11C-labeled analogs of indomethacin esters and amides for brain cyclooxygenase-2 imaging: Radiosynthesis, in vitro evaluation and in vivo characteristics in mice

Article abstract of DOI:10.1248/cpb.59.938

There is great potential in the use of positron emission tomography (PET) and suitable radiotracers for the study of cyclooxygenase type 2 (COX-2) enzyme in living subjects. In the present study, we prepared and evaluated five ¹¹C-labeled ester and amide analogs derived from indomethacin as potential PET imaging agents for the in vivo visualization of the brain COX-2 enzyme. Five ¹¹C-labeled COX-2 inhibitors, with different lipophilicities and moderate COX-2 inhibitory activity, were prepared by treatment of the corresponding O-desmethyl precursors with [¹¹C] methyl triflate and purified by HPLC (radiochemical yields of 55-71%, radiochemical purity of >93%, and the specific activities of 22-331 GBq/μmmol). In mice, radioactivity in the brain for all radiotracers was low, with very low brain-to-blood ratios. A clear inverse relationship was observed between brain uptake at 1 min postinjection and the lipophilicity (experimental log P_7.4) of the studied ¹¹C-radiotracers. Pretreatment of mice with cyclosporine A to block P-glycoproteins caused a significant increase in brain uptake of radioactivity following injection of the ¹¹C-radiotracer compared to control. HPLC analysis showed that each radiotracer was rapidly metabolized, and a few metabolites, which were more polar than the original radiotracers, were found in both plasma and brain. No specific binding of the tracers towards the COX-2 enzyme in the brain was clearly revealed by in vivo blocking study. Further structural refinement of the tracer agent is necessary for better enhancement of brain uptake and for sufficient metabolic stability.

Full text of DOI:10.1248/cpb.59.938

938
Regular Article
Chem. Pharm. Bull. 59(8) 938—946 (2011)
Vol. 59, No. 8
11C-Labeled Analogs of Indomethacin Esters and Amides for Brain
Cyclooxygenase-2 Imaging: Radiosynthesis, in Vitro Evaluation and
in Vivo Characteristics in Mice
Yumi YAMAMOTO,a Jun TOYOHARA,b Kiichi ISHIWATA,b Kohei SANO,a Fumihiko YAMAMOTO,a
,a,c
,a
*
*
Takahiro MUKAI,
and Minoru MAEDA
a Graduate School of Pharmaceutical Sciences, Kyushu University; 3–1–1 Maidashi, Higashi-ku, Fukuoka 812–8582,
Japan: b Positron Medical Center, Tokyo Metropolitan Institute of Gerontology; 1–1 Naka-cho, Itabashi-ku, Tokyo
173–0022, Japan: and c Kobe Pharmaceutical University; 4–19–1 Motoyamakita-machi, Higashinada-ku, Kobe 658–8558,
Japan. Received January 5, 2011; accepted May 13, 2011; published online May 19, 2011
There is great potential in the use of positron emission tomography (PET) and suitable radiotracers for the
study of cyclooxygenase type 2 (COX-2) enzyme in living subjects. In the present study, we prepared and evalu-
ated five 11C-labeled ester and amide analogs derived from indomethacin as potential PET imaging agents for the
in vivo visualization of the brain COX-2 enzyme. Five 11C-labeled COX-2 inhibitors, with different lipophilicities
and moderate COX-2 inhibitory activity, were prepared by treatment of the corresponding O-desmethyl precur-
sors with [11C]methyl triflate and purified by HPLC (radiochemical yields of 55—71%, radiochemical purity of
ꢀ93%, and the specific activities of 22—331 GBq/mmol). In mice, radioactivity in the brain for all radiotracers
was low, with very low brain-to-blood ratios. A clear inverse relationship was observed between brain uptake at
1 min postinjection and the lipophilicity (experimental log P7.4) of the studied 11C-radiotracers. Pretreatment of
mice with cyclosporine A to block P-glycoproteins caused a significant increase in brain uptake of radioactivity
following injection of the 11C-radiotracer compared to control. HPLC analysis showed that each radiotracer was
rapidly metabolized, and a few metabolites, which were more polar than the original radiotracers, were found in
both plasma and brain. No specific binding of the tracers towards the COX-2 enzyme in the brain was clearly
revealed by in vivo blocking study. Further structural refinement of the tracer agent is necessary for better
enhancement of brain uptake and for sufficient metabolic stability.
Key words cyclooxygenase type 2; indomethacin; carbon-11; brain; biodistribution; mouse
Cyclooxygenase (COX) is the crucial enzyme in the con- better design of COX-2 imaging probes.
version of arachidonic acid to prostaglandins. It exists at
Currently, many selective COX-2 inhibitors with various
least in two isoforms, a constitutive form (COX-1) and an in- structural features, as exemplified by vicinal diaryl heterocy-
ducible form (COX-2),1—3) although a third distinct COX cles such as celecoxib and refecoxib, have been developed as
isozyme has also been reported.4) The COX-1 enzyme is re- non-steroidal anti-inflammatory drugs (NSAIDs) with im-
sponsible for maintaining homeostasis whereas COX-2 is in- proved gastric safety profiles.19) Most of them have high
duced during inflammation by various stimuli and, in some lipophilicity. Indomethacin is a traditional non-selective in-
tissues including the brain, kidney and placenta, is also con- hibitor of the COX isozyme which is widely used as a
stitutively expressed. Recent studies indicate that elevated NSAID, but with poor brain penetration largely due to the
expression of COX-2 has been implicated in many pathologi- presence of the carboxylic acid moiety.20) Interestingly, re-
cal events, including rheumatoid arthritis, cancer, heart dis- cent studies concerning ester and amide derivatives of in-
ease, and neurodegenerative disorders.5)
domethacin have found very high COX-2 inhibition potency
In the brain, COX-2 is mainly expressed in the cortex, hy- and a high degree of COX-2-selectivity.21—23) Marnett and
pothalamus, and hippocampus,6) and is upregulated in neuro- colleagues synthesized 123I-labeled N-iodobenzyl-containing
logical disorders such as Parkinson’s disease and Alzheimer’s amide derivative of indomethacin as a COX-2 imaging agent,
disease,7—9) although the functions of COX-2 in pathophysio- which exhibited sufficient stability in COX-2 expressing
logical processes are not yet well-understood. There is great nude mouse tumor.24) More recently, using organic fluoro-
potential that COX-2-targeted imaging by positron emission phores tethered to indomethacin through an amide linkage,
tomography (PET) may provide useful information on the the feasibility of specific in vivo targeting of COX-2 in in-
role of this enzyme, especially in various neurological disor- flammatory lesions in mice has been demonstrated by the
ders, and also may help to evaluate the efficacy of selective same group.25)
COX-2 inhibitors.10) Thus, many COX-2 inhibitors have been
The purpose of this study is to find suitable radioligands
labeled with a suitable radionuclide for mapping the enzymes for imaging brain COX-2 in vivo by PET. We initially chose
in vivo. However, previously evaluated PET radiotracers, in- two potent and selective COX-2 inhibitors of indomethacin
cluding 11C-labeled COX inhibitors of the diarylheterocyclic ester analog for 11C-radiolabeling: N-Pentyl-(1-p-chloro-
class which were synthesized by us,11,12) have only achieved benzoyl-5-methoxy-2-methylindole)-3-acetate (1) (COX-1,
limited success, due to a lack of sufficient specific-binding to IC50ꢀ66 mM; COX-2, IC50ꢁ50 nM)21) and N-octyl-(1-p-
the COX-2 enzyme and/or sufficient brain penetration, in chlorobenzoyl-5-methoxy-2-methylindole)-3-acetate (2)
spite of promising in vitro pharmacological data.13—18) There (COX-1, IC50ꢁ66 mM; COX-2, IC50ꢁ40 nM).21) Furthermore,
is a need for systematic studies correlating the physicochem- on the basis of the reported structure–activity studies concern-
ical properties of agents with their in vivo behavior for a ing the ester and amide derivatives of indomethacin 21—23)
we
,
∗ To whom correspondence should be addressed. e-mail: m-maeda@daiichi-cps.ac.jp; mukai@phar.kyushu-u.ac.jp © 2011 Pharmaceutical Society of Japan

August 2011
939
Chart 1. Synthesis Scheme for Indomethacin Esters and Amides (3—6), and Labeling Precursors (7—10, 12, 13)
next designed and synthesized one additional amide deriva- 13, with an alkoxy moiety in the side chain were obtained in
tive, two amide compounds with a –OCH2– linkage, and one a two-step procedure, in which demethylation of the methoxy
carbamate ester derivative of indomethacin, in order to exam- group in indomethacin was achieved by treating it with BBr3,
ine the effects of different functional groups on COX-2 in- and subsequent condensation of the carboxyl group with 2-
hibitory potency and lipophilicity. As a consequence, five ethoxyethylamine or 3-methoxypropylamine gave the desired
11C-labeled COX-inhibitors were prepared via demethyl pre- compounds (Chart 1). All of the synthetic compounds gave
cursors, and examined for in vivo characteristics in mice.
satisfactory spectroscopic data, which were in full accor-
dance with their depicted structures.
Results and Discussion
COX Inhibitory Activity The inhibitory activity of the
Synthesis Compounds 1 and 2 were prepared as previ- compounds (2—6) for the COX enzyme and selectivity index
ously described.21) The target compounds (3, 5, 6) for HPLC (SI) was determined in vitro, including celecoxib and in-
standards and in vitro studies were obtained in satisfactory domethacin as a reference COX inhibitor, using a colorimet-
yields by treatment of indomethacin with the appropriate ric COX (ovine) inhibitor screening assay,26) in which arachi-
amine or alcohol according to a simlar literature proce- donic acid was used as the substrate and N,N,Nꢂ,Nꢂ-tetra-
dure,21) as illustrated in Chart 1. Additionally, the carbamate methylphenylenediamine (TMPD) as the cosubstrate. The
butyl ester (4) of indomethacin was obtained by Curtius re- IC50 value was calculated from the concentration–inhibition
arrangement of the indomethacin azide, followed by treat- response curve. Additionally, the ratio of IC50 value of in-
ment of intermediate isocyanate with 1-butanol. The pheno- domethacin as a reference compound-to-IC50 values of the
lic precursors (7—10) for 11C-radiolabeling were prepared in test compounds was calculated to know the relative COX-2
reasonable yields by demethylation of the corresponding in- inhibitory activity. In vitro inhibition studies showed that
domethacin derivatives (1—4) using BBr3 as a dealkylating compound 3 had the highest affinity and selectivity for COX-
agent. On the other hand, two phenolic compounds, 12 and 2, followed by compound 2 (Table 1). These two compounds

940
Vol. 59, No. 8
have higher activity than celecoxib and indomethacin. The [11C]-O-methylation of precursor 7, the use of an excess
introduction of a –OCH2 linkage instead of the methylene amount of NaOH led to a more polar radioactive byproduct,
moiety in the side chain of indomethacin amide 3 signifi- consistent with the identity of 11C-labeled indomethacin. Ef-
cantly decreased the COX-2 inhibition potency with reduced fective methylation of precursor 7 proceeded by the use of an
selectivity (compounds 5, 6), which indicates that ether-oxy- equimolar amount of NaOH.11) For [11C]-O-methylations of
gen may be detrimental to COX inhibition. Furthermore, the other four precursors, the use of a larger amount of
compound 4 showed a significant loss of inhibition activity NaOH successfully gave the required products. The absence
against both COX isoforms.
of any residual of starting demethyl precursors in the purified
Radiolabeling Five 11C-labeled indomethacin analogs product was verified by HPLC analysis.
[11C]1—3, 5, 6 were prepared by [11C]-O-methylation of the
Lipophilicity The partition coefficient of five 11C-la-
corresponding demethyl precursors in acetone with beled compounds between 1-octanol and phosphate buffered
[11C]methyl triflate in the presence of NaOH as the base at saline was measured.27) The log P7.4 values were 1.98—3.94
room temperature, followed by purification with HPLC, as (Table 2). Lipophilicity increased with increasing length of
shown in Chart 2. Table 2 summarizes the results of the the side chain as seen in [11C]2 and [11C]3, and replacement
radiosyntheses. The presence of NaOH was essential for an of the methylene of the side chain with oxygen as in [11C]5
efficient procedure for the desired methylation; however, for and [11C]6 decreased log P7.4 values due to its increased po-
larity. The measured values are somewhat in the outside
range of those generally considered optimal for good blood
brain barrier (BBB) penetration of drugs, although the well-
known parabolic relationship between the lipophilicity of a
molecule and its uptake by the brain is not universally appli-
cable.28,29) It should be noted that [11C]3 features a relatively
low lipophilicity (log P7.4ꢁ2.40) while retaining high COX-2
inhibitory potency and high selectivity, as shown in Table 1.
In this study there was no correlation between the measured
IC50 and log P7.4 values for the compounds.
Biodistribution To explore the in vivo characteristics of
the five 11C-labeled compounds, biodistribution studies were
performed in normal male ddY mice. The radioactivity con-
centrations in various tissues as a function of time following
intravenous injection of the radiotracers are represented in
Table 3. Four 11C-radiotracers except for [11C]1 exhibited a
relatively fast clearance of radioactivity in the blood within
60 min. The half-lives of the blood clearance were 12.3 min,
Table 1. COX Inhibitory Activity of Six Indomethacin Esters and Amides,
and Two Reference Compounds
IC50 (mM)a)
Relative
COX-2
Compound
Selectivityc)
inhibitory
COX-1
COX-2
potencyb)
d)
d)
1
2
3
4
5
6
—
—
—
56
303
ꢃ0.1
5
6
21
1
—
ꢀ476
ꢀ2564
—
ꢀ42
ꢀ47
ꢀ100
ꢀ100
ꢀ100
ꢀ100
ꢀ100
ꢀ100
0.79
0.21
0.039
ꢀ100
2.35
2.10
0.56
Celecoxib
Indomethacin
ꢀ178
0.067
11.8
a) Values represent the average of three independent determinations by colorimetric
inhibition screening assay. b) Ratio of IC50 (indomethacin)/IC50 (test compound). c)
Ratio of IC50 (COX-1)/IC50 (COX-2). d) IC50 of compound 1 could not be determin-
ed due to the lack of its chemical stability in the assay procedure.
Chart 2. Radiosynthesis of 11C-Labeled Indomethacin Esters and Amides ([11C]1, [11C]2, [11C]3, [11C]5, [11C]6)
Table 2. Summary of Radiosynthesis and Lipophilicity of 11C-Labeled Indomethacin Esters and Amides
Radiochemical yielda)
%
Radiochemical purity
%
Specific activity
GBq/mmol
b)
Compound
Log P7.4
[11C]1
[11C]2
[11C]3
[11C]5
[11C]6
65ꢄ8 (nꢁ4)
55ꢄ12 (nꢁ5)
61ꢄ7 (nꢁ3)
71ꢄ5 (nꢁ4)
67ꢄ12 (nꢁ4)
ꢀ97 (nꢁ4)
ꢀ94 (nꢁ4)
ꢀ93 (nꢁ7)
ꢀ96 (nꢁ4)
ꢀ96 (nꢁ4)
22—61 (nꢁ5)
29—97 (nꢁ6)
38—135 (nꢁ7)
166—295 (nꢁ4)
149—331 (nꢁ4)
3.94
3.78
2.40
2.12
1.98
a) Preparation time for 5 compounds was 21—27 min after the end of [11C]CO2 production. Decay-corrected yield based on [11C]methyl triflate, averageꢄS.D. (run numbers).
b) The octanol/phosphate buffered saline (pHꢁ7.4) partition coefficient (log P7.4) using a conventional flask-shake technique, average of three independent assays.

August 2011
941
Table 3a. Tissue Distribution of Radioactivity after Intravenous Injection Table 3e. Tissue Distribution of Radioactivity after Intravenous Injection
of [11C]1 into Mice
of [11C]6 into Mice
Uptake (%ID/g)a)
5 min 30 min
Uptake (%ID/g)a)
5 min 30 min
Tissue
Tissue
1 min
60 min
1 min
60 min
Blood
Brain
Heart
Lung
Liver
Kidney
Small intestine
10.73ꢄ0.94 8.80ꢄ1.02 6.36ꢄ0.37 4.95ꢄ0.29
0.47ꢄ0.09 0.45ꢄ0.06 0.34ꢄ0.06 0.23ꢄ0.02
6.81ꢄ1.18 5.11ꢄ0.75 2.69ꢄ0.37 2.17ꢄ0.23
10.06ꢄ1.54 4.38ꢄ0.15 3.63ꢄ0.39 3.49ꢄ0.11
12.43ꢄ1.75 6.54ꢄ1.38 6.22ꢄ0.43 5.53ꢄ0.33
7.57ꢄ0.74 8.25ꢄ0.83 7.88ꢄ0.46 6.94ꢄ1.13
2.20ꢄ0.36 3.56ꢄ0.80 3.73ꢄ0.74 3.15ꢄ0.47
Blood
Brain
Heart
Lung
Liver
Kidney
Small intestine
3.01ꢄ0.28
1.84ꢄ0.16
6.09ꢄ0.25
5.38ꢄ0.40
1.92ꢄ0.20 1.05ꢄ0.13 0.84ꢄ0.08
1.12ꢄ0.11 0.65ꢄ0.17 0.41ꢄ0.07
2.73ꢄ0.23 0.92ꢄ0.08 0.57ꢄ0.04
2.91ꢄ0.29 1.07ꢄ0.14 0.73ꢄ0.02
11.68ꢄ2.35 14.90ꢄ2.45 13.95ꢄ1.24 9.15ꢄ0.69
10.28ꢄ1.39
5.12ꢄ1.10
4.67ꢄ0.62 2.59ꢄ0.27 1.82ꢄ0.06
8.39ꢄ2.47 8.32ꢄ1.78 7.26ꢄ2.14
a) MeanꢄS.D. (nꢁ4).
a) MeanꢄS.D. (nꢁ4).
Table 3b. Tissue Distribution of Radioactivity after Intravenous Injection
of [11C]2 into Mice
Uptake (%ID/g)a)
5 min 30 min
Tissue
1 min
60 min
Blood
Brain
Heart
Lung
3.75ꢄ0.50
0.18ꢄ0.08
5.65ꢄ0.57
14.48ꢄ3.00
1.15ꢄ0.11 0.83ꢄ0.11 0.79ꢄ0.13
0.19ꢄ0.06 0.26ꢄ0.07 0.25ꢄ0.05
3.95ꢄ0.75 0.74ꢄ0.05 0.59ꢄ0.08
6.60ꢄ0.75 1.86ꢄ0.96 1.43ꢄ0.73
Liver
Kidney
Small intestine
16.11ꢄ0.62 14.58ꢄ2.44 8.07ꢄ0.88 5.59ꢄ0.22
5.79ꢄ0.40 5.13ꢄ0.66 2.16ꢄ0.42 1.84ꢄ0.25
1.71ꢄ0.30 12.01ꢄ0.68 6.90ꢄ0.85 5.36ꢄ2.07
a) MeanꢄS.D. (nꢁ4).
Fig. 1. Relationship between Lipophilicity (log P7.4) and Mouse Brain Up-
Table 3c. Tissue Distribution of Radioactivity after Intravenous Injection
take (%ID/g) at 1 min after Injection of the 11C-COX Inhibitors
of [11C]3 into Mice
Uptake (%ID/g)a)
5 min 30 min
Tissue
tration of radioactivity observed in the small intestine may
reflect hepatobiliary clearance of the parent tracers and/or
their radioactive metabolites. [11C]1 displayed very high lev-
els of radioactivity in the blood and most organs such as the
liver, lungs and kidneys in contrast to the other four radio-
tracers.
1 min
60 min
Blood
Brain
Heart
Lung
3.50ꢄ0.18
1.06ꢄ0.09
7.05ꢄ0.36
6.26ꢄ0.53
1.80ꢄ0.23 0.73ꢄ0.06 0.72ꢄ0.13
0.98ꢄ0.09 0.51ꢄ0.09 0.48ꢄ0.13
2.36ꢄ0.18 0.80ꢄ0.15 0.59ꢄ0.11
4.18ꢄ1.82 2.79ꢄ2.44 0.76ꢄ0.13
Liver
12.98ꢄ1.52 10.03ꢄ1.70 5.40ꢄ0.85 4.87ꢄ0.98
Whole brain uptake for all tracers was low, with values
ranging from 0.18% injected dose (ID)/g for [11C]2 to 1.84%
ID/g for [11C]6 at 1 min after administration which, with ex-
ception of [11C]2, then decreased slowly over 60 min. In the
case of [11C]2 which showed the lowest brain uptake, a slight
increase of radioactivity in the brain was observed at later
time points. The brain-to-blood ratios of all tracers were con-
sequently much less than 1.0 unity (0.04—0.77) throughout
the observation period. These results clearly demonstrate that
all tracers investigated have low BBB permeability. Notably,
the lipophilicity of the radiotracers inversely correlated well
(r2ꢁ0.90, pꢁ0.01) with brain uptake at 1 min post injection
(Fig. 1), which is most likely to be reflective of the initial ex-
traction of the radiotracers from the blood. This relationship
may be due to binding to plasma proteins which is expected
to increase with larger log P7.4 values of the indomethacin
analogs.30) These radiotracers likely bind to the plasma pro-
teins similar to other NSAIDs, including indomethacin bind-
ing with high affinity to serum albumin.31,32) Non-specific
Kidney
Small intestine
10.95ꢄ1.19
3.49ꢄ0.41
5.40ꢄ0.43 1.84ꢄ0.30 1.42ꢄ0.27
6.15ꢄ1.39 2.50ꢄ0.56 3.16ꢄ0.85
a) MeanꢄS.D. (nꢁ4).
Table 3d. Tissue Distribution of Radioactivity after Intravenous Injection
of [11C]5 into Mice
Uptake (%ID/g)a)
5 min 30 min
Tissue
1 min
60 min
Blood
Brain
Heart
Lung
Liver
Kidney
Small intestine
2.13ꢄ0.09
1.53ꢄ0.09
4.61ꢄ0.25
3.73ꢄ0.69
7.65ꢄ1.23
7.76ꢄ0.54
4.03ꢄ0.44
1.50ꢄ0.10 0.80ꢄ0.12 0.86ꢄ0.22
1.10ꢄ0.05 0.52ꢄ0.04 0.52ꢄ0.25
2.30ꢄ0.10 0.77ꢄ0.09 0.61ꢄ0.10
2.29ꢄ0.21 0.93ꢄ0.14 0.76ꢄ0.17
10.70ꢄ1.27 5.25ꢄ0.73 4.64ꢄ1.69
3.54ꢄ0.19 1.61ꢄ0.16 1.26ꢄ0.24
10.71ꢄ2.71 4.12ꢄ1.35 3.20ꢄ1.79
a) MeanꢄS.D. (nꢁ4).
1.3 min, 2.9 min, 4.2 min, and 3.8 min for [11C]1, [11C]2, binding to albumin and other plasma proteins is known to
[11C]3, [11C]5, and [11C]6, respectively. The highest initial correlate positively and linearly in vitro with increasing
concentrations of radioactivity were observed in the liver for lipophilicity. However, it remains controversial the extent to
all tracers, followed by a gradual increase and/or relatively which non-specific binding to plasma proteins inhibits BBB
slow clearance of radioactivity. In addition, the peak concen- penetration. This theory cannot explain why some radiotrac-

942
Vol. 59, No. 8
ers exhibiting high (ꢀ90%) plasma protein binding may methacin (polar 1) were found. The findings seemingly indi-
enter the brain nearly as well as those that exhibit much cated hydrolysis of the ester or amide moiety of the tracers
lower plasma protein binding.29) Certainly, some tracers in- by esterase in the blood and peripheral organs. Unidentified
cluding [11C]palmitic acid33) and [11C]arachidonic acid34,35) more polar metabolites (polar 3) were predominant for both
with calculated log Pꢀ7 have been used successfully in PET [11C]3 and [11C]5. Remmel et al. have shown for indo-
experiments although their maximal brain uptakes are lower methacin phenethylamide in rodents that the initial step in
than those of commonly used neuroreceptor radioligands. the metabolism is the oxidative pathway on its amide side
Therefore, lipophilicity is, in any case, an imperfect proxy chain, resulting in the formation of hydroxylated and O-
for BBB permeability. Currently, there is no general under- demethylated metabolites in addition to their glucuronides.37)
standing of the relationship between lipophilicity and the dis- If this route of metabolism is also the case for the radiotrac-
tribution characteristics for radiotracers of COX-2 inhibitors. ers studied here, then more polar metabolites than the parent
Unfortunately, high lipophilicity is inherent in currently compounds would be produced.
available COX inhibitors because of hydrophobic interac-
In the brain samples only 9—26% of the radioactivity was
detected as intact forms and the presence of two radiolabeled
tions with the COX binding sites in both isozymes.36)
Metabolism in Plasma and Brain The metabolic be- metabolites, [11C]indomethacin (polar 1) and an unidentified
havior of each radiotracer in vivo was studied in mice at fraction (polar 2), was found. A polar 3 metabolite observed
30 min after injection. As summarized in Table 4, each radio- in plasma extracts was absent in the brain samples. The me-
tracer was rapidly metabolized. HPLC analysis of plasma for tabolism probably occurred in the brain because plasma
[11C]1 showed no unchanged form and one polar peak (polar metabolites are expected to poorly penetrate the BBB. How-
1) of which the retention time corresponds to that of in- ever, this consideration requires further investigation. Thus,
domethacin. In the other four tracers, 14—32% of the ra- one of the reasons for the low accumulation of radioactivity
dioactivity was recovered as unchanged parent compounds, in the brain has been found in the marked metabolism of the
and a few radiolabeled metabolites including [11C]indo- studied 11C-labeled radiotracers, which occurred in the
plasma and brain tissue. However, a low initial uptake of
Table 4. 11C-Labeled Metabolites in Plasma and Brain at 30 min after In-
travenous Injection of 11C-Labeled Indomethacin Esters and Amides into
Mice
each radiotracer was not explained by the rapid metabolism.
Specific Binding To determine the specific binding of
radiotracers to COX-2 enzymes, blocking experiments were
carried out using several COX-2 selective inhibitors with
BBB permeability, celecoxib and N-[2-(cyclohexyloxy)-4-
nitrophenyl]methanesulfonamide (NS-398), a nonselective
COX inhibitor, indomethacin, and carrier-loading. None of
any COX inhibitors and each unlabeled compound caused
significant inhibition in the brain uptake and brain-to-blood
ratio of [11C]1, [11C]2, [11C]5 or [11C]6 as summarized in
Table 5 which involves the representative blocking results
with celecoxib and NS-398, suggesting the absence of COX-
2-specific and saturable binding. On the other hand, in the
case of [11C]3, a small but statistically significant inhibition
(ca. 15% reduction) was observed in the brain uptake of ra-
dioactivity with celecoxib or NS-398 (Table 5). This finding
seems to indicate a small fraction of displaceable binding in
the brain, but additional evidence is necessary. The rapid
degradation of all 11C-COX inhibitors and the presence of ra-
diometabolites in the brain as described above can also con-
tribute to the lack of specific binding for COX-2 enzyme in
vivo blocking.
Percentage of metabolitea)
Polar 1
(indomethacin)
Parent
Polar 2
Polar 3
[11C]1 Retention timeb)
8.0 min
ND
2.9 min
100
2.1 min
ND
Plasma
Brain
26ꢄ6
13.2 min 5.2 min
15ꢄ4
22ꢄ4
6.7 min 4.8 min
32ꢄ5
21ꢄ7
12.7 min 8.1 min
17ꢄ4
9ꢄ2
11.0 min 8.4 min
52ꢄ3
22ꢄ9
4.1 min
13ꢄ1
NDc)
3.3 min
14ꢄ5
ND
5.3 min
9ꢄ5
ND
[11C]2 Retention timeb)
Plasma
72ꢄ3
Brain
[11C]3 Retention timeb)
Plasma
NDc)
2.1 min
43ꢄ7
73ꢄ7
3.9 min
46ꢄ9
84ꢄ5
4.2 min
25ꢄ17
84ꢄ3
11ꢄ2
6ꢄ4
Brain
[11C]5 Retention timeb)
Plasma
28ꢄ8
7ꢄ4
Brain
[11C]6 Retention timeb)
Plasma
6.6 min
20ꢄ8
ND
14ꢄ5
9ꢄ2
41ꢄ6
7ꢄ1
Brain
a) MeanꢄS.D. (nꢁ3), NDꢁnot detected. b) Retention time on HPLC analysis.
c) Besides a parent peak, 78% of the radioactivity was eluted with the broadening of
peaks including polar 1 and polar 2 and metabolite profiles could not be accurately de-
termined.
Effect of Cyclosporine A The BBB expresses various
Table 5. Radioactivity of 11C-COX Inhibitors in the Brain and the Blood of Mice Co-injected with COX-2 Inhibitor, Celecoxib and NS-398a)
Brain uptake (%ID/g)b)
Celecoxib
Brain-to-blood ratiob)
Celecoxib
Control
NS-398
Control
NS-398
[11C]1
[11C]2
[11C]3
[11C]5
[11C]6
0.54ꢄ0.13
0.40ꢄ0.04
0.68ꢄ0.08
0.73ꢄ0.12
0.80ꢄ0.06
0.46ꢄ0.07
0.34ꢄ0.09
0.58ꢄ0.09*
0.72ꢄ0.09
0.76ꢄ0.04
0.42ꢄ0.07
0.48ꢄ0.13
0.71ꢄ0.09
0.79ꢄ0.08
0.73ꢄ0.09
0.08ꢄ0.01
0.46ꢄ0.02
0.86ꢄ0.20
0.66ꢄ0.08
0.62ꢄ0.03
0.08ꢄ0.01
0.41ꢄ0.08
0.67ꢄ0.06*
0.60ꢄ0.18
0.60ꢄ0.03
0.07ꢄ0.01
0.52ꢄ0.11
0.67ꢄ0.06*
0.88ꢄ0.27
0.51ꢄ0.12
a) Mice received the following blocker solution (1 mg/kg per animal, 0.1 ml), celecoxib, and NS-398 dissolved in DMSO, co-injected with each radiotracer. In the control
mice the same amount of DMSO was co-injected. The radioactivity in the tissue was measured at 30 min for [11C]2 and [11C]3, and at 15 min for [11C]5 and [11C]6 after each
11C-COX inhibitor injection. b) MeanꢄS.D. (nꢁ5). ∗ Significant decrease (pꢃ0.05) compared to the control (Student’s t-test).

August 2011
943
Table 6. Effect of P-Glycoprotein Modulation with Cyclosporine A on the Brain Uptake of Radioactivity after Intravenous Injection of 11C-COX Inhibitors
into Micea)
Brain uptake (%ID/g)b)
CysA
Brain-to-blood ratiob)
CysA
Control
CysA/Control
Control
CysA/Control
[11C]2
[11C]3
[11C]5
[11C]6
0.40ꢄ0.04
0.68ꢄ0.08
0.73ꢄ0.12
0.80ꢄ0.06
1.55ꢄ1.04*
0.99ꢄ0.09**
1.11
3.88
1.46
1.52
2.13
0.46ꢄ0.02
0.86ꢄ0.20
0.66ꢄ0.08
0.62ꢄ0.03
1.27ꢄ0.85*
0.90ꢄ0.05
0.92
2.76
1.05
1.39
1.94
1.71ꢄ0.25**
1.20ꢄ0.25**
a) Mice were pretreated with cyclosporine A (CysA) (50 mg/kg, i.v.) 30 min before injection of the tracers, and the radioactivity in the tissue was measured at 30 min for
11C]2 and [11C]3, and at 15 min for [11C]5 and [11C]6 after each 11C-COX inhibitor injection. b) MeanꢄS.D. (CysA; nꢁ4 for [11C]2 and [11C]6, nꢁ5 for [11C]3, nꢁ2 for [11C]5.
[
control; nꢁ5 for each group). ∗ pꢃ0.05 and ∗∗ pꢃ0.01, compared to the control (Student’s t-test).
transporters such as P-glycoprotein (P-gp) (also called as here, although the small number of radiotracers (nꢁ5), sug-
ABCB1) and multidrug resistance-associated proteins gesting that the lipophilicity could be used to predict the
(MRPs).38) Recently, the modulation of several PET-radioli- BBB permeability for derivatives of indomethacin and re-
gands by P-gp in vivo has been evaluated in rodents, showing lated compounds. In this study, we found that one compound
that P-gp plays an important role in restricting brain uptake [11C]3 features relatively low lipophilicity (log P7.4ꢁ2.40)
of radioligands.39,40) Low brain uptakes of four radiotracers while retaining high COX-2 inhibitory potency and high se-
may attribute to the P-gp. Zrieki et al. reported that the hep- lectivity. This might be good news for guiding the design of
tyl ester derivative of indomethacin is neither a substrate nor COX-2 selective radiotracers. Further structural refinement
a competitive inhibitor of P-gp in Caco-2 cells.41) On the of the radiotracers is necessary to achieve better enhance-
other hand, in a previous study, we found that 2-(4-aminosul- ment of brain uptake and sufficient metabolic stability.
fonylphenyl)-3-(8-methoxyphenyl)-indole with high COX-2-
Experimental
inhibiting activity was found to be a substrate for the P-gp in
General N-Pentyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-
in vitro monolayer efflux assays, but its 11C-labeled analog
acetate (1), N-octyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-acet-
amide (2)21) and celecoxib45) were prepared as previously described. Other
chemical reagents and solvents were of special grade and were used without
further purification unless otherwise noted.
was not sensitive to P-gp modulation with cyclosporine (Cys)
A in vivo.12) In this study, we investigated P-gp modulation of
four radiotracers, although they were not applied to the in
vitro monolayer efflux assays. As shown in Table 6, pretreat-
ment with CysA (50 mg/kg) seemed to slightly enhanced
brain uptake and/or the brain-to-blood ratio for [11C]2,
[11C]3, [11C]5 and [11C]6 with and without statistical signifi-
cance, thus seemingly indicating a possible contribution of P-
gp-mediated transport of these 11C-tracers. CysA has several
peripheral effects, not mediated by P-gp blockade, such as
inhibition of cytochrome P450 and toxicity to liver and kid-
ney.42,43) In this study we found that each radiotracer suffered
extensive metabolism and that two or three radiometabolites
were present in both brain and plasma tissues with high lev-
els of radioactivity, although the metabolic profile was meas-
1H-NMR spectra were recorded on a Varian Unity 400 (400 MHz, Califor-
nia, U.S.A.) and the chemical shifts are referenced to trimethylsilane (TMS)
(dꢁ0 ppm). IR spectra were recorded with a Shimadzu FTIR-8400 spec-
trometer (Kyoto, Japan) and mass spectra were obtained with a JEOL JMS
DX-610 (FAB Mass, Tokyo, Japan) or an Applied Biosystems Mariner Sys-
tem 5299 spectrometer (electrospray ionization (ESI) MS, California,
U.S.A.). All melting points were determined on a Yanaco melting point ap-
paratus (Yanagimoto Ind. Co., Kyoto, Japan) and are uncorrected. Column
chromatography was performed on Silica gel 60N (63—210 mesh, Kanto
Chemical Co., Tokyo, Japan). The progress of the reaction was monitored by
TLC on Silica gel 60F 254 plates (Merck, Darmstadt, Germany) and spots
were visualized with UV light or iodine. In the synthetic procedures, the or-
ganic extracts were routinely dried over anhydrous Na2SO4 and evaporated
with a rotary evaporator.
Preparation of [11C]methyl triflate from [11C]CO2 via [11C]methyl iodide
ured at a one-time point (15 or 30 min postinjection). Con- and subsequent 11C-methylation were carried out automatically by using a
synthetic apparatus for 11C-labeled compounds at the Tokyo Metropolitan
sidering these findings, this modest increase in cerebral up-
take of radioactivity induced by CysA, was probably caused
by altered cytochrome P450-mediated metabolism, rather
Institute of Gerontology. Preparative HPLC was done using a Shimadzu Liq-
uid Chromatograph System (SPD-10A, Kyoto, Japan), equipped with a UV
detector set at 260 nm and a semiconductor radiation detector system under
than a P-gp-mediated efflux of the parent radiotracer it-
the indicated conditions. Radiochemical purity and specific radioactivity of
the labeled compounds were determined by the same HPLC system under
the indicated conditions. The radioactivity was also quantified with an auto-
well gamma counter (LKB Wallac Compugamma 1282CS, Turku, Finland).
HPLC for metabolite analysis was done with a Shimadzu Liquid Chromato-
graph System (SPD-6AVA, Kyoto, Japan) having a UV detector set at
260 nm, and a packed flow scintillation analyzer 150TR (PerkinElmer,
Waltham, MA, U.S.A.).
self.37,44)
Conclusion
This paper describes the effect of altering the structure of
the alkyl side chain of indomethacin esters and amides in re-
gard to COX-2 inhibitory potency and in vivo brain uptake in
order to find a suitable radiotracer candidate for imaging
brain COX-2. The in vivo behavior of all radiotracers in mice
is characterized by low brain uptake and fast metabolic sus-
ceptibility. No specific binding of the radiotracers toward the
COX-2 enzyme in the brain was clearly revealed, although
[11C]3 seemed to show a small specific signal. It should be
noted that there was an inverse correlation between the brain
uptake of radioactivity at 1 min postinjection and the
Male ddY mice (7—8 weeks old) were obtained from Tokyo Laboratory
Animals (Tokyo, Japan). The animal studies were approved by the Animal
Care and Use Committee of the Tokyo Metropolitan Institute of Gerontology.
N-Pentyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-acetamide
(3) According to the procedure reported by Marnett and colleagues,21) a re-
action mixture containing indomethacin (1000 mg, 2.8 mmol) and bis-(2-
oxo-3-oxazolidinyl)phosphinic chloride (727 mg, 2.8 mmol, 1 eq) in anhy-
drous CH2Cl2 (20 ml) was treated with Et3N (770 ml, 5.5 mmol, 2 eq) and al-
lowed to stir at room temperature for 10 min. The mixture was then treated
with 1-pentylamine (371 ml, 3.2 mmol, 1.1 eq) and stirred overnight at room
lipophilicity values (log P7.4) of the 11C-tracers investigated temperature. The reaction mixture was filtered and the filtrate was concen-

944
Vol. 59, No. 8
trated in vacuo. The residue was purified by chromatography on silica gel
with hexane : EtOAcꢁ3 : 1 to provide 3 (843 mg, 70.6%) as a pale yellow
solid: mp 137—138 °C; H-NMR (CDCl3) d: 7.66 (2H, d, Jꢁ8.4 Hz), 7.49
(2H, d, Jꢁ8.4 Hz), 6.87 (2H, m), 6.70 (1H, dd, Jꢁ2.4, 9.2 Hz), 5.56 (1H, s),
ica gel with hexane : EtOAcꢁ3 : 1 to give 9 (55 mg, 57.3%) as a dark brown
solid: mp 194—195 °C; 1H-NMR (CDCl3) d: 7.67(2H, d, Jꢁ8.8 Hz), 7.48
(2H, d, Jꢁ8.4 Hz), 6.85 (2H, m), 6.64 (1H, dd, Jꢁ2.4, 8.8 Hz), 6.18 (1H, s),
5.61 (1H, s), 3.62 (3H, s), 3.20 (2H, m), 2.37 (3H, s), 1.40 (2H, m), 1.26
(4H, m), 0.82 (3H, t, Jꢁ7.2 Hz); FT-IR (KBr) cmꢅ1: 3212, 1665, 1641;
FAB-MS m/z: 413.2 (MꢆH)ꢆ. HR-MS: Calcd for C23H26ClN2O3 (MꢆH)ꢆ:
413.1626, 415.1607. Found 413.1672, 415.1653.
Butyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-ylmethylcar-
bamate (10) Compound 10 was obtained in a similar manner as 7 by
demethylation of 4 using BBr3. The crude product was purified by chro-
matography on silica gel with hexane : EtOAcꢁ4 : 1 to give 10 (106 mg,
73.3%) as a beige powder: mp 91—93 °C; 1H-NMR (CDCl3) d: 7.65 (2H, d,
Jꢁ8.4 Hz), 7.47 (2H, d, Jꢁ8.4 Hz), 7.00 (1H, s), 6.81 (1H, d, Jꢁ8.8 Hz),
6.60 (1H, dd, Jꢁ2.4, 8.8 Hz), 5.19 (1H, s), 4.77 (1H, s), 4.44 (2H, d,
Jꢁ5.2 Hz), 4.12 (2H, m), 2.40 (3H, s), 1.38 (2H, m), 1.26 (2H, m), 0.92
(3H, m); FT-IR (neat) cmꢅ1: 3327, 1686; FAB-MS m/z: 415.26 (MꢆH)ꢆ.
1-p-Chlorobenzoyl-5-hydroxy-2-methylindole-3-acetatic Acid (11)
Compound 11 was obtained in a similar manner as 7 by demethylation of in-
domethacin using BBr3. The crude product was purified by chromatography
on silica gel with CH3Cl : acetoneꢁ2 : 1 to give 11 (543 mg, 56.5%) as a white
powder: mp 223—225 °C; 1H-NMR (DMSO-d6) d: 9.16 (1H, s), 7.65 (4H, q,
Jꢁ6.0, 8.8 Hz), 6.86 (1H, d, Jꢁ8.8 Hz), 6.82 (1H, d, Jꢁ2.0 Hz), 6.55 (1H, dd,
Jꢁ2.4, 8.8 Hz), 3.57 (2H, s), 2.19 (3H, s); FAB-MS m/z: 344.2 (MꢆH)ꢆ. HR-
MS: Calcd for C18H14ClNO4 (MꢆH)ꢆ: 344.0684. Found 344.0712.
1
3.82 (3H, s), 3.63 (2H, s), 3.20 (2H, m), 2.38 (3H, s), 1.39 (2H, m), 1.21
(4H, m), 0.83 (3H, t, Jꢁ7.2 Hz); Fourier transform (FT)-IR (KBr) cmꢅ1
:
3300, 1670, 1638; FAB-MS m/z: 427.25 (MꢆH)ꢆ. HR-MS: Calcd for
C24H28ClN2O3 (MꢆH)ꢆ: 427.1783, 429.1764. Found 427.1795, 429.1804.
Butyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-yl Methylcar-
bamate (4)
A solution of indomethacin (500 mg, 1.4 mmol) and
diphenylphosphoryl azide (1.2 ml, 5.6 mmol, 4 eq) in dry benzene (22 ml) in
the presence of Et3N (3.8 ml, 27.5 mmol, 20 eq) was heated under reflux for
1 h. After adding 1-butanol (25 ml, 272.5 mmol, 194 eq), the reaction solu-
tion was further heated for 1.5 h, cooled, and concentrated in vacuo. The
residue was dissolved in EtOAc, filtered through a short path of celite, and
the solvent was removed in vacuo. The residue was purified by chromatogra-
phy on silica gel with CHCl3 : hexane : acetoneꢁ5 : 90 : 5 to provide 4
(414 mg, 69.1%) as a pale yellow solid: mp 103—107 °C; 1H-NMR (CDCl3)
d: 7.65 (2H, d, Jꢁ8.4 Hz), 7.47 (2H, d, Jꢁ8.0 Hz), 7.03 (1H, m), 6.83 (1H,
d, Jꢁ8.8 Hz), 6.67 (1H, dd, Jꢁ2.4, 8.8 Hz), 4.74 (1H, s), 4.47 (2H, d,
Jꢁ5.6 Hz), 4.10 (2H, m), 3.83 (3H, s), 2.42 (3H, s), 1.59 (2H, m), 1.38 (2H,
m), 0.92 (3H, t, Jꢁ7.2 Hz); FT-IR (KBr) cmꢅ1: 3379, 1721; FAB-MS m/z:
428.2 (M)ꢆ. HR-MS: Calcd for C23H26ClN2O4 (MꢆH)ꢆ: 429.1576. Found
429.1613.
N-2-Ethoxyethyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-
acetamide (5) Compound 5 was obtained in a similar manner to the syn-
thesis of 3 from indomethacin using 2-ethoxy ethylamine. The crude product
was purified by chromatography on silica gel with hexane : EtOAcꢁ2 : 1 to
provide 5 (580 mg, 48.4%) as a pale yellow solid: mp 99—102 °C; 1H-NMR
(CDCl3) d: 7.72 (2H, d, Jꢁ8.4 Hz), 7.48 (2H, d, Jꢁ8.8 Hz), 6.90 (1H, d,
Jꢁ2.4 Hz), 6.86 (1H, d, Jꢁ9.2 Hz), 6.69 (1H, dd, Jꢁ2.6 Hz, 9.0 Hz), 6.00
(1H, s), 3.82 (3H, s), 3.64 (2H, s), 3.40 (4H, s), 3.34 (2H, m), 2.39 (3H, s),
1.00 (3H, t, Jꢁ7.0 Hz); FT-IR (KBr) cmꢅ1: 3325, 2924, 1681, 1637; FAB-
MS m/z: 429.3 (MꢆH)ꢆ. HR-MS: Calcd for C23H26ClN2O4 (MꢆH)ꢆ:
429.1576, 431.1557. Found 429.1560, 431.1587.
N-2-Ethoxyethyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-
acetamide (12) Compound 12 was obtained in a similar manner to the
synthesis of 3 from 11 using 2-ethoxy ethylamine. The crude product was
purified by chromatography on silica gel with CHCl3 : acetoneꢁ9 : 1 to give
12 (74 mg, 33.7%) as a pale yellow solid: mp 112—114 °C; 1H-NMR
(CDCl3) d: 7.66 (2H, d, Jꢁ8.8 Hz), 7.30 (2H, d, Jꢁ8.4 Hz), 6.86 (1H, d,
Jꢁ2.0 Hz), 6.80 (1H, d, Jꢁ8.8 Hz), 6.66 (1H, s), 6.62 (1H, dd, Jꢁ2.4,
8.8 Hz), 6.12 (1H, s), 3.63 (2H, s), 3.39 (6H, s), 2.38 (3H, s), 1.02 (3H, t,
Jꢁ7.0 Hz); FT-IR (KBr) cmꢅ1: 3244, 2929, 1658, 1653; ESI-MS m/z: 415.2
(MꢆH)ꢆ. HR-MS: Calcd for C22H24ClN2O4 (MꢆH)ꢆ: 415.1419, 417.1399.
Found 415.1424, 417.1431.
N-3-Methoxypropyl-(1-p-chlorobenzoyl-5-methoxy-2-methylindole)-3-
acetamide (6) Compound 6 was obtained in a similar manner to the syn-
thesis of 3 from indomethacin using 3-methoxy propylamine. The crude
product was purified by chromatography on silica gel with hexane :
EtOAcꢁ3 : 2 to provide 6 (600 mg, 50%) as a pale yellow solid: mp 119—
121 °C; 1H-NMR (CDCl3) d: 7.67 (2H, d, Jꢁ8.6 Hz), 7.48 (2H, d,
Jꢁ8.5 Hz), 6.88 (2H, m), 6.70 (1H, dd, Jꢁ2.4, 8.9 Hz), 6.23 (1H, s), 3.83
(3H, s), 3.61 (2H, s), 3.33 (4H, m), 3.01 (3H, s), 2.36 (3H, s), 1.66 (2H, q,
Jꢁ6.0 Hz); FT-IR (KBr) cmꢅ1: 3269, 2922, 1695, 1643; FAB-MS m/z: 429.1
(MꢆH)ꢆ. HR-MS: Calcd for C23H26ClN2O4 (MꢆH)ꢆ: 429.1576, 431.1557.
Found 429.1600, 431.1591.
N-3-Methoxypropyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-
acetamide (13) Compound 13 was obtained in a similar manner to the
synthesis of 3 from 11 using 3-methoxy propylamine. The crude product
was purified by chromatography on silica gel with CHCl3 : acetoneꢁ9 : 1 to
1
give 13 (132 mg, 54.9%) as a pale yellow solid: mp 132—134 °C; H-NMR
(CDCl3) d: 7.73 (1H, s), 7.66 (2H, d, Jꢁ9.0 Hz), 7.47 (2H, d, Jꢁ7.6 Hz),
6.90 (1H, d, Jꢁ2.0 Hz), 6.84 (1H, d, Jꢁ8.8 Hz), 6.66 (1H, dd, Jꢁ2.4,
8.8 Hz), 6.46 (1H, t, Jꢁ5.2 Hz), 3.61 (2H, s), 3.33 (4H, m), 3.02 (3H, s),
2.36 (3H, s), 1.67 (2H, q, Jꢁ6.0 Hz); FT-IR (KBr) cmꢅ1: 3201, 2928, 1664,
1635; ESI-MS m/z: 415.2 (MꢆH)ꢆ. HR-MS: Calcd for C22H24ClN2O4
(MꢆH)ꢆ: 415.1419, 417.1399. Found 415.1384, 417.1408.
N-Pentyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-acetate (7)
Boron tribromide (BBr3) (1 M in CH2Cl2, 0.8 ml, 0.8 mmol, 6.8 eq) was
added dropwise to an ice-cooled solution of 1 (50 mg, 117 mmol) in anhy-
drous CH2Cl2 (3 ml). The reaction mixture was stirred at 5 °C for 0.5 h. After
stirring overnight at room temperature, the reaction mixture was diluted with
water and extracted with CH2Cl2. The combined organic phase was washed
with water, dried and filtered, and the solvent was removed in vacuo. The
residue was purified by chromatography on silica gel with CHCl3 :
methanolꢁ99 : 1 to provide 7 (44 mg, 90.2%) as a dark brown solid: mp
72—74 °C: 1H-NMR (CDCl3) d: 7.65 (2H, d, Jꢁ8.4 Hz), 7.47 (2H, d,
Jꢁ8.0 Hz), 6.92 (1H, s), 6.86 (1H, d, Jꢁ8.6 Hz), 6.58 (1H, d, Jꢁ9.2 Hz),
4.60 (1H, s), 4.09 (2H, m), 3.62 (2H, s), 2.36 (3H, s), 1.61 (2H, m), 1.29
(4H, m), 0.86 (3H, t, Jꢁ6.6 Hz); FT-IR (neat) cmꢅ1: 3400, 2929, 1681; FAB-
MS m/z: 413.2 (M)ꢆ. HR-MS: Calcd for C23H25ClNO4 (MꢆH)ꢆ: 414.1467,
416.1448. Found 414.1448, 416.1428.
N-Octyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-acetamide
(8) Compound 8 was obtained in a similar manner as 7 by demethylation
of 2 using BBr3. The crude product was purified by chromatography on sil-
ica gel with CHCl3 : acetoneꢁ17 : 1 to give 8 (134 mg, 92.0%) as a pale yel-
low solid: mp 192—194 °C; 1H-NMR (DMSO-d6) d: 9.09 (1H, s), 7.87 (1H,
m), 7.68 (2H, d, Jꢁ8.8 Hz), 7.63 (2H, d, Jꢁ8.4 Hz), 6.89 (2H, m), 6.55 (1H,
d, Jꢁ8.8 Hz), 3.41 (2H, s), 3.03 (2H, m), 2.18 (3H, s), 1.39 (2H, m), 1.22
(10H, m), 0.84 (3H, t, Jꢁ6.6 Hz); FT-IR (KBr) cmꢅ1: 3332, 3219, 2925,
2856; FAB-MS m/z: 455.3 (MꢆH)ꢆ. HR-MS: Calcd for C26H32ClN2O3 (Mꢆ
H)ꢆ: 455.2096. Found 455.2131.
Inhibition of COX Activity Inhibition of COX activity was assayed by
using the Colorimetric COX (ovine) Inhibitor Screening Assay kit (Cayman
Chemical, No. 760111, lot; 194562, 040471, 0407481, 0408012, Michigan,
U.S.A.). This assay measures the heme-catalyzed hydroperoxidase activity
of ovine COX by monitoring the appearance of oxidized N,N,Nꢂ,Nꢂ-tetra-
methyl-p-phenylenediamine (TMPD).26) Dimethylsulfoxide (DMSO) (10 ml,
control) or a solution of the studied compound in DMSO (10ꢅ2—10ꢅ8 M)
was added to a 96-well plate with 0.1 M Tris–HCl assay buffer (pH 8.0)
(150 ml), 4.4% solution of heme in DMSO (10 ml) and a solution of ovine
COX-1 or COX-2 in 80 mM Tris–HCl (pH 8.0) containing 0.1% Tween 20
and 300 mM diethyldithiocarbamate. After 5 min of preincubation at room
temperature, a solution of TMPD (20 ml) and arachidonic acid (20 ml) dis-
solved in 1.1 M ethanol containing 5 mM KOH was added to the mixture. The
mixture was incubated for a further 5 min and the absorbance was measured
on a plate reader at 595 nm. Celecoxib and indomethacin were used as the
reference compounds. The results are shown in Table 1.
11C-Labeling of Compounds 1, 2, 3, 5 and 6 [11C]Methyl triflate was
trapped in acetone (250 ml) containing each precursor (0.25 mg, 0.54—
0.60 mmol) at room temperature. Immediately before starting the radiosyn-
thesis, a solution of aqueous NaOH was added in the reaction vessel: 3 ml of
0.2 M NaOH (0.60 mmol, 1 eq) for 1 and 5 ml of 1 M NaOH (5 mmol) for the
other four compounds. After trapping [11C]methyl triflate, a mixture (1.4 ml)
of H2O and the mobile phase of HPLC (1 : 1) was added and then the reac-
tion mixture was applied to a semi-preparative HPLC column (YMC-Pack
ODS-A, 10 mmꢇ150 mm, YMC, Kyoto, Japan). At a 5 ml/min flow rate, a
mixture of acetonitrile (MeCN) and H2O was used for the mobile phase:
MeCN : H2Oꢁ9 : 1 for [11C]1 (retention time (tR): 6.1 min for [11C]1 and
3.8 min for 7) and [11C]2 (tR: 5.6 min for [11C]2 and 3.9 min for 8); 7 : 3 for
N-Pentyl-(1-p-chlorobenzoyl-5-hydroxy-2-methylindole)-3-acetamide
(9) Compound 9 was obtained in a similar manner as 7 by demethylation
of 3 using BBr3. The crude product was purified by chromatography on sil-

August 2011
945
[11C]3 (tR: 7.2 min for [11C] 3 and 4.0 min for 9); and 6 : 4 for [11C]5 (tR: for [11C]3, [11C]5 and [11C]6 at a flow rate of 2—3 ml/min. The results are
6.5 min for [11C] 5 and 3.6 min for 12) and [11C]6 (tR: 5.9 min for [11C] 6 and
3.3 min for 13). Each [11C]product fraction was collected in a flask contain-
ing a solution of L-ascorbic acid (0.1 ml, 100 mg/ml) in water and evaporated
to dryness. The residue was dissolved in physiological saline containing
0.125% Tween 80. The labeled compound was analyzed by HPLC: TSKgel
Super-ODS column (4.6 mm ꢇ100 mm, Tosoh, Tokyo, Japan); mobile
phase, MeCN : H2Oꢁ8 : 2 for [11C]1 and [11C]2; 7 : 3 for [11C]3; and 6 : 4 for
[11C]5 and [11C]6; and 1.0 ml/min flow rate. Retention times were 3.6 min
shown in Table 4.
Acknowledgments The financial support of the program for the Cre-
ation of Innovation Centers for Advanced Interdisciplinary Research Areas
from Special Coordination Funds for Promoting Science and Technology
(SCF) commissioned by the Ministry of Education, Culture, Sports, Science
and Technology (MEXT) of Japan is gratefully acknowledged. We appreci-
ate for the technical supports from the Research Support Center, Graduate
for [11C]1, 3.0 min for [11C]2, 5.3 min for [11C]3, 3.0 min for [11C]5 and School of Medical Sciences, Kyushu University, and Mr. Kunpei Hayashi for
2.7 min for [11C]6. The results are shown in Table 2.
radiosynthesis.
Octanol–Water Partition Coefficient A mixture of each radiotracer
(1.5—2.8 MBq) in octanol (4 ml for [11C]1, [11C]2, [11C]5 and [11C]6, or
2 ml for [11C]3) and an equivalent volume of phosphate buffered saline
(PBS; 1/15 M, pH 7.4) was vortexed at 20 s ꢇ3 times and then centrifuged
(1500 rpm for 1 min). An aliquot was taken from the organic phase (0.2 ml)
and the aqueous phase (0.2 ml for [11C]1 and [11C]2, 0.4 ml for [11C]3, 1 ml
for [11C]5 and [11C]6) and the 11C radioactivity was measured in an auto-
gamma-counter. The octanol–water partition coefficient was calculated as
the radioactivity ratio between the octanol and aqueous phases, and its mean
logarithm (log P7.4) was determined to express as lipophilicity. The results
are shown in Table 2.
Tissue Distribution of Radiotracers in Mice Male ddY mice (8—9
weeks old, body weight; 34—41 g) were used for the following experiments:
(A) tissue distribution, (B) a blocking study to determine specific uptake and
(C) cyclosporine A (CysA) treatment, and all mice were injected with each
radiotacer (2.0—2.8 MBq/8—97 pmol) in a physiological saline (0.2 ml)
containing 0.125% Tween 80 through the tail vein.
Group A mice were killed by cervical dislocation 1, 5, 15, 30 and 60 min
after radiotracer injection. Group B mice received the following blocker so-
lution (1 mg/kg per animal, 0.1 ml), carrier compounds (1, 2, 3, 5, or 6), NS-
398, celecoxib, and indomethacin dissolved in DMSO, co-injected with each
radiotracer. In the control mice the same amount of DMSO was co-injected.
The mice were killed at 30 min after injection for [11C]1, [11C]2 and [11C]3
and at 15 min for [11C]5 and [11C]6. Group C mice were treated with an in-
travenous injection of cyclosporine A (CysA, Sandimmun®, Tokyo, Japan)
(50 mg/kg)46,47) dissolved in a physiological saline 30 min prior to the tracer
injection. Each radiotracer was intravenously injected into CysA-treated and
control mice, and the radioactivity levels of the brain and blood were meas-
ured at 30 min for [11C]2 and [11C]3, and at 15 min for [11C]5 and [11C]6.
The 11C in the samples obtained was counted with an auto-gamma-
counter and the tissues were weighed. The tissue uptake of 11C was ex-
pressed as a percentage of injected dose per gram of tissue (%ID/g). Stu-
dent’s unpaired t-test was used to compare the groups in the pharmacologi-
cal experiments; a value of pꢃ0.05 was considered statistically significant.
The results are shown in Table 3, Table 5 and Table 6.
References
1) Crofford L. J., J. Rheumatol., 49, 15—19 (1997).
2) Fitzpatrick F. A., Curr. Pharm. Des., 10, 577—588 (2004).
3) Blobaum A. L., Marnett L. J., J. Med. Chem., 50, 1425—1441 (2007).
4) Chandrasekharan N. V., Dai H., Roos K. L., Evanson N. K., Tomsik J.,
Elton T. S., Simmons D. L., Proc. Natl. Acad. Sci. U.S.A., 99, 13926—
13931 (2002).
5) Mitchell J. A., Warner T. D., Br. J. Pharmacol., 128, 1121—1132
(1999).
6) Hewett S. J., Bell S. C., Hewett J. A., Pharmacol. Ther., 112, 335—
357 (2006).
7) Minghetti L., J. Neuropathol. Exp. Neurol., 63, 901—910 (2004).
8) Lleo A., Galea E., Sastre M., Cell. Mol. Life Sci., 64, 1403—1418
(2007).
9) FizGerald G.A., Nat. Rev. Drug Discov., 2, 879—890 (2003).
10) de Vries E. F. J., Curr. Pharm. Des., 12, 3847—3856 (2006).
11) Fujisaki Y., Kawamura K., Wang W.-F., Ishiwata K., Yamamoto F.,
Kuwano T., Ono M., Maeda M., Ann. Nucl. Med., 19, 617—625
(2005).
12) Tanaka M., Fujisaki Y., Kawamura K., Ishiwata K., Qinggeletu,
Yamamoto F., Mukai T., Maeda M., Biol. Pharm. Bull., 29, 2087—
2094 (2006).
13) McCarthy T. J., Sheriff A. U., Graneto M. J., Talley J. J., Welch M. J., J.
Nucl. Med., 43, 117—124 (2002).
14) de Vries E. F. J., van Waarde A., Buursma A. R., Vaalburg W., J. Nucl.
Med., 44, 1700—1706 (2003).
15) Toyokuni T., Kumar J. S. D., Walsh J. C., Shapiro A., Talley J. J.,
Phelps M. E., Herschman H. R., Barrio J. R., Satyamurthy N., Bioorg.
Med. Chem. Lett., 15, 4699—4702 (2005).
16) Prabhakaran J., Underwood M. D., Parsey R. V., Arango V., Majo V. J.,
Simpson N. R., Van Heertum R., Mann J. J., Kumar J. S. D., Bioorg.
Med. Chem., 15, 1802—1807 (2007).
17)
Wuest F., Kniess T., Bergmann R., Pietzsch J., Bioorg. Med. Chem.,
16, 7662—7670 (2008).
18) de Vries E. F. J., Doorduin J., Dierckx R. A., van Waarde A., Nucl.
Med. Biol., 35, 35—42 (2008).
19) Singh P., Mittal A., Mini Rev. Med. Chem., 8, 73—90 (2008).
20) Clark D. E., Drug Discov. Today, 8, 927—933 (2003).
21) Kalgutkar A. S., Marnett A. B., Crews B. C., Remmel R. P., Marnett L.
J., J. Med. Chem., 43, 2860—2870 (2000).
22) Kalgutkar A. S., Crews B. C., Saleh S., Prudhomme D., Marnett L. J.,
Bioorg. Med. Chem., 13, 6810—6822 (2005).
23) Kalgutkar A. S., Crews B. C., Rowlinson S. W., Marnett A. B., Kozak
K. R., Remmel R. P., Marnett L. J., Proc. Natl. Acad. Sci. U.S.A., 97,
925—930 (2000).
24) Uddin Md. J., Crews B. C., Blobaum A. L., Kingsley P. J., Ghebrese-
lasie K., Saleh S. S., Clanton J. A., Baldwin R. M., Marnett L. J., J.
Labelled Compd. Radiopharm., 52, 387—393 (2009).
25) Uddin Md. J., Crews B. C., Blobaum A. L., Kingsley P. J., Gorden D.
L., Mclntyre J. O., Matrisian L. M., Subbaramaiah K., Dannenberg A.
J., Piston D. W., Marnett L. J., Cancer Res., 70, 3618—3627 (2010).
26) Sharma R., Sainy J., Chaturvedi S. C., Acta Pharm., 58, 317—326
(2008).
27) Wilson A. A., Jin L., Garcia A., DaSilva J. N., Houle S., Appl. Radiat.
Isot., 54, 203—208 (2001).
28) Dischino D. D., Welch M. J., Kilbourn M. R., Raichle M. E., J. Nucl.
Med., 24, 1030—1038 (1983).
Metabolite Analysis Each tracer (20—100 MBq/230—1500 pmol) was
intravenously injected into the mice (8—9 weeks old), and 30 min later they
were killed by cervical dislocation (nꢁ3). The blood was removed by heart
puncture using a heparinized syringe and the brain was removed.
The blood was centrifuged at 7000ꢇg for 1 min at 4 °C to obtain the
plasma, which (50 ml) was denatured with 0.5 ml of 10% trichloroacetic acid
(TCA) in MeCN or a 1 : 1 mixture of MeCN and 50 mM CH3COOH–
CH3COONa (50 : 50). The mixture was centrifuged under the same condi-
tions, and the precipitate was re-suspended in 0.5 ml of the same solution,
followed by centrifugation. This procedure was repeated twice. The com-
bined supernatant was diluted with an equal portion of 50 mM CH3COOH–
CH3COONa (50 : 50). No decomposition of the radiotracers under the condi-
tions treated was confirmed.
The cerebral cortex (50 mg) was homogenized in 1.0 ml of ice-cold 10%
TCA in MeCN or an ice-cold 1 : 1 mixture of MeCN and 50 mM CH3COOH–
CH3COONa (50 : 50). The homogenate was then treated as described in the
plasma sample.
After protein elimination, recovery of the radioactivity in the soluble frac-
tions was 82—98% for the plasma samples and 73—95% for brain samples
of each radiotracer.
The supernatant samples obtained from the plasma and brain ho-
mogenates were applied to HPLC analysis using a 10-mm Radial-Pak C18
column and a 6-mm Prep Nova-pak 60A equipped with an RCM 8ꢇ10 mod-
ule (8 mmꢇ100 mm, Waters, Millford, MA, U.S.A.) for [11C]1 and [11C]3
and for [11C]1 and [11C]2, respectively, and a reversed phase column (YMC-
Pack ODS-A, 10 mmꢇ250 mm, YMC, Kyoto, Japan). The mobile phase
consisted of a 8 : 2 mixture of MeCN and a 50 mM CH3COOH–CH3COONa
buffer (50 : 50) for [11C]1 and [11C]2, and a 7 : 2 mixture of the same solvent
29) Waterhouse R. N., Mol. Imaging Biol., 5, 376—389 (2003).
30) Boyer D., Bauman J. N., Walker D. P., Kapinos B., Karki K., Kalgutkar
A. S., Drug Metab. Dispos., 37, 999—1008 (2009).
31) Davis N. M., Skjodt N. M., Clin. Pharmacokinet., 38, 377—392
(2000).

946
Vol. 59, No. 8
32) Parepally J. M. R., Mandula H., Smith Q. R., Pharm. Res., 23, 873—
881 (2006).
48, 81—87 (2007).
41) Zrieki A., Farinotti R., Buyse M., Pharm. Res., 25, 1991—2001
33) Rapoport S. I., Arai T., Wakabayashi S., Channing M. A., Dunn B. B.,
(2008).
Der M. G., Bell J. M., Herscovitch P., Eckelman W. C., Chang M. C., J. 42) Niwa T., Yamamoto S., Saito M., Shiraga T., Takagi A., Yakugaku
Nucl. Med., 36, 2261—2267 (1995).
Zasshi, 127, 209—216 (2007).
34) Chang M. C., Arai T., Freed L. M., Wakabayashi S., Channing M. A.,
Dunn B. B., Der M. G., Bell J. M., Sasaki T., Herscovitch P., Eckelman
W. C., Rapoport S. I., Brain Res., 755, 74—83 (1997).
43) Wacher V. J., Silverman J. A., Zhang Y., Benet L. Z., J. Pharm. Sci.,
87, 1322—1330 (1998).
44) Kajosaari L. I., Niemi M., Neuvonen M., Laitila J., Neuvonen P. J.,
Backman J. T., Clin. Pharmacol. Ther., 78, 388—399 (2005).
35) Giovacchini G., Chang M. C., Channing M. A., Toczek M., Mason A.,
Bokde A. L., Connolly C., Vuong B.-K., Ma Y., Der M. G., Doudet D. 45) Penning T. D., Talley J. J., Bertenshaw S. R., Carter J. S., Collins P. W.,
J., Herscovitch P., Eckelman W. C., Rapoport S. I., Carson R. E., J.
Cereb. Blood Flow Metab., 22, 1453—1462 (2002).
36) Habgood M. D., Begley D. J., Abbott N. J., Cell. Mol. Neurobiol., 20,
231—253 (2000).
37) Remmel R. P., Crews B. C., Kozak K. R., Kalgutkar A. S., Marnett L.
J., Drug Metab. Dispos., 32, 113—122 (2004).
38) Cecchell R., Berezowski V., Lundquist S., Culot M., Renftel M., De-
houck M. P., Fenart L., Nat. Rev. Drug Discov., 6, 650—661 (2007).
39) Elsinga P. H., Hendrikse N. H., Bart J., Vaalburg W., van Waarde A.,
Curr. Pharm. Des., 10, 1493—1503 (2004).
Docter S., Graneto M. J., Lee L. F., Malecha J. W., Miyashiro J. M.,
Rogers R. S., Rogier D. J., Yu S. S., Anderson G. D., Burton E. G.,
Cogburn J. N., Gregory S. A., Koboldt C. M., Perkins W. E., Seibert
K., Veenhuizen A. W., Zhang Y. Y., Isakson P. C., J. Med. Chem., 40,
1347—1365 (1997).
46) de Vries E. F. J., Kortekaas R., van Waarde A., Dijkstra D., Elsinga P.
H., Vaalburg W., J. Nucl. Med., 46, 1384—1392 (2005).
47) Evens N., Bosier B., Lavey B. J., Kozlowski J. A., Vermaelen P.,
Baudemprez L., Busson R., Lambert D. M., Van Laere K., Verbruggen
A. M., Bormans G. M., Nucl. Med. Biol., 35, 793—800 (2008).
40) Ishiwata K., Kawamura K., Yanai K., Hendrikse N. H., J. Nucl. Med.,