Design of selective COX-2 inhibitors in the (aza)indazole series. Chemistry, in vitro studies, radiochemistry and evaluations in rats of a [18F] PET tracer

Article abstract of DOI:10.1080/14756366.2018.1501043

A series of novel derivatives exhibiting high affinity and selectivity towards the COX-2 enzyme in the (aza) indazole series was developed. A short synthetic route involving a bromination/arylation sequence under microwave irradiation and direct C–H activation were established in the indazole and azaindazole series respectively. In vitro assays were conducted and structural modifications were carried out on these scaffolds to furnish compound 16 which exhibited effective COX-2 inhibitory activity, with IC₅₀ values of 0.409 μM and an excellent selectivity versus COX-1. Radiolabeling of this most potent derivative [¹⁸F]16 was achieved after boron ester release and the tracer was evaluated in vivo in a rat model of neuroinflammation. All chemistry, radiochemistry and biological experimental data are discussed.

Full text of DOI:10.1080/14756366.2018.1501043

Journal of Enzyme Inhibition and Medicinal Chemistry
ISSN: 1475-6366 (Print) 1475-6374 (Online) Journal homepage: http://www.tandfonline.com/loi/ienz20
Design of selective COX-2 inhibitors in the
(aza)indazole series. Chemistry, in vitro studies,
radiochemistry and evaluations in rats of a [18F]
PET tracer
Jonathan Elie, Johnny Vercouillie, Nicolas Arlicot, Lucas Lemaire, Rudy
Bidault, Sylvie Bodard, Christel Hosselet, Jean-Bernard Deloye, Sylvie Chalon,
Patrick Emond, Denis Guilloteau, Frédéric Buron & Sylvain Routier
To cite this article: Jonathan Elie, Johnny Vercouillie, Nicolas Arlicot, Lucas Lemaire, Rudy
Bidault, Sylvie Bodard, Christel Hosselet, Jean-Bernard Deloye, Sylvie Chalon, Patrick Emond,
Denis Guilloteau, Frédéric Buron & Sylvain Routier (2019) Design of selective COX-2 inhibitors
in the (aza)indazole series. Chemistry, inꢀvitro studies, radiochemistry and evaluations in rats
of a [18F] PET tracer, Journal of Enzyme Inhibition and Medicinal Chemistry, 34:1, 1-7, DOI:
10.1080/14756366.2018.1501043
To link to this article: https://doi.org/10.1080/14756366.2018.1501043
© 2018 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group.
View supplementary material
Submit your article to this journal
View Crossmark data
Published online: 26 Oct 2018.
Article views: 33
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=ienz20

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
2018, VOL. 34, NO. 1, 1–7
https://doi.org/10.1080/14756366.2018.1501043
RESEARCH PAPER
Design of selective COX-2 inhibitors in the (aza)indazole series. Chemistry, in vitro
studies, radiochemistry and evaluations in rats of a [18F] PET tracer
b,c,d,e
a
b
b
Jonathan Eliea,b, Johnny Vercouillieb,c,d,e , Nicolas Arlicot
, Lucas Lemaire , Rudy Bidault , Sylvie Bodard ,
ꢀ ꢀ
ꢀ
ꢀ
Christel Hosseletb,c, Jean-Bernard Deloyec,f, Sylvie Chalonb, Patrick Emondb, Denis Guilloteaub,c,d,e, Frederic Buron
a
and Sylvain Routiera
a
b
c
ꢀ
ICOA, UMR CNRS 7311, University of Orleans, Orleans, France; UMR 1253, iBrain , Universite de Tours, Inserm, Tours, France; CERRP, Centre
d
e
d’Etude et de Recherche sur les Radiopharmaceutiques, Tours, France; CHRU, de Tours, Tours, France; INSERM CIC 1415, University of
Franc¸ois-Rabelais de Tours, Tours, France; Biopole Clermont-Limagne, Laboratoires Cyclopharma, Saint-Beauzire, France
f
^
ABSTRACT
ARTICLE HISTORY
Received 20 April 2018
Revised 11 July 2018
Accepted 11 July 2018
A series of novel derivatives exhibiting high affinity and selectivity towards the COX-2 enzyme in the (aza)
indazole series was developed. A short synthetic route involving a bromination/arylation sequence under
microwave irradiation and direct C–H activation were established in the indazole and azaindazole series
respectively. In vitro assays were conducted and structural modifications were carried out on these scaf-
folds to furnish compound 16 which exhibited effective COX-2 inhibitory activity, with IC50 values of
0.409mM and an excellent selectivity versus COX-1. Radiolabeling of this most potent derivative [18F]16
was achieved after boron ester release and the tracer was evaluated in vivo in a rat model of neuroinflam-
mation. All chemistry, radiochemistry and biological experimental data are discussed.
KEYWORDS
Cyclooxygenase; neuroin-
flammation; NSAID;
radiolabeling; PET; boronic
ester; (aza)indazoles
GRAPHICAL ABSTRACT
Introduction
new therapeutic approaches. In this context, a large panel of
molecular targets can be envisaged for positron emission tomog-
raphy (PET) exploration7.
An inflammatory reaction is a ubiquitous effective protective
mechanism, which includes the cascade activation of coordinated
chemical and cellular events. The role of inflammation is to restore
tissue homeostasis1. However, it can have either a beneficial effect
when it promotes repair or adverse consequences when it is
excessive or long-lasting. Compared to other organs, the brain is
characterised by a low regenerative capacity and specific immune
processes due to the presence of the blood-brain-barrier (BBB).
The immune response of the brain to various injuries is called
neuroinflammation, and includes a number of events, the main
one being the activation of microglial cells2. During this process,
microglia change from a resting to an activated state, which can
either mediate protective and regenerating mechanisms or on the
contrary aggravate injury, contributing to neurodegeneration3.
Neuroinflammation is involved in a number of neurodegenerative
disorders such as Alzheimer’s disease (AD) and Parkinson’s disease
(PD) as well as in several neuropsychiatric disorders such as aut-
ism, schizophrenia, and depression4–6. The in vivo detection and
quantification of neuroinflammation in several brain diseases can
therefore, be of high value for a better understanding of patho-
Among all the molecular pathways involved in the inflamma-
tion process, the cyclooxygenase (COX) enzyme that contributes
to the subsequent production of prostaglandins clearly plays a
central role. COX-2 is an inducible enzyme, which is expressed at
high concentrations at inflammation sites and malignant transfor-
mations compared to most normal tissues. This, associated with
the availability of COX-2-selective inhibitors, makes this enzyme
an ideal target in order to image inflammation8,9
.
Several selective COX-2 inhibitors have been reported in the
literature, among them, the Coxib family (celecoxib, rofecoxib), a
non-steroidal anti-inflammatory drug class (NSAIDs) which has
been extensively studied10. Several [18F]PET radiotracers to
image COX-2 have been developed over the past decade
(Figure 1)8,9. Although some of them have proved useful to
explore inflammation in a pre-clinical colorectal cancer model,
no specific COX-2 radiotracers are available to visualise brain
inflammation because of a poor brain penetrance which remains
the main challenge for all PET agents targeting the central ner-
physiological mechanisms, early diagnosis, and identification of vous system (CNS)11,12
.
ꢀ
ꢀ ꢀ
CONTACT Nicolas Arlicot
nicolas.arlicot@univ-tours.fr
UMR Inserm U930, CHRU de Tours 2 Bd Tonnelle – 37044 Tours Cedex 9, France; Frederic Buron
fred-
eric.buron@univ-orleans.fr; Sylvain Routier
These authors contributed equally to the project.
ß 2018 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
sylvain.routier@univ-orleans.fr ICOA, UMR CNRS 7311, University of Orleans, Orleans, France
ꢀ
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use,
distribution, and reproduction in any medium, provided the original work is properly cited.

2
J. ELIE ET AL.
Figure 1. Structure of some [18F]coxib tracers.
Figure 2. Objectives of the study.
Scheme 1. Synthesis of di(het)arylated indazoles 14–27. Reagents and conditions: (a) 2- bromobenzaldehyde (1.0 equiv), aniline (1.01 equiv), MgSO4 (5.0 equiv), THF,
100 ꢁC, lW, 1: 10 min, 99%; 2: 15min, 99%; 3: 1.5 h, 95%; 4: 1.5 h, 90%; (b) NaN3 (2.5 equiv), CuI (0.1 equiv), DMSO, 150ꢁC, lW, 5 min, 5: 99%; 6: 99%; 7: 17%; 10:
quant TLC; (c) DMF.DMA (2.0 equiv), THF, r.t., 1 h, 98%; (d) Oxone (2.1 equiv), MeOH/H2O (2/1), r.t., 20h, 83%; (e) NaOH (3.0 equiv), 150 ꢁC, lW, 5 min then HCl conc.,
85% from 9; (f) Br2 (1.0 equiv), AcOH/MeOH/CH2Cl2 (2/1/0.1), r.t., 4 h, 11: 95%; 12: 99%; 13: 92%; (g) (Het)Ar2-B(OH)2 (1.2 equiv.), Cs2CO3 (3.0 equiv), Pd(PPh3)4
(0.1 equiv), dioxane, 150ꢁC, lW, 1 h, structures and yields are indicated in Table 2.
The present study, therefore aimed to develop a novel series However, no reaction occurred in the presence of methanesulfonyl
of selective COX-2 inhibitors using novel 2,3-di(het)arylated 3 or methanesulfonamide 4. This lack of reactivity prompted us to
oxidise the thiomethyl function of 6 under classical conditions to
sulfone 7 which was isolated in 83% yield. Concerning the access
to 8, the presence of the sulfonamide acidic function on 4 seems
to inhibit the intramolecular reaction. To counteract this effect, we
decided to use N, N-dimethylformamidine as the sulfonamide pro-
tecting group. After synthesising 9 from 4 using DMF.DMA as a
reagent, the near quantitative annelation furnished 10 (TLC) which
was immediately subjected to deprotection in basic media and
led successfully to the desired indazole 8 in an 85% overall yield.
The bromination step occurred smoothly and the desired
products 11–13 were isolated in excellent yields. For the next C-
3 (het)arylations, we successfully achieved the Suzuki-Miyaura
cross-coupling reactions using several commercially available aryl-
boronic acids (1.2 equiv.) under classical conditions15–17. The
desired products 14–27 were isolated in moderate to excellent
yields (Table 2). Most of the reactions appeared to be complete
during the monitoring and variations in the efficiency of the
cross-coupling reaction were mostly related to the purification
step. Details of the structures and yield in final products are
given in Table 2.
(aza)indazole series derivatives and to establish structure-activity
relationships (Figure 2). After developing the synthetic pathways
and proving their efficiency in a library-building strategy, each
new compound was evaluated for its activity on COX-2 as well as
its selectivity over COX-1 when relevant. The most potent deriva-
tive compound was converted to 18 F PET tracer to perform in
vivo studies. Each step of this study, aimed at the de novo concep-
tion of potent 18 F radiolabeled ligands targeting COX-2, is pre-
sented and the results discussed.
Results and discussion
Chemistry
The 2H-indazoles 5–8 were synthesised using a two-step approach
including a Schiff base formation followed by intramolecular cyc-
lisation13,14 using microwave irradiation to reduce the reaction
time of each step from 12 h to a few minutes. The condensation
of aniline with 2-bromobenzaldehyde in presence of MgSO4
under microwave activation allowed access to imine 1–4 in
excellent yields. N-2 arylated-indazoles were obtained using a cop-
per-catalysed reaction with sodium azide, and compounds 5 and
To be able to define SAR, we next studied the incorporation of
a nitrogen atom in the central heterocyclic scaffold and a switch
6 were isolated in 95% and 56% yields, respectively (Scheme 1). to the azaindazole series. To obtain the expected derivatives, we

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
3
Table 1. Synthesis of 2-aryl azaindazoles 36–38.
Step 1
Step 2
Product (Yield %)[a]
Step 3
Entry
1
Product (Yield %)[a],[b]
Product (Yield %)[a]
28 (99)
29 (62)
30 (99)
31 (99)
32 (29)
33 (ND)
34 (45)
35 (23)
36 (66)
2
3
4
37 (34)
38 (58)
[a]Yields are indicated in isolated products
[b]Irradiation time 15 min. ND: Not Detected. Step 1: (a) 2- bromobenzaldehyde (1.0 equiv), aniline (1.01 equiv), MgSO4 (5.0 equiv), THF, 100 ꢁC, lW, 15 min. Step 2:
NaN3 (2.5 equiv), CuI (0.1 equiv), DMSO, 150 ꢁC, lW, 5 min. Step 3: Oxone (2.1 equiv), MeOH/H2O (2/1), r.t., 20 h.
adapted the previous synthetic strategy and employed, as starting and C-3 respectively) led to the most active derivative 18 (entry
materials, pyridines that were brominated and formylated in vici- 6). In C-4, an ethyl ether group considerably diminished the activ-
nal positions. While the first synthetic pathway reaction afforded ity (entry 8), which was partially restored when fluorine was
imines in near quantitative yields, building the azaindazoles and replaced by a CF3 group. To prove that the oxygen electron
the following sulfur oxidation steps appeared hazardous but fortu- donating effect is involved in the activity loss, we replaced the
nately afforded 36–38 as pure compounds (Table 1).
The last step consisted in introducing an aryl moiety in C-3
position. In the literature, the sole methodology to describe this
ether residue by a chlorine atom and observed that 21 (entry 9)
was as active as 18. In order to modulate the electronic density in
C-9, we replaced the C-3 aryl group with a heterocyclic scaffold. 2
functionalization in the azaindazole series used a direct arylation or 2,6-fluoropyridines and 3-methoxy groups led only to inactive
strategy15–17. In the present study, the reaction required employ- derivatives (entries 10,12) whereas the 2-alkoxy pyridines 24 and
ing a catalytic system including Pd(dppf)Cl2.CH2Cl2, Ag2CO3 as 26 exhibited moderate activity (entries 10–12 vs 13–14). Finally
base and water as solvent. This powerful system was used starting pyrimidine does not appear suitable to improve the pharmaco-
from compounds 36–38 and led to 2,3-diarylazaindazole deriva- phoric model as 27 (entry 15) was totally inactive. In conclusion,
tives 39–41 in fairly good yields (Scheme 2).
these modulations indicated that each modulation in the C-3 inda-
zolic position needs to be finely tuned to conserve activity on the
COX-2 target, which is very sensitive to electronic, lipophilicity
and steric parameters.
In vitro enzymatic assays
To complete our study, we replaced the methylsulfonyl group
with a sulfonamide group. The analog of 14 afforded the most
active derivative 16, which showed a very interesting IC50 at
400 nM. Again, modification of the C-3 aryl functional group pos-
ition led to a strong variability in efficiency as the fluorine dis-
placement from the para to the meta position (17, entry 17) led
to a 10-fold decrease in activity.
The derivatives 14–27 and compounds 39–41 were evaluated
using a COX Inhibitor Screening Kit from Cayman Chemical
Company (Table 2) to measure COX activity. Our set of molecules,
designed for COX-2 inhibition, appeared to be very selective
towards COX-1 as no efficient inhibition of this enzyme was
detected at 30 mM for any of the derivatives tested. In the indazole
series, all the derivatives showed an activity against the COX-2
enzyme. The SAR study showed that the best position for the 4-
sulfonylaryl moiety remained the N-2 position since its shift to C-3
led to a decrease in COX-2 activity (entries 1–2). Compound 14
Radiolabeling studies
showed an interesting IC50 around 1 mM whereas its isomer analog From our in vitro experiments, the sulfonylamide derivative 16
15 was active with an IC50 up to 2 mM. Despite our synthetic was found to be the most selective COX-2 inhibitor in the de novo
effort, the azaindazole derivatives 41 and 42 were 4-fold less effi- built (aza)indazole library. This activity, which is the direct reflec-
cient than their aromatic analog 14 (entries 3–4). The position of tion of the specific biological target affinity, indicated that it could
be a potent 18 F probe candidate. Three different labeling strat-
nitrogen is related to the efficiency as only 4- or 5- azaindazoles
egies were envisaged to obtain the desired radiotracer (Scheme
3). As the easiest way of precursor preparation, we first tried start-
40 and 41 versus 39 exhibited an IC50 value against COX-2 at an
acceptable level.
In the indazole series, the next modulations concerned the C-3 ing from the nitro derivative 42, which was prepared from 13
aromatic moiety. Its bi-substitution by methyl and fluorine (in C-4 after a Suzuki-Miyaura cross-coupling reaction in a 53% yield.

4
J. ELIE ET AL.
Table 2. Cox-2 inhibition.
Entry
1
Product (Yield %)[b]
IC50 COX-2 (mM)[a] Entry
Product (Yield %)[b]
IC50 COX-2 (mM)[a]
>25
14 (95)
15 (47)
1.459 0.5
11
23 (24)
24 (85)
2
2.296 0.5
12
2.977 0.4
3
4
39 (63)
40 (75)
>25
13
14
25 (46)
26 (94)
>25
4.421 0.6
4.461 0.4
5
6
7
8
41 (79)
18 (95)
19 (98)
20 (72)
4.750 0.6
1.028 0.2
>25
15
16
17
18
27 (85)
16 (73)
17 (59)
DuP-697
>25
0.409 0.3
4.200 0.4
0.039 0.024
1. 923 0.4
9
21 (91)
22 (85)
1.100 0.3
19
Celecoxib
0.076 0.054
10
>25
[a]Values are means of triplicates excepted for DuP-697 (n ¼ 27) and Celecoxib (n ¼ 5) which serves as reference in the assays
[b]Yields are thus obtained at the final step reported in Scheme 1.
Under thermal conditions direct nucleophilic substitution always was next transformed into tri n-butyl stannyl analog 45 under pal-
failed. In fact, traces of [18F]16 were only detected under micro- ladium catalysis in a moderate yield. Finally, Koser reagent
wave irradiation (2–3%). Then we switched to the use of a more smoothly afforded the attempted iodonium salt 46 in a satisfying
complex iodonium salt derivative 46, which was obtained after 4 manner. Based on the literature19,20, the radiolabeling reaction
steps from 13. First a Suzuki-Miyaura cross-coupling reaction led was performed in DMF using Cu(OTf)2 as catalyst under micro-
to 43 in a 69% yield. Electrophilic iodination with ICl afforded18, in wave irradiation at 100 ꢁC. In warm conditions the incorporation
a near quantitative yield, the 4-iodo phenyl derivative 44, which of 18 F, up to 60%, occurred but surprisingly in hot conditions

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
5
with automation, the pure formulated derivative [18F]16 was excitotoxicity mediated via NMDA receptor activation. This animal
model is well known in the literature to produce an inflammatory
reaction induced by COX-2 activation in lesioned cerebral
obtained with a very poor yield of 9%. So, a third strategy requir-
ing the use of boronic ester as leaving group was conducted.
Starting from 13, the mono Suzuki reaction led to boronic ester
46 in a 52% yield.
parenchyma23–25
.
The cerebral biodistribution of [18F]16 in rats lesioned with QA
was expressed as percent of injected dose per gram of tissue
(%ID/g) and is shown in Figure 3. For the control group, the cere-
bral accumulation of [18F]16 in non-activated brain regions one
hour after i.v. injection was low (0.046 0.004 to 0.078 0.030%ID/
g tissue). Furthermore, in activated brain areas, similar accumula-
tions were found whatever the cerebral regions (cerebellum, stri-
atum, cortex, hippocampus), with no statistically significant
difference compared to the contralateral striatum and cortex, the
brain areas that are the most affected by the neuroinflamma-
tory process.
Based on recent work21,22, the nucleophilic substitution was
achieved using an eluent solution containing the 18 F from QMA
and KOTf and K2CO3 in CH3CN/water and pyridine. This solution
was added to a solution which contained precursor 47, DMF as
solvent and Cu(OTf)2 as a catalyst. Using this method in warm
conditions, the incorporation of 18 F gave yields of up to 92%.
With automation for a 20 min reaction at 130 ꢁC, a decrease in effi-
ciency was observed but unlike the use of iodonium salt, the yield
remained satisfying (40%, decay corrected yield). Using the latter
method, with a view to preclinical evaluation, fully automated pro-
duction was achieved in 75 min with a molar activity of 90 GBq/
mmole on average.
These preliminary results suggest a low passage of [18F]16
through the blood-brain-barrier (BBB), even if the hypothesis of a
fast cerebral kinetics of the tracer, characterised by a rapid periph-
eral release, could be investigated using in vivo micro-PET.
Blocking in vivo experiments, consisting of a pretreatment with
either celecoxib or cold 16 (3 mg/kg) 15 min before injection of
the radiotracer, are also presented in Figure 3. In none of these
experiments was a decrease in [18F]16 accumulation observed.
This confirmed that the low in vivo accumulation observed con-
sisted of nonspecific binding. Finally, we observed a slight [18F]16
bone uptake in the femur (0.154 0.032%ID/g tissue), suggesting
negligible defluorination in vivo.
In vivo studies
The potential of the radiolabelled derivative of compound 16,
[18F]16, to image COX-2 by PET imaging was evaluated in a
rodent model of neuroinflammation. This model consists of a uni-
lateral intrastriatal injection of quinolinic acid to induce
The low brain uptake of [18F]16 might be related to its rather
low binding kinetic constant to the enzyme combined to a poor
brain uptake. Recent attempts to develop an adequate COX-2 PET
tracer faced the same unresolved issue at the pre-clinical evalu-
ation step26,27. Therefore, whereas numerous COX-2 inhibitors
have been proposed as therapeutic applications to modulate
inflammation, there is still to date a lack of effective radiopharma-
ceutical drug to explore COX-2 by in vivo brain molecular imaging
Scheme 2. Synthesis of fluorinated azaindazoles. Reagents and conditions: 4-
Iodofluorobenzene(1.1 equiv), Ag2CO3 (1.0 equiv), Pd(dppf)Cl2.CH2Cl2 (0.05 equiv),
PPh3 (0.10 equiv), H2O, 50 ꢁC, 18 h, 39: 4-aza 63%; 40: 6-aza 75%; 41: 7-aza 79%.
Scheme 3. Methods for radiolabeling. Rþeagents and conditions: (a) corresponding boronic acid (1.2 equiv), Cs2CO3 (3.0 equiv), Pd(PPh3)4 (0.1 equiv), dioxane, 150 ꢁC,
lW, 1 h, 42 53%, 43 69%; (b) [K/K222
]
18F–, DMF, 130 ꢁC, lW, 20 min; (c) ICl (2.0 equiv), CH2Cl2, 0 ꢁC then t.þa., 1.5h, 99%; (d) n-Bu6Sn2 (3.3 equiv), Pd(PPh3)4
(0.1 equiv), dioxane, 90 ꢁC, 2 h, 33%; (e) Moser reagent (1.5 equiv), CH3CN/CH2Cl2 (1/1), r.t., 18 h, 73%; (f) [K/K222
]
18F–, DMF, Cu(OTf)2 cat, 100 ꢁC, lW, 10 min; (g)
1,4-bis(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzene (3.0 equiv), Na2CO3 (6.5 equiv.), Pd(PPh3)4 (0.02 equiv), dioxane, 100ꢁC, lW, 40 min, 52%; (h) KOTf; K2CO3,
18F]KF, Cu(OTf)2 cat., DMF, pyridine, 130ꢁC, 20 min.
[

6
J. ELIE ET AL.
Figure 3. Cerebral biodistribution of [18F]16, 60 min after i.v. injection in rats, 7 days after focal intrastriatal injection of 150 nmol of quinolinic acid. In the “celecoxib”
and “cold 16” groups, animals received an i.v. pre-injection (15min) of the corresponding compound (3mg/kg). Results are expressed as % injected dose/g
tissue SEM. Cx: cortex; Stri: striatum; Hip: hippocampus, Cb: cerebellum.
but other applications could be envisioned in cancer
References
models11,12,28
.
1. Medzhitov R. Origin and physiological roles of inflammation.
Nature 2008;454:428–35.
2. Jacobs AH, Tavitian B. Noninvasive molecular imaging of
neuroinflammation. J Cereb Blood Flow Metab 2012;32:
1393–415.
3. Czeh M, Gressens P, Kaindl AM. The yin and yang of micro-
glia. Dev Neurosci 2011;33:199–209.
4. Perry V, Nicoll J, Holmes C. Microglia in neurodegenerative
disease. Nat Rev Neurol 2010;6:193–201.
Conclusion
The aim of the present study was to develop a series of novel
derivatives exhibiting high affinity and selectivity towards the
COX-2 enzyme. A straightforward synthesis was achieved, involv-
ing a bromination/arylation sequence under microwave irradiation
in the indazole series, whereas in the azaindazole series direct
C–H activation was achieved. The newly designed library was eval-
uated by an enzymatic cyclooxygenase inhibition assay in vitro to
assess the inhibitory potencies of COX-2 and COX-1 peroxydases
and SAR was discussed. The most promising compound 16 was
then radiolabeled with fluorine 18 using a very interesting boronic
ester displacement and the [18F]16 evaluated as a tracer in vivo in
a rat model of neuroinflammation. Although showing a submicro-
molar affinity for COX-2 and an excellent selectivity versus COX-1,
cerebral biodistribution studies using [18F]16 revealed a low brain
penetrance combined with poor accumulation specificity associ-
ated with neuroinflammation. Because of its good in vitro charac-
teristics; this radioligand deserves to be evaluated in peripheral
inflammation and cancer animal models. In parallel, efforts are
currently in progress regarding other original series to work
around these limitations and improve the brain availability of new
PET candidates. This work gives promising first results as a
molecular scaffold able to inhibit COX-2 activity at the sub-micro-
molar level. However, more work is needed to explore this scaf-
fold to increase the enzyme activity as well as to enhance its
brain penetrance.
5. Prata J, Santos S, Almeida M, et al. Bridging autism spectrum
disorders and schizophrenia through inflammation and
biomarkers
-
pre-clinical and clinical investigations.
J Neuroinflammation 2017;14:179.
6. Singhal G, Baune B. Microglia: an interface between the loss
of neuroplasticity and depression. Front Cell Neurosci 2017;
11:270.
7. Tronel C, Largeau B, Santiago Ribeiro M, et al. Molecular
targets for PET imaging of activated microglia: the cur-
rent situation and future expectations. Int
2017;18:802.
J Mol Sci
8. Vries E. Imaging of cyclooxygenase-2 (Cox-2) expression:
potential use in diagnosis and drug evaluation. Curr Pharm
Des 2006;12:3847–56.
9. Laube M, Kniess T, Pietzsch J. Radiolabeled Cox-2 inhibitors
for non-invasive visualization of Cox-2 expression and activ-
ity-a critical update. Molecules 2013;18:6311.
10. Jackson L, Hawkey C. Cox-2 selective nonsteroidal anti-
inflammatory drugs: do they really offer any advantages?.
Drugs 2000;59:1207–16.
11. Tietz O, Wuest M, Marshall A, et al. Pet imaging of cyclooxy-
genase-2 (Cox-2) in a pre-clinical colorectal cancer model.
EJNMMI Res 2016;6:37.
Disclosure statement
No potential conflict of interest was reported by the authors.
12. Uddin M, Crews B, Ghebreselasie K, et al. Fluorinated Cox-2
inhibitors as agents in PET imaging of inflammation and
cancer. Cancer Prev Res (Phila) 2011;4:1536–45.
Funding
13. Khatun N, Gogoi A, Basu P, et al. CuO nanoparticle catalysed
synthesis of 2H-indazoles under ligand free conditions. RSC
Adv 2014;4:4080–4.
This research was supported by grants from the Labex IRON [ANR-
11-LABX-0018–01] and InMIND program which are associated to
this work.

JOURNAL OF ENZYME INHIBITION AND MEDICINAL CHEMISTRY
7
14. Kumar M, Park A, Park N, Lee S. Consecutive condensation, 22. Ye Y, Schimler S, Hanley P, Sanford M. Cu(OTf)2-mediated
C-N and N-N bond formations: a copper- catalyzed one-pot
three-component synthesis of 2H-indazole. Org Lett 2011;13:
3542–5.
fluorination of aryltrifluoroborates with potassium fluoride.
J Am Chem Soc 2013;135:16292–5.
23. Kalonia H, Kumar A. Suppressing inflammatory cascade by
cyclo-oxygenase inhibitors attenuates quinolinic acid
induced huntington’s disease-like alterations in rats. Life Sci
2011;88:784–91.
15. Copin C, Henry N, Buron F, Routier S. Synthesis of 2,6-disub-
stituted imidazo[2,1-b][1,3,4]thiadiazoles through cyclization
and Suzuki–Miyaura cross-coupling reactions. Eur J Org
Chem 2012;2012:3079–83.
16. Meur M, Bourg S, Massip S, et al. First palladium-catalyzed
direct regioselective C-5 (het)arylation of mono- or disubsti-
tuted thiazolo[3,2-b][1,2,4]triazoles. Eur J Org Chem 2014;
2014:3704–14.
17. El Kazzouli S, Guillaumet G. Functionalization of indazoles by
means of transition metal-catalyzed cross-coupling reactions.
Tetrahedron 2016;72:6711–27.
18. Langhals H, Knochel P, Walter A, Zimdars S.
Benzothiadiazoloperylenes and benzoxadiazoloperylenes:
amorphous functional materials. Synthesis 2012;44:3465–76.
24. Shukuri M, Takashima-Hirano M, Tokuda K, et al. In vivo expres-
sion of cyclooxygenase-1 in activated microglia and macro-
phages during neuroinflammation visualized by PET with 11C-
ketoprofen methyl ester. J Nucl Med 2011;52:1094–101.
25. Salzberg-Brenhouse H, Chen E, Emerich D, Baldwin S, et al.
Inhibitors of cyclooxygenase-2, but not cyclooxygenase-1
provide structural and functional protection against quino-
linic acid-induced neurodegeneration. J Pharmacol Exp Ther
2003;306:218–28.
26. Kaur J, Tietz O, Bhardwaj A, et al. Design, synthesis, and
evaluation of an 18F-labeled radiotracer based on celecox-
ib–nbd for positron emission tomography (PET) imaging of
cyclooxygenase-2 (Cox-2). ChemMedChem 2015;10:1635–40.
27. Kniess T, Laube M, Bergmann R, et al. Radiosynthesis of a 18F-
labeled 2,3-diarylsubstituted indole via Mcmurry coupling for
functional characterization of cyclooxygenase-2 (Cox-2) in
vitro and in vivo. Biorg Med Chem 2012;20:3410–21.
28. Pacelli A, Greenman J, Cawthorne C, Smith G. Imaging Cox-2
expression in cancer using PET/SPECT radioligands: current
status and future directions. J Labelled Compd Radiopharm
2014;57:317–22.
ꢁ
19. Neumann K, Qin L, Vavere A, et al. Efficient automated syn-
theses of high specific activity 6-[18F]fluorodopamine using
a
diaryliodonium salt precursor.
J
Labelled Compd
Radiopharm 2016;59:30–4.
20. Ichiishi N, Canty A, Yates B, Sanford M. Mechanistic investi-
gations of Cu-catalyzed fluorination of diaryliodonium salts:
elaborating the cuI/cuIII manifold in copper catalysis.
Organometallics 2014;33:5525–34.
21. Mossine A, Brooks A, Makaravage K, et al. Synthesis of
[18F]arenes via the copper-mediated [18F]fluorination of bor-
onic acids. Org Lett 2015;17:5780–3.