions and basicity of the reaction medium, and
further optimisation would be required.
The radiolabelling of the precursor 18 was carried out in
acetonitrile at 100 °C for 10 min resulting in high fluorine-18
Scheme 4 Synthesis of the labelling precursors 14d, 15d, 15e and 15g. Reagents and conditions: (i) N,N-Dimethylamine, DIPEA, ethanol, 80 °C
(avg. 86%); (ii) CH3I, CH3OH, 40 °C (77%); (iii) TFA, 80 °C; (iv) 4-Br-2-Cl-pyridine, Pd2(dba)3, Xantphos, Cs2CO3, toluene, 100 °C (43–87% over two
steps); (v) m-CPBA, DCM, 0 °C → RT (avg. 92%); (vi) TFAA, TEA, DCM, 0 °C → RT (75%).
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
Scheme 5 Synthesis of the tosylate precursors 16e and 18 for the radiolabelling of ethyl sulfone derivatives. Reagents and conditions: (i) TFA,
80 °C; (ii) 2-bromoethanol or ethylbromide, TBAI, NaOH (aq.), MeOH, 60 °C (88–97% over two steps); (iii) m-CPBA, DCM, 0 °C → RT (avg. 92%); (iv)
TsCl, pyridine, DMAP, DCM, 40 °C (40–43%); (v) 2-bromoacetic acid, NaH, TBAI, THF, 60 °C; MeOH, H2SO4, 60 °C (74% over two steps); (vi)
DIBAL-H, THF, −78 °C → RT (avg. 88%).
incorporation yield of 74% (HPLC). These conditions were and the use of Na-ascorbate as a radical scavenger was used
transferred to Synthra RNplus module, resulting in a high, iso- for the final production runs. Na-Ascorbate was added in the
lated radioactive yield of [18F]-3l (43 4%, n.d.c., n = 4) at the HPLC-purification (0.5% w/v), SPE reformulation (5 mg mL−1
)
end of synthesis. However, the product was observed to and in the final formulated product (0.5% w/v). Radiochemical
undergo radiolysis with higher starting activities, and conse- purity of the product [18F]-3l was found to be >97% and was
quently, a lower starting activity (max. 6 GBq of [18F]fluoride) observed to be stable for over two hours in the final formu-
Fig. 2 Representative autoradiograms of rat spleen tissues after incubation with [18F]-3l (top) and [18F]-1d (bottom). TB = total binding; OB = off-
target binding determined by co-incubation of the radioligands with GW-405833 (10 μM); NB = non-displaceable binding determined by co-incu-
bation with 19F-standards 3l and 1d (10 μM).
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
lation. The molar activity of [18F]-3l was measured to be 115
64 GBq μmoL−1 at the end of synthesis.
18F-Labelled radiotracer candidates could be readily syn-
thesised from 2-chloro pyridyl or alkyl tosylate precursors in
high radiochemical yield and purity. In the in vitro autoradiog-
raphy on rat spleen, CB2-specific binding of the selected radi-
oligands accounted for 30–40% of the total binding, however,
a substantial amount of the total binding was assigned to off-
targets. Yet, the observed off-target binding may be highly
dependent on species and organ/tissue under study, and may
not translate into high CB2-unrelated binding in general or in
the target organ brain. Thus, further in vivo evaluation of the
CB2-radiotracer candidates presented here in a neuroinflam-
mation model is warranted.
In vitro autoradiography
As a preliminary evaluation, the CB2-specific binding of two
representative compounds, one from each ligand group [18F]-
1d and [18F]-3l, was investigated using in vitro autoradiography
on rat spleen that has a high level of CB2-receptors under
basal conditions (Fig. 2).50 Even though pyridyl sulfone [18F]-
1e had the highest CB2-affinity of the tested ligands, [18F]-1d
was selected, as it not only has high CB2-affinity (Ki = 2.6 1.0
nM) but also excellent selectivity over CB1 subtype. CB2-
specific binding was determined by incubating the radioli-
gands in the presence of a known CB2-selective agonist
GW-405833 (10 μM, Ki rCB2 = 3.6 1.1 nM; Ki rCB1 = 273
42.6 nM).51 Non-displaceable, non-specific binding was deter-
mined by co-incubation of the spleen sections with the radioli-
gands and the corresponding 19F-standards (10 μM). CB2-
specific binding accounted for 32 14% and 44 7% of the
total binding, whereas the non-displaceable binding was
Experimental
The ESI† contains details of the chemical synthesis, radio-
synthesis, HPLC-analytical conditions and experimental
details of in vitro autoradiography and pharmacological assays.
observed to be 34 10% and 11 9% of the total binding for Chemistry
[18F]-1d and [18F]-3l, respectively (1σ). The CB2-specific com-
Example of synthesis of compound 1a via route A.
N-(2-Amino-5-((4-methoxybenzyl)thio)phenyl)pivalamide (5)
ponent represented thus approximately 48% of the displace-
able binding in both cases.
General method i. 5-Chloro-2-nitrobenzeneamine
(23 mmol, 4.0 g) was dissolved in ethanol (60 mL) under N2
atmosphere. Potassium hydroxide (46 mmol, 2.6 g) and
4-methoxytoluene thiol (3.2 mL, 23 mmol) were added to the
stirring solution that was then heated to mild reflux. After 2 h,
reaction mixture was cooled down to room temperature and
the yellow precipitate formed was filtered off. The solids were
washed with small amount of ethanol. The crude product was
dried under vacuum to give 5-((4-methoxybenzyl)thio)-2-nitroa-
niline as yellow crystalline solid that was used without purifi-
cation for the next step.
4
A few CB2-radioligands developed thus far, e.g. [11C]NE40
and [11C]RS-016, have shown high CB2-specific binding in
murine spleen autoradiography in vitro.23,52 However, many
CB2-radioligands have displayed high non-specific binding in
spleen autoradiograms.53–55 For instance, high non-specific
binding was reported for [18F]MA3, which was suggested to
result from either suboptimal autoradiography protocol or a
high koff value of the radioligand.27 Despite the inconclusive
autoradiography data, [18F]MA3 was evaluated in a rat model
with local striatal overexpression of hCB2, and was observed to
show CB2-specific binding in vivo.27 Whether the high off-
target binding observed in rat spleen in this study translates
into off-target binding in the brain in inflammatory conditions
in other species, requires further evaluation.
General method ii. 5-((4-Methoxybenzyl)thio)-2-nitroaniline
(3.0 g, 10 mmol) was weighed in a flask and dissolved in dry
dichloromethane (50 mL) under N2 atmosphere. Pyridine
(1.2 mL, 16 mmol) and 4-dimethylaminopyridine (0.13 g,
10 mol%) were added to a stirring solution at room tempera-
ture. Pivaloyl chloride (1.5 mL, 1.2 equiv.) was added at RT,
after which solution was heated to mild reflux overnight. Once
the reaction was complete, the solution was diluted with DCM
Conclusions
In summary, a new series of fluorinated pyridyl and ethyl sul- and water. Aqueous phase was extracted with DCM and com-
fonyl benzimidazole CB2-ligands were synthesised by varying bined organic phase was washed with 0.1 M HCl (25 mL) and
the N1-alkyl chain. The ligands were found to be agonists with brine. Crude product was dried over MgSO4 and concentrated
high efficacy, comparable to the standard CP-55940. Most of to give a solid that was purified on silica gel chromatography
the ligands showed low nanomolar potency at CB2 and excel- using gradient Hex → EtOAc : Hex (1 : 9) → EtOAc : Hex (1 : 5).
lent subtype selectivity over the CB1 receptor. Based on the This afforded N-(5-((4-methoxybenzyl)thio)-2-nitrophenyl)piva-
in vitro evaluation, the N1-area can be used for adjusting lamide as a light yellow crystalline solid (2.69 g, 72% over two
physicochemical properties of the ligands without losing steps).
activity, however, carboxylic acid or amide H-bond donor
M.p. 94.5 °C; 1H NMR (300 MHz, CDCl3): δ 1.36 (9H, s),
groups are not tolerated. The linear alkyl chains in the N1-area 3.79 (3H, s), 4.24 (2H, s), 6.85 (2H, d, J = 7.5 Hz), 6.93 (1H, d,
were observed to result in high CB2-selectivity. The ethyl sul- J = 9 Hz), 7.33 (2H, d, J = 7.8 Hz) 8.10 (1H, d, J = 9 Hz), 8.90
fones showed, in general, slightly lower CB2 potency than (1H, s), 11.02 (1H, bs); 13C NMR (75 MHz): δ 27.6 (3 × C), 36.3,
2-fluoropyridyl sulfones.
40.9, 55.4, 114.3 (2 × C), 116.8, 120.3, 126.1, 127.1, 130.4 (2 × C),
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
132.7, 136.2, 150.6, 159.3, 178.4; HRMS (ESI-TOF) m/z calcu- hydrate (0.17 g, 0.90 mmol) was added. Solution was heated to
lated for C19H22N2O4SN [M + Na]+ 397.1193, found 397.1194.
130 °C in a sealed flask for over night. After completion of the
General method iii. N-(5-((4-Methoxybenzyl)thio)-2-nitro- reaction, solution was cooled down to room temperature and
phenyl)pivalamide (2.00 g, 5.47 mmol) was weighed and dis- diluted with EtOAc. The organic phase was washed with satu-
solved in ethanol (50 mL) under N2 atmosphere. The solution rated NaHCO3 and brine, and dried over anhydrous MgSO4.
was heated to 40 °C. NH4Cl (2.80 g, 10 equiv.) was dissolved in Filtration and concentration under reduced pressure gave the
water (10 mL) and was added to the stirring solution. Iron crude product that was purified on silica gel using EtOAc : Hex
powder (3.06 g, 10 equiv.) was added portionwise to a vigor- gradient (1 : 6 → 1 : 3) afforded the product as light yellow
ously stirring solution. After 2 h, solution was cooled down to solid (0.17 g, 70%).
1
room temperature, and was diluted with DCM and water.
M.p. 78.1 °C; H NMR (300 MHz, CDCl3): δ 0.97 (3H, d J =
Solution was filtered through a pad of basic alumina and 6.9 Hz), 0.99 (3H, t, J = 6.6 Hz), 1.38 (4H, m), 1.66 (9H, s), 2.42
Celite® and eluted with DCM/methanol mixture. Filtrate was (1H, m), 3.88 (3H, s), 4.17 (2H, s), 4.14–4.32 (2H, 2 × dd, J = 7.2
concentrated and diluted with DCM. The aqueous layer was Hz, J = 14.7 Hz), 6.88 (2H, d, J = 8.4 Hz), 7.27 (4H, m), 7.93
separated and the organic phase was washed with brine and (1H, s); 13C NMR (75 MHz, CDCl3): δ 14.3, 17.6, 20.2, 30.5 (3 ×
dried over MgSO4. Crude product was concentrated and dried C), 33.5, 34.7, 37.1, 40.6, 51.7, 55.3, 110.8, 113.9 (2 × C), 122.7,
under vacuum to give the title product (5) as light orange crys- 126.1, 128.3, 130.05 (2 × C), 130.15, 136.3, 142.4, 158.7, 162.0;
talline solid (1.65 g, 87%).
M.p. 124.8 °C; H NMR (300 MHz): δ 1.33 (9H, s), 3.78 (3H, 411.2465, found 411.2463.
s), 3.92 (2H, s), 6.68 (1H, d, J = 8.4 Hz), 6.79 (2H, d, J = 8.1 Hz),
7.01 (1H, d, J = 8.1 Hz), 7.10 (2H, d, J = 8.4 Hz), 7.14 (1H, s),
7.28 (1H, bs); 13C NMR (75 MHz): δ 27.9 (3 × C), 39.5, 40.9,
55.4, 110.1, 113.9 (2 × C), 118.7, 124.8, 126.7, 129.9, 130.2 (2 ×
C), 131.9, 139.9, 158.7, 177.3; HRMS (ESI-TOF) m/z: calculated
for C19H24N2O2S [M + Na]+ 367.1451, found 367.1451.
HRMS (ESI-TOF) m/z: calculated for C25H34N2OS [M + H]+
1
rac-2-(tert-Butyl)-5-((2-fluoropyridin-4-yl)thio)-1-(2-methylpentyl)-
1H-benzo[d]imidazole (8a)
General method vi. rac-2-(tert-Butyl)-5-((4-methoxybenzyl)thio)-
1-(2-methylpentyl)-1H-benzo[d]imidazole (83 mg, 0.20 mmol)
rac-N-(5-((4-Methoxybenzyl)thio)-2-((2-methylpentyl)amino) was dissolved in trifluoroacetic acid (4 mL) and was refluxed at
phenyl)pivalamide (6a) 80 °C for overnight. Reaction mixture was cooled down to
General method iv. N-(2-Amino-5-((4-methoxybenzyl)thio) room temperature and evaporated to dryness under N2 flow.
phenyl)pivalamide 5 (0.50 g, 1.46 mmol) was dissolved in dry The resulting light yellow residue was used directly for the next
DCM (25 mL) in a two-necked flask under N2 atmosphere. rac- step.
2-Methyl pentanal (0.2 mL, 1.6 mmol) and a few drops of
General method vii. A two-necked flask and a condenser were
glacial acetic acid were added to a stirring solution at room dried under vacuum with a heatgun and the system was left to
temperature. Solution was refluxed for 4 h, after which reac- cool down to room temperature under N2 atmosphere.
tion was cooled down to room temperature. NaBH4 (0.22 g, 4 Toluene (4 mL) was added to the flask and degassed.
equiv.) and MeOH (dry, 7 mL) were added gradually over Xantphos® (6 mg, 5 mol%) and Pd2(dba)3 (9 mg, 5 mol%)
1.5 hours to a stirring solution at room temperature. After were added to the flask and the mixture was degassed.
completion of the reaction, the mixture was diluted with water 4-Bromo-2-fluoropyridine (21 μL, 0.20 mmol) was added to the
and DCM. The water layer was separated and extracted with reaction. Lastly, the crude 2-(tert-butyl)-1-(2-methylpentyl)-1H-
DCM (3 × 20 mL). The combined organic phase was washed benzo[d]imidazole-5-thiol (in 3 mL toluene) and Cs2CO3
with brine and dried over MgSO4. Purification on basic (98 mg, 0.3 mmol) were added. The reaction mixture was
alumina using Hex → EtOAc : Hex (1 : 10) → (1 : 8) → (1 : 6) gave refluxed at 110 °C for over night. After completion of the reac-
the title product as orange solid (0.39 g, 62%).
tion, the reaction was diluted with EtOAc and sat. NaHCO3
1
M.p. 72.9 °C; H NMR (300 MHz, CDCl3): δ 0.91 (3H, t, J = and filtered through a pad of silica and Celite® with EtOAc.
6.9 Hz), 0.98 (3H, d, J = 6.6 Hz), 1.20 (2H, m), 1.33 (9H, s), 1.40 The organic phase was separated, dried over anhydrous
(2H, m), 1.74 (1H, m), 2.86 and 2.97 (2H, 2 × dd, J1 = 6 Hz, J2 = MgSO4, filtered and concentrated to give a residue that was
11.7 Hz), 3.77 (3H, s), 3.92 (2H, s), 6.64 (1H, d, J = 8.4 Hz), 6.79 purified on silica gel using EtOAc : Hex (1 : 6), which afforded
(2H, d, J = 8.4 Hz), 7.08 (1H, m), 7.12 (2H, d, J = 8.7 Hz), 7.21 the title product 8a as light yellow oil (50 mg, 65% over
(1H, d, J = 1.8 Hz), 7.26 (1H, bs); 13C NMR (75 MHz, CDCl3): 2 steps).
δ 14.4, 18.2, 20.2, 27.9, 32.9, 37.2, 39.5, 41.1, 50.6, 55.4, 110.1,
1H NMR (300 MHz, CDCl3): δ 0.89 (3H, t, J = 6.9 Hz), 0.92
113.7, 113.8, 124.6, 130.15, 130.2 (2 × C), 130.4, 132.3, 142.7, (3H, d, J = 8.1 Hz), 1.31 (4H, m), 1.57 (9H, s), 2.33 (1H, m), 4.13
158.7, 177.4; HRMS (ESI-TOF) m/z: calculated for (1H, dd, J1 = 14.7 Hz, J2 = 9.0 Hz), 4.24 (1H, dd, J1 = 14.7 Hz,
C25H36N2O2SH [M + H]+ 429.2570, found 429.2571.
J2 = 6.9 Hz), 6.40 (1H, s), 6.81 (1H, d, J1 = 5.4 Hz), 7.35 (1H, d,
rac-2-(tert-Butyl)-5-((4-methoxybenzyl)thio)-1-(2-methylpentyl)- J = 9.6 Hz), 7.41 (1H, d, J = 8.4 Hz), 7.91 (1H, d, J = 5.4 Hz),
1H-benzo[d]imidazole (7a)
7.97 (1H, s); 13C NMR (75 MHz, CDCl3): δ 14.3, 17.7, 20.2, 30.5
General method v. rac-N-(5-((4-Methoxybenzyl)thio)-2-((2- (3 × C), 33.6, 34.9, 37.1, 51.9, 105.3 (1C, d, J = 40.7 Hz), 112.3,
methylpentyl)-amino)phenyl)pivalamide (0.26 g, 0.60 mmol) 118.1 (1C, d, J = 3.7 Hz) 119.9, 127.6, 129.1, 138.3, 142.9, 146.8
was dissolved in toluene in a pressure flask and p-TsOH mono- (1C, d, J = 16.1 Hz), 157.7 (1C, d, J = 8.4 Hz), 164.2 (1C, d, J =
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
237 Hz), 163.3; 19F NMR (258 MHz): −68.8 (s); LRMS (ESI) m/z:
estimated 386.21 [M + H]+ and 408.19 [M + Na]+; found 386.24
and 408.21, respectively; HRMS (ESI-TOF) m/z: calculated for
C22H28FN3S [M + H]+ 386.2061, found 386.2060.
rac-2-(tert-Butyl)-5-((2-fluoropyridin-4-yl)sulfonyl)-1-(2-methyl-
pentyl)-1H-benzo[d]imidazole (1a)
General method x. rac-2-(tert-Butyl)-5-((2-fluoropyridin-4-yl)
thio)-1-(2-methylpentyl)-1H-benzo[d]imidazole (50 mg, 0.13 mmol)
was dissolved in dry DCM (4 mL) under N2 flow. m-CPBA (77%,
64 mg, 0.29 mmol) was added gradually at 0 °C. Once com-
5 J. D. Cherry, J. A. Olschowka and J. K. O’Banion,
J. Neuroinflammation, 2014, 11, 1–15.
6 J. A. Kabba, Y. Xu, H. Christian, W. Ruan, K. Chenai,
Y. Xiang, L. Zhang, J. M. Saavedra and T. Pang, Cell. Mol.
Neurobiol., 2018, 38, 53–71.
7 R. M. Ransohoff, Nat. Neurosci., 2016, 19, 987–991.
8 C. Cerami, L. Iaccarino and D. Perani, Int. J. Mol. Sci., 2017,
18, 1–14.
9 B. Pulli and J. W. Chen, J. Clin. Cell. Immunol., 2014, 5, 1–
44.
plete, the reaction was diluted with DCM, and aqueous 10 B. Janssen, D. J. Vugts, U. Funke, G. T. Molenaar,
sodium metabisulfate (5%, 4 mL) and sat. NaHCO3 (5 mL)
were added. The organic phase was washed with sat. NaHCO3
(2 × 5 mL) and brine (5 mL), and was dried over anhydrous
P. S. Kruijer, B. N. van Berckel, A. A. Lammertsma and
A. D. Windhorst, Biochim. Biophys. Acta, 2016, 1862, 425–
441.
MgSO4. Filtration and concentration under reduced pressure 11 S. Lavisse, M. Guillermier, A. S. Hérard, F. Petit,
gave the crude material that was purified on silica gel using
EtOAc : Hex (1 : 6 → 1 : 4) gave the product 1a as yellow viscous
oil (42 mg, 78%).
M. Delahaye, N. Van Camp, L. Ben Haim, V. Lebon,
P. Remy, F. Dollé, T. Delzescaux, G. Bonvento, P. Hantraye
and C. Escartin, J. Neurosci., 2012, 32, 10809–10818.
1H NMR (300 MHz, CDCl3): δ 0.85 (3H, d, J = 6.6 Hz), 0.86 12 C. Wimberley, S. Lavisse, V. Brulon, M. A. Peyronneau,
(3H, t, J = 6.9 Hz), 1.35 (4H, m), 1.55 (9H, s), 2.28 (1H, m), 4.18
(1H, dd, J1 = 14.7 Hz, J2 = 6.9 Hz), 4.18 (1H, dd, J1 = 14.7 Hz,
C. Leroy, B. Bodini, P. Remy, B. Stankoff, I. Buvat and
M. Bottlaender, J. Nucl. Med., 2018, 59, 307–314.
J2 = 9 Hz), 7.41 (1H, m), 7.44 (1H, d, J = 8.7 Hz), 7.65 (1H, dt, 13 D. R. Owen, A. J. Yeo, R. N. Gunn, K. Song, G. Wadsworth,
J1 = 5.1 Hz, J2 = 1.5 Hz), 7.75 (1H, dd, J1 = 8.7 Hz, J2 = 1.8 Hz),
8.35 (1H, d, J = 5.4 Hz), 8.37 (1H, d, J = 1.8 Hz); 13C NMR
(75 MHz, CDCl3): δ 14.2, 17.6, 20.2, 30.3 (3 × C), 33.7, 35.1,
37.0, 52.0, 108.0 (1C, d, J = 40.2 Hz), 111.8, 118.8 (1C, d, J =
A. Lewis, C. Rhodes, D. J. Pulford, I. Bennacef, C. A. Parker,
P. L. StJean, L. R. Cardon, V. E. Mooser, P. M. Matthews,
E. A. Rabiner and J. P. Rubio, J. Cereb. Blood Flow Metab.,
2012, 32, 1–5.
5.0 Hz), 121.1, 121.5, 131.9, 141.0, 141.8, 149.5 (1C, d, J = 14.4 14 W. C. Kreisl, K. J. Jenko, C. S. Hines, C. H. Lyoo,
Hz), 156.1 (1C, d, J = 6.6 Hz), 163.8 (1C, d, J = 242.6 Hz), 165.1;
19F NMR (282 Hz): δ −63.4 (s); HRMS (ESI-TOF) m/z: [M + Na]+
calculated for C22H28FN3O2S 440.1779, found 440.1778; HPLC-
purity: >99%, RT = 23.6 min.
W. Corona, C. L. Morse, S. S. Zoghbi, T. Hyde,
J. E. Kleinman, V. W. Pike, F. J. McMahon and R. B. Innis,
J. Cereb. Blood Flow Metab., 2012, 33, 53–58.
15 D. R. Owen, R. N. Gunn, E. A. Rabiner, I. Bennacef,
M. Fujita, W. C. Kreisl, R. B. Innis, V. W. Pike, R. Reynolds,
P. M. Matthews and C. A. Parker, J. Nucl. Med., 2011, 52,
24–32.
Conflicts of interest
16 K. Maresz, E. J. Carrier, E. D. Ponomarev, C. J. Hillard and
B. N. Dittel, J. Neurochem., 2005, 95, 437–445.
17 G. A. Cabral and F. Marciano-Cabral, J. Leukocyte Biol.,
2005, 78, 1192–1997.
The authors declare no competing financial interest.
18 T. Hosoya, D. Fukumoto, T. Kakiuchi, S. Nishiyama,
S. Yamamoto, H. Ohba, H. Tsukada, T. Ueki, K. Sato and
Y. Ouchi, J. Neuroinflammation, 2017, 14, 1–9.
19 Y. Wang, X. Yue, D. O. Kiesewetter, G. Niu, G. Teng and
X. Chen, Eur. J. Nucl. Med. Mol. Imaging, 2014, 41, 1440–
1449.
References
1 R. Ransohoff, Science, 2016, 353, 777–783.
2 M. T. Heneka, M. J. Carson, J. El Khoury, G. E. Landreth,
F. Brosseron, D. L. Feinstein, A. H. Jacobs, T. Wyss-Corey,
J. Vitorica, R. M. Ranshohoff, K. Herrup, S. A. Frautschy, 20 I. Israel, A. Ohsiek, E. Al-Momani, C. Albert-Weissenberger,
B. Finsen, G. C. Brown, A. Verkhratsky, K. Yamanaka,
J. Koistinaho, E. Latz, A. Halle, G. C. Petzold, T. Town,
C. Stetter, S. Mencl, A. K. Buck, C. Kleinschnitz, S. Samnick
and A.-L. Sirén, J. Neuroinflammation, 2016, 13, 1–13.
D. Morgan, M. L. Shinohara, V. H. Perry, C. Holmes, 21 R. Ahmad, M. Koole, N. Evens, K. Serdons, A. Verbruggen,
N. G. Bazan, D. J. Brooks, S. Hunot, B. Joseph,
N. Deigendesch, O. Garaschuk, E. Boddeke, C. A. Dinarello,
G. Bormans and K. Van Laere, Mol. Imaging Biol., 2013, 15,
384–390.
C. J. Breitner, G. M. Cole, D. T. Golenbock and 22 R. Ahmad, A. Postnov, G. Bormans, J. Versijpt,
M. P. Kummer, Lancet Neurol., 2015, 14, 388–405.
3 G. Martino, L. Adorini, P. Rieckmann, J. Hillert,
M. Vandenbulcke and K. Van Laere, Eur. J. Nucl. Med. Mol.
Imaging, 2016, 43, 2219–2227.
B. Kallmann, G. Comi and M. Filippi, Lancet Neurol., 2002, 23 R. Slavik, A. Müller Herde, D. Bieri, M. Weber, R. Schibli,
1, 499–509.
S. D. Krämer, S. M. Ametamey and L. Mu, Eur. J. Med.
Chem., 2015, 92, 554–564.
4 J. W. Finnie, Inflammopharmacology, 2013, 21, 309–320.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2019

View Article Online
Organic & Biomolecular Chemistry
Paper
24 R. Slavik, U. Grether, A. Müller Herde, L. Gobbi, J. Fingerle, 41 B. M. P. Verbist, M. A. J. De Cleyn, M. Surkyn, E. Fraiponts,
C. Ullmer, S. D. Krämer, R. Schibli, L. Mu and
S. M. Ametamey, J. Med. Chem., 2015, 58, 4266–4277.
J. Aerssens, M. J. M. A. Nijsen and H. J. M. Gijsen, Bioorg.
Med. Chem. Lett., 2008, 18, 2574–2579.
25 G. Pottier, V. Gómez-Vallejo, D. Padro, R. Boisgard, 42 J. M. Frost, M. J. Dart, K. R. Tietje, T. R. Garrison,
F. Dollé, J. Llop, A. Winkeler and A. Martín, J. Cereb. Blood
Flow Metab., 2017, 37, 1163–1178.
26 S. Yrjölä, M. Sarparanta, A. J. Airaksinen, M. Hytti,
G. K. Grayson, A. V. Daza, O. F. El-Kouhen, B. B. Yao,
G. C. Hsieh, M. Pai, C. Z. Zhu, P. Chandran and
M. D. Meyer, J. Med. Chem., 2010, 53, 295–315.
A. Kauppinen, S. Pasonen-Seppänen, B. Adinolfi, P. Nieri, 43 S. D. Banister, J. Stuart, R. C. Kevin, A. Edington,
C. Manera, O. Keinänen, A. Poso, T. J. Nevalainen and
T. Parkkari, Eur. J. Pharm. Sci., 2015, 67, 85–96.
27 B. Attili, S. Celen, M. Ahamed, M. Koole, C. Van de Haute,
M. Longworth, S. M. Wilkinson, C. Beinat, A. S. Buchanan,
D. E. Hibbs, M. Glass, M. Connor, I. S. McGregor and
M. Kassiou, ACS Chem. Neurosci., 2015, 6, 1445–1458.
W. Vanduffel and G. Bormans, Br. J. Pharmacol., 2019, 176, 44 M. Longworth, M. Connor, S. D. Banister and M. Kassiou,
1481–1491. ACS Chem. Neurosci., 2017, 8, 1673–1680.
28 F. Spinelli, L. Mu and S. M. Ametamey, J. Labelled Compd. 45 X.-Q. Xie, J. Z. Chen and E. M. Billings, Proteins: Struct.,
Radiopharm., 2018, 61, 299–308. Funct., Genet., 2003, 53, 307–319.
29 M. Aghazadeh Tabrizi, P. G. Baraldi, P. A. Borea and 46 V. M. Showalter, D. R. Compton, B. R. Martin and
K. Varani, Chem. Rev., 2016, 116, 519–560. M. E. Abood, J. Pharmacol. Exp. Ther., 1996, 278, 989–999.
30 G. Navarro, P. Morales, C. Rodriguez-Cueto, J. Fernández- 47 Y.-C. Cheng and W. H. Prusoff, Biochem. Pharmacol., 1973,
Ruiz, N. Jagerovic and R. Franco, Front. Neurosci., 2016, 10,
1–11.
31 C. Turcotte, M. R. Blanchet, M. Laviolette and N. Flamand,
Cell. Mol. Life Sci., 2016, 73, 4449–44470.
32 T. Cassano, S. Calcagnini, L. Pace, F. De Marco, A. Romano
and S. Gaetani, Front. Neurosci., 2017, 11, 1–10.
33 T. Kato, K. Kon-I, Y. Yuki and K. Ando, Pfizer Inc.,
WO2007102059A1, 2007.
22, 3099–3108.
48 H. Sun and S. G. DiMagno, J. Fluorine Chem., 2007, 128,
806–812.
49 L. Dolci, F. Dollé, S. Jubeau, F. Vaufrey and C. Crouzel,
J. Labelled Compd. Radiopharm., 1999, 42, 975–985.
50 S. J. Govaerts, E. Hermans and D. M. Lambert,
Eur. J. Pharm. Sci., 2004, 23, 233–243.
51 K. Valenzano, L. Tafesse, G. Lee, J. E. Harrison,
J. M. Boulet, S. L. Gottshall, L. Mark, M. S. Pearson,
W. Miller, S. Shan, R. Rabadi, Y. Rotshteyn,
S. M. Chaffer, P. I. Turchin, D. A. Elsemore, M. Toth,
L. Koetzner and G. T. Whiteside, Neuropharmacology,
2005, 248, 658–672.
34 H. J. M. Gijsen, M. A. J. De Cleyn, M. Surkyn and
B.
M.
P.
Verbist,
Janssen
Pharmaceutica,
WO2008003665A1, 2008.
35 C. Watson, D. R. Owen, D. Harding, K. Kon-I, M. L. Lewis,
H. J. Mason, M. Matsumizu, T. Mukaiyama, M. Rodriguez-
Lens, A. Shima, M. Takeuchi, I. Tran and T. Young, Bioorg. 52 N. Evens, C. Vandeputte, C. Coolen, P. Janssen, R. Sciot,
Med. Chem. Lett., 2011, 21, 4284–4287.
V. Baekelandt, A. M. Verbruggen, Z. Debyser, K. V. Laere
36 A. N. French, E. Napolitano, H. F. V. Brocklin,
and G. M. Bormans, Nucl. Med. Biol., 2012, 39, 389–399.
R. N. Hanson, M. J. Welch and J. A. Katzenellenbogen, 53 L. Mu, R. Slavik, A. Müller, K. Popaj, S. Cermak, M. Weber,
J. Labelled Compd. Radiopharm., 1991, 30, 431–433.
R. Schibli, S. D. Krämer and S. M. Ametamey,
37 A. N. French, E. Napolitano, H. F. V. Brocklin,
Pharmaceuticals, 2014, 7, 339–352.
R. N. Hanson, M. J. Welch and J. A. Katzenellenbogen, 54 R. Slavik, A. Müller Herde, A. Haider, S. D. Krämer,
Nucl. Med. Biol., 1993, 20, 31–47.
M. Weber, R. Schibli, S. M. Ametamey and L. Mu,
J. Neurochem., 2016, 138, 874–886.
55 R. Slavik, D. Bieri, S. Cermak, A. Müller, S. D. Krämer,
M. Weber, R. Schibli, S. M. Ametamey and L. Mu, Chimia,
2014, 68, 208–210.
38 M. A. Phillips, J. Chem. Soc., 1928, 2393–2399.
39 V. W. Pike, Curr. Med. Chem., 2016, 23, 1818–1869.
40 S. A. Hitchcock and L. D. Pennington, J. Med. Chem., 2006,
49, 7559–7583.
This journal is © The Royal Society of Chemistry 2019
Org. Biomol. Chem.