Synthesis of 2-methyl-3-indolylacetic derivatives as anti-inflammatory agents that inhibit preferentially cyclooxygenase 1 without gastric damage

Article abstract of DOI:10.1021/jm0608199

Novel substituted 2-methyl-3-indolylacetic derivatives were synthesized and evaluated for their activity in vitro and in vivo on COX-1 and COX-2. Active compounds were screened to determine their gastrointestinal tolerability in vivo in the rat. Results s

Full text of DOI:10.1021/jm0608199

7774
J. Med. Chem. 2006, 49, 7774-7780
Synthesis of 2-Methyl-3-indolylacetic Derivatives as Anti-Inflammatory Agents That Inhibit
Preferentially Cyclooxygenase 1 without Gastric Damage
Elisa Perissutti,† Ferdinando Fiorino,† Christian Renner,‡ Beatrice Severino,† Fiorentina Roviezzo,§ Lidia Sautebin,§
Antonietta Rossi,§ Giuseppe Cirino,§ Vincenzo Santagada,† and Giuseppe Caliendo*,†
Dipartimento di Chimica Farmaceutica e Tossicologica and Dipartimento di Farmacologia Sperimentale, UniVersita` di Napoli “Federico II”,
Via D. Montesano, 49, 80131 Napoli, Italy, and Max-Planck-Institute of Biochemistry, Am Klopferspitz 18, 82152 Martinsried,
Munich, Germany
ReceiVed July 12, 2006
Novel substituted 2-methyl-3-indolylacetic derivatives were synthesized and evaluated for their activity in
vitro and in vivo on COX-1 and COX-2. Active compounds were screened to determine their gastrointestinal
tolerability in vivo in the rat. Results showed that 3 and 4 preferentially inhibited COX-1 in vitro and in
vivo. MD simulations indicated an induced fit for COX-1 but not for COX-2, probably because of a lower
plasticity of the latter.
Introduction
We have designed and synthesized novel substituted 2-meth-
yl-3-indolylacetic derivatives in an attempt to obtain COX-1
inhibitors sparing the gastrointestinal (GI) side effects. These
new derivatives present an acidic function related to a basic
moiety. In particular, the synthesized structures possess in
position 1 a N-methyl-4-piperidinyl ring as a basic and non-
aromatic pharmacophoric portion present in some drugs that
have shown COX-2 selectivity (i.e., Nimesulide,7 NS-398,8
RWJ-633569), with the purpose of verifying if these structural
changes would shift selectivity toward COX-2. All synthesized
compounds were screened in vitro for COX-1 and COX-2
inhibitory effects and then tested in vivo in carrageenin-induced
mouse paw edema. Compounds showing a favorable profile
were screened for GI tolerability in vivo in the rat and in specific
pain assays. Molecular modeling simulations were performed
to rationalize the inhibitory activity and to analyze the relevant
interactions between enzyme and ligand.
Nonsteroidal anti-inflammatory drugs (NSAIDs) are potent
cyclooxygenase (COX) inhibitors that are widely used for the
treatment of autoimmune and chronic inflammatory diseases.1
COX performs the first step in the conversion of arachidonic
acid into inflammatory prostaglandins. It adds two oxygen
molecules to arachidonic acid, initiating a set of reactions that
will ultimately create a host of unusual molecules. Two distinct
isoforms have been identified separately: COX-1 and COX-2.
COX-1 is known as a housekeeping enzyme constitutively
expressed in most mammalian tissues and is responsible for
keeping the stomach lining intact and maintaining functional
kidneys. While COX-2 is constitutively expressed only in
kidney, brain, and ovaries, it can be rapidly expressed during
inflammatory conditions by proinflammatory stimuli such as
IL-1, TNF-R, and LPS and agents such as carrageenan, releasing
metabolites that are used to induce inflammation, fever, and
pain.2
Chemistry
NSAIDs can block the binding of arachidonic acid in the
COX active site; it is thought to result from initial binding to
the Arg120 and acetylation of the Ser530 hydroxyl group (in the
case of aspirin) in the primary COX binding site. After this
interaction the enzyme is blocked and the inflammatory reaction
and the associated pain perception are controlled. However, the
chronic usage of these drugs is associated with the induction
of gastrointestinal mucosal lesions, perforations, and bleeding
in part of the population. Decreased renal functions have been
observed in some patients. The reason is that these COX
inhibitors also inhibit COX-1 to produce the necessary pros-
taglandins.3 Selective COX-2 inhibitors such as celecoxib elicit
effective anti-inflammatory activity devoid of the ulcerogenic
effects associated with the use of NSAIDs such as aspirin, which
inhibit COX-1 and COX-2.4,5 However, the lack of effect on
COX-1 does not allow COX-2 inhibitors to exert an antiplatelet
effect. Recently COX-2 inhibitors have been shown to produce
cardiovascular side effects and some of them such as rofecoxib
have been retired from the market.6
The synthetic route for the preparation of substituted 2-meth-
yl-3-indolylacetic derivatives (3-6) is reported in Scheme 1.
The aniline derivative 1 was obtained by a reductive amination
of p-anisidine with 1-methylpiperidin-4-one in presence of
sodium cyanoborohydride10 and molecular sieves,11 which serve
simultaneously as a dehydrating agent and as a catalyst. The
nitrosation followed by a reduction step with LiAlH4 afforded
the hydrazine 2 in 90% yield. The Fischer indole reaction of
hydrazine 2 was performed as a “one-pot” reaction under acidic
conditions yielding 3 (yield 71%) (Scheme 1). Starting from 3,
we obtained 4 by ethyl ester saponification and 6 by reaction
with HI and acetic anhydride at room temperature. Compound
5 was obtained by treating 4 with HI and acetic anhydride at
130 °C (Scheme 2). Analytical purification of each product
(Table 1) was obtained by chromatography on silica gel column
and by crystallization from the appropriate solvent. All new
compounds gave satisfactory elemental analysis results (C, H,
1
N) and were characterized by H NMR spectroscopy.
* To whom correspondence should be addressed. Phone: 0039-081-
678649. Fax: 0039-081-678649. E-mail: caliendo@unina.it.
† Dipartimento di Chimica Farmaceutica e Tossicologica, Universita` di
Napoli “Federico II”.
Results and Discussion
All the synthesized compounds were screened for COX
selectivity in an in vitro cell based assay. Stimulation of J774
macrophages with arachidonic acid (15 µM) for 30 min induced
a significant increase (p < 0.001) of PGE2 (3.6 ( 0.3 ng × l06
‡ Max-Planck-Institute of Biochemistry.
§ Dipartimento di Farmacologia Sperimentale, Universita` di Napoli
“Federico II”.
10.1021/jm0608199 CCC: $33.50 © 2006 American Chemical Society
Published on Web 12/21/2006

Synthesis of 2-Methyl-3-indolylacetic DeriVatiVes
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7775
Scheme 1
cells) levels in comparison to unstimulated cells (<0.0125 ng
× l06 cells). In the presence of increasing concentrations (0.01-
100 µM) of indomethacin (a nonselective COX-1/COX-2
inhibitor) a concentration-dependent inhibition of PGE2 bio-
synthesis was observed (Figure 1A). 3 and 4 (0.01-100 µM)
derivatives, related to indomethacin, exerted a similar concentra-
tion-dependent inhibition on PGE2 generation, although they
were less potent than indomethacin (Figure 1A). At the highest
concentration (100 µM), PGE2 production was significantly
inhibited by indomethacin (88%, p < 0.001), 4 (64%, p <
0.001), and 3 (58%, p < 0.001), with 4 and 3 displaying
equivalent inhibitory activity on COX-1. At lower concentration
(10 µM), indomethacin still significantly (p < 0.001) inhibited
PGE2 production to about 75%, displaying a more powerful
inhibition than 4 (57%, p < 0.001) and 3 (37%, p < 0.001). 5
and 6 were inactive at all the tested concentrations (0.01-100
Figure 1. (A) Effect of derivatives 3-6 vs indomethacin on COX-1
activity. 3 and 4 display a good inhibitory activity on COX-1, while 5
and 6 are weak inhibitors. (B) Effect of derivatives 3-6 vs DFU, a
selective COX-2 inhibitor, on COX-2 activity. (C) Effect of derivatives
3 and 4 vs indomethacin on mouse paw edema.
µM). Conversely, a significant (p < 0.00l) concentration-
dependent inhibition of PGE2 production (Figure 1B) was
induced by 5,5-dimethyl-3-(3-fluorophenyl)-4-(4-methylsulpho-
nyl)phenyl-2(5H)-furanone (DFU), a selective COX-2 inhibitor
(0.1-100 µM) while the new synthesized derivatives were
ineffective with the only exception of 3, which exerted 24% of
inhibition (p < 0.05) at the highest tested concentration (Figure
1B). Stimulation with LPS followed by wash-out and 30 min
of incubation with arachidonic acid (15 pM) induced a
significant (p < 0.001) increase of PGE2 generation (31.6 ( 2.
ng × 106 cells) in comparison to unstimulated cells (0.62 (
0.072 ng × l06 cells). After this preliminary screening, 3 and 4
were selected for the subsequent in vivo screening. They were
screened for their anti-inflammatory activity in the mouse edema
assay, while the analgesic activity was assessed by using two
different assays, e.g., formalin paw licking and acetic acid
induced writhing. 3 and 4 displayed significant anti-inflamma-
tory activity in carrageenin-induced mouse paw edema; 3 was
especially equipotent with indomethacin (Figure 1C). 3 and 4
displayed an analgesic activity comparable to indomethacin in
the second phase of formalin paw licking (Figure 2A), while
they were ineffective in the first phase (data not shown). In the
writhing assay, 3 and 4 displayed a dose related effect with 4
Table 1.
mp,
°C
cryst
yield,
%
compd
R′
CH3
CH3
H
R′′
formulaa
solventb
3
4
5
6
CH2CH3 C20H28N2O3
68-70
a
a
71
98
40
30
H
H
C18H24N2O3 172-173
C17H22N2O3 204-206
a + b
a
COCH3 CH2CH3 C21H28N2O4
98-99
a Satisfactory microanalyses obtained: C, H, N values are within (0.4%
of theoretical values. b Crystallization solvents: (a) diethyl ether; (b) ethyl
alcohol.

7776 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Perissutti et al.
Figure 2. (A) Analgesic effect of indomethacin, 3, and 4 on the second
phase of paw licking assay. Data are expressed as percent of inhibition.
Control value was 25 ( 4 s. (B) Analgesic effect of indomethacin, 3,
and 4 on acetic acid induced writhing. Data are expressed as percent
of inhibition. Control value was 39.8 ( 5.3.
Figure 3. (A) Crystal structure 4COX of COX-2 in complex with
indomethacin in light-gray with the flexible binding site of the COX-
1/indomethacin and COX-1/4 complex models in blue and green,
respectively. (B) Ligand and the conformation of the loop Val-349 to
Tyr-355 at the end of the 300 K simulation for the COX-2/indomethacin
(dark-gray), COX-1/indomethacin (blue), COX-2/4 (yellow), and COX-
1/4 (green) complexes. The X-ray structure 4COX is again shown in
light-gray as reference.
showing the better profile (Figure 2B). Interestingly, we have
also found that the ester derivative 3 is completely hydrolyzed
after 30 min to the corresponding acid derivative 4 by liver
metabolism. This evidence was verified by specific assays
performed on liver or mice plasma using HPLC analysis.
Importantly, 3 and 4 were devoid of GI damaging activity in
the rat GI damage assay. In the GI damage assay, indomethacin
caused a gastric damage of 50 ( 8 mm and 3 of 0.6 ( 0.03
mm, while 4 did not cause any damage in any of the six treated
rats. These data imply that the structural modification operated
on 4 preserved the anti-inflammatory and analgesic features of
indomethacin but markedly reduced the GI side effects. The
finding that a selective COX-1 inhibitor with anti-inflammatory
activity (1 and 2) does not display GI side effects could be
contrary to present knowledge. However, it has been shown
that SC560, a selective COX-1 inhibitor, does not cause GI
damage even though it inhibits PG synthesis in rat stomach and
platelet COX-1 activity.12 More recently SC560 has been
confirmed to have anti-inflammatory activity independent of
COX-1 inhibition13 and to be analgesic in rats.14
Molecular Modeling. Available crystal structures of COX-1
and COX-2 in complex with ligands are a solid experimental
basis for computer modeling calculations of the new 3-6 in
complex with COX-1 or COX-2. Unbiased starting models were
generated on the basis of the X-ray structure of the COX-2/
indomethacin complex15 (PDB code 4COX; for details see the
Experimental Section). During the molecular dynamics (MD)
simulation the ligand and a 7 Å binding site on the protein were
allowed to move freely with the flexible binding site comprising
all amino acid residues of the COX enzyme that had in the
4COX structure at least one atom closer than 7 Å to any atom
of indomethacin (see Figure 3A). The parts of the protein far
from the ligand binding site were fixed to avoid disintegration
of the global fold at higher temperatures. Steric clashes and
unfavorable interactions of the initial model were relaxed during
minimization and simulation at low temperature (100 K) for
50 ps. The subsequent productive period of 100 ps at 300 K
was followed by 50 ps at 500 K to test for stability of the
resulting complexes. Generally, rearrangements were observed
upon heating to 300 K, but the resulting structures remained
fairly stable at 300 K. Further heating to 500 K led to increased
fluctuations, but ligands never left the binding site and the
binding mode was conserved throughout. Note that conforma-
tional sampling at 50 ps at 500 K is quite effective and
corresponds to some microseconds at 300 K. Figure 3B shows
that 4 induces rearrangements within the protein environment
that are significantly more pronounced for COX-1 compared
to COX-2. This differential effect is exerted by 3-6 ligands
but only to minor extents by indomethacin, in agreement with
the low specificity of indomethacin with respect to the two COX
isoforms. The presence of the charged basic piperidine ring of
3-6 in place of the phenyl ring of indomethacin, which occupies
the hydrophobic pocket in the 4COX X-ray structure (see Figure
4A), leads to a displacement of the 3-6 ligands that in turn
appear to induce conformational changes to the protein. For
COX-2 these changes are small and the loss of binding to the
hydrophobic pocket renders 3-6 as ineffective inhibitors of
COX-2 (Figure 1B). Conversely, in simulations of the COX-1
complexes a new extended hydrophobic patch is formed because

Synthesis of 2-Methyl-3-indolylacetic DeriVatiVes
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7777
Figure 5. Binding site of the COX enzyme. Protein side chains and
the ligand, highlighted by a transparent green surface, are in atom colors,
and the protein backbone is in gray. (A) Binding of indomethacin to
COX-2 in the 4COX crystal structure. The hydrophobic pocket is
indicated by the transparent white surface. (B) Binding of 4 to COX-1
in the calculated model. The newly formed hydrophobic path is
indicated by the transparent white surface.
stituent (OH, OMe, OAc) and are likely due to the differential
hydrophobicity and size of the three substituents. A remarkable
feature is that the specificity thus arises not from structural
differences of COX-1 and COX-2 (the binding sites are basically
identical with very few differences in sequence) but from
different plasticities of the protein structures, at least around
the binding site (Figure 5). Higher B factors for COX-1 (PDB
code 1DIY, average B factor of 45.0 ( 12.4) than for COX-2
(PDB code 1CVU, average B factor of 27.6 ( 8.3), both
crystallized with the identical ligand arachidonic acid, support
the idea of higher flexibility in the former protein. The generally
accepted concept of the “induced fit” is thus transformed into
a “selectively induced fit” for 3-6. It is speculated that the
structural changes are related to the fact that less severe side
effects are observed for 3-6. In agreement with this hypothesis
is the observation that indomethacin in the MD simulations does
not induce pronounced changes to the COX-2 structure.
Figure 4. Energies and rmsd values of COX complexes in modeling
calculations. The rmsd values are calculated for heavy atoms. In the
case of the enzyme only, they were calculated for heavy atoms of the
flexible binding site. The temperature curve in black is always shown
for reference with a scale given on the right side of the graph. (A) The
rmsd values indicating structural rearrangements for both isoforms of
the COX/4 complex. (B) Energies of the complexes with 4 and
indomethacin reported as the energy difference of the COX-2 complex
minus the COX-1 complex . The kinetic energy difference (corrected
for the slightly different numbers of moving atoms) is fluctuating around
zero, as expected. C) flexibility of 4 in both COX complexes. (D) The
rmsd values of COX-1 and COX-2 complexes with the unselective
indomethacin (cf. panel A).
Conclusions
MD simulations show that the 3-6 inhibitors, because of the
presence of the polar piperidine, cannot bind to the hydrophobic
pocket in the COX binding site. However, COX-1 structural
rearrangements induced by the ligands seem to form a new
hydrophobic interaction site close to the indole moiety that
allows recovery of inhibitory activity, at least for suitably
5-substituted compounds. The lower plasticity of COX-2 does
not allow these conformational changes to occur and thus
prevents activity. Selectivity is thus achieved by a selective
of structural changes brought about by the 3-6 ligands, most
prominently in loop 351-355 (Figure 3). In that case the ligands
can recover their inhibitory activity by hydrophobic interactions
between this patch and the indole moiety (Figure 4B). The
differences among the 3-6 activities correlate with the 5-sub-

7778 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Perissutti et al.
mL). The organic layers were washed with brine, dried on
anhydrous MgSO4, and filtered, and the solvent was evaporated.
After chromatography on silica gel column (eluent, 94:0.5:1
dichloromethane/methanol/NH4OH), a yield of 19.7 g of 3 was
obtained (71%): mp 68-70 °C.
[5-Methoxy-2-methyl-1-(1-methyl-piperidin-4-yl)-1H-indol-3-
yl]acetic Acid (4). To a solution of 3 (8 g, 0.022 mol) in ethanol
(140 mL), an amount of 80 mL of 1 N NaOH was added. The
mixture was stirred for 2 h at room temperature. Then the solution
was neutralized with 1 N HCl. Ethanol was removed and the
extraction performed with ethyl acetate (3 × 150 mL). The organic
layers were washed with brine, dried on anhydrous MgSO4, and
filtered, and the solvent was evaporated, yielding 6.8 g of 4 (98%)
as a white solid: mp 172-173 °C.
[5-Hydroxy-2-methyl-1-(1-methylpiperidin-4-yl)-1H-indol-3-
yl]acetic Acid‚HI (5). To a mixture of 4 (3.2 g, 0.010 mol) and
acetic anhydride (16 mL) was added 57% hydroiodic acid (7.8 mL).
The mixture was stirred for 2 h at 130 °C. Successively, the mixture
was cooled to 0 °C, alkalinized with 5% NaHCO3, and then
extracted with dichloromethane (3 × 50 mL). The organic layers
were washed with brine, dried on anhydrous MgSO4, and filtered,
and the solvent was evaporated. The crude compound was
recrystallized from diethyl ether, yielding 1.2 g of 5 (40%) as a
solid: mp 204-206 °C.
[5-Acetoxy-2-methyl-1-(1-methylpiperidin-4-yl)-1H-indol-3-
yl]acetic Acid Ethyl Ester (6). To a mixture of 3 (3.2 g, 0.010
mol) and acetic anydride (16 mL) was added 57% hydroiodic acid
(7.8 mL). The mixture was stirred for 2 h at room temperature.
Successively, the mixture was cooled to 0 °C, alkalinized with 5%
NaHCO3, and then extracted with dichloromethane (3 × 50 mL).
The organic layers were washed with brine, dried on anhydrous
MgSO4, and filtered, and the solvent was evaporated. The crude
compound was recrystallized from diethyl ether, yielding 1.1 g of
6 (30%) as a solid: mp 98-99 °C.
Pharmacology. Cell Culture. The murine monocyte/macrophage
J774 cell line was grown in Dulbecco’s modified Eagle medium
(DMEM) supplemented with 2 mM glutamine, 25 mM HEPES,
penicillin (100 µg/mL), streptomycin (100 µg/mL), 10% fetal bovine
serum (FBS), and 1.2% sodium pyruvate (Bio Whittaker, Europe).
Cells were plated in 24-well culture plates at 2.5 × 105 cells/mL
or in 10 cm diameter culture dishes (1 × l07 cells/l0 mL per dish)
and allowed to adhere at 37 °C in 5% CO2/95% O2 for 2 h.
Immediately before the experiments, the culture medium was
replaced by fresh medium without FBS to avoid interference with
the radioimmunoassay and cells were stimulated as previously
described.16
Assessment of COX-1 Activity. Cells were pretreated with
reference or test compounds (0.01-100 µM) for 15 min and further
incubated at 37 °C for 30 min with 15 µM arachidonic acid to
activate the constitutive COX.8 Stock solutions of reference and
test compounds were prepared in ethanol, and equivalent amounts
of ethanol were included in the control samples. At the end of the
incubation the supernatants were collected for measurement of
PGE2 by radioimmunoassay.17
Assessment of COX-2 Activity: Exogenous Substrate. J774
cells were stimulated for 24 h with S. thyphosa lipopolysaccharide
(LPS, 10 µg/mL) to induce COX-2. The supernatant of the cells
was replaced with fresh medium containing arachidonic acid (15
µM) in the absence (controls) or presence of test compounds (0.01-
100 µM). Then cells were incubated for 30 min at 37 °C and the
supernatant was collected for the measurement of PGE2 by
radioimmunoassay.13 Stock solutions of reference and test com-
pounds were prepared in ethanol, and equivalent amounts of ethanol
were included in the control samples.
Chemical Stability of 3. Blood was withdrawn by cardiac
puncture from mice (18-20 g) and plasma obtained by centrifuga-
tion (12 000 rpm, 4 °C). Citrate (3.8%v/w) was used as anticoagu-
lant. Liver was obtained from mice and homogenized in RIPA
buffer (1 mg/1 mL). 3 (1 mg/mL) was dissolved in DMSO and
left to incubate with plasma or liver homogenate. The compounds
were incubated at 37 °C, and samples for analysis were taken after
propensity of the protein to the induced fit. In this respect, our
data have shed new light on the chemical requisites for designing
new COX inhibitors. 3 and 4 derivatives, which present an acidic
function related to a basic moiety, preferentially inhibit COX-1
in vitro and are analagesic and anti-inflammatory in vivo. In
addition and most importantly, both compounds appear to be
devoid of gastric effects.
Experimental Section
Chemistry. All reagents were commercial products purchased
from Aldrich. Melting points were determined using a Kofler hot-
stage apparatus and are uncorrected. 1H NMR spectra were recorded
on a Bruker AMX-500 instrument. All spectra were recorded in
CDCl3 or CD3OD. Chemical shifts are reported in ppm using Me4-
Si as the internal standard. The following abbreviations are used
to describe peak patterns when appropriate: s (singlet), d (doublet),
t (triplet), q (quadruplet), m (multiplet). Mass spectra of the final
products were obtained with a LCQ Thermoquest ion trap mass
spectrometer. Where analyses are indicated only by the symbols
of the elements, results are within (0.4% of the theoretical values.
All reactions were followed by TLC, carried out on Merck silica
gel 60 F254 plates with fluorescent indicator, and the plates were
visualized with UV light (254 nm). Preparative chromatographic
purifications were performed using a Kieselgel 60 silica gel column.
Solutions were dried over Na2SO4 and concentrated with a Bu¨chi
rotary evaporator at low pressure.
N-(1-Methyl-4-piperidinyl)-4-anisidine (1). A solution of p-
anisidine (30 g, 0.243 mol) and 1-methylpiperidin-4-one (21.3 g,
0.189 mol) in dry methanol (250 mL) was stirred for 6 h at room
temperature in the presence of molecular sieves (3 Å, 21 g; 5 Å,
21 g). Then acetic acid (150 mL) was added and the mixture was
stirred for 14 h. NaBH3CN (7.74 g, 0.123 mol) was added, and the
resulting solution was stirred for another 3 h. After filtration, the
molecular sieves were washed with methanol. The solvent was
evaporated, and the crude material was then dissolved in ethyl
acetate (350 mL) and washed with brine (3 × 50 mL) and water
(50 mL). The organic layer was dried on anhydrous Na2SO4 and
filtered, and the solvent was evaporated. After chromatography on
a silica gel column (eluent, 6:4 dichloromethane/methanol) 28 g
of 1 was obtained (68%) as an oil.
N-(1-Methyl-4-piperidinyl)-4-anisylhydrazine (2). Compound
1 (19.5 g, 0.089 mol) in absolute ethanol (240 mL) was cooled to
0 °C and mechanically stirred. Then 37% HCl (190 mL) was added
slowly. A solution of NaNO2 (7.93 g, 0.115 mol) in H2O (50 mL)
was added dropwise, keeping the mixture below 0 °C. After
completion of the addition, the solution was stirred at this
temperature for 3 h. Then water (600 mL) was added, and the
mixture was made alkaline with 6 N NaOH. The solution was
extracted with ethyl acetate (3 × 200 mL), washed with water,
dried on anhydrous Na2SO4, and filtered, and the solvent was
evaporated. The crude nitroso compound was used without further
purification. A suspension of LiAlH4 (9.5 g) in THF (84 mL) was
heated to reflux. The crude nitroso compound (25 g) in THF (60
mL) was added dropwise. The mixture was stirred at reflux for 1
h and then cooled to room temperature. Successively, water (15
mL) and aqueous 3 N NaOH (15 mL) were added dropwise. The
mixture was filtered on Celite, and the residue was washed several
times with dichloromethane. The solvent was removed, obtaining
18.7 g of 2 (90%).
[5-Methoxy-2-methyl-1-(1-methylpiperidin-4-yl)-1H-indol-3-
yl]acetic Acid Ethyl Ester (3). A solution of the hydrazine
derivative 2 (18.7 g, 0.080 mol) and ethyl levulinate (11.52 g, 0.080
mol) in absolute ethanol (430 mL) was stirred and heated to reflux
for 18 h. After evaporation of the solvent, ethyl acetate (300 mL)
was added. The organic phase was washed with brine, dried on
anhydrous MgSO4, and filtered, and the solvent was evaporated.
Sulfuric acid (96%, 4 mL) was added to a residue dissolved in
absolute ethanol, and the mixture was heated to reflux for 18 h.
Then the solution was made alkaline with 1 N NaOH. Ethanol was
removed and the extraction performed with ethyl acetate (3 × 150

Synthesis of 2-Methyl-3-indolylacetic DeriVatiVes
Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26 7779
30 or 60 min of incubation. Levels of 4 were monitored by HPLC
analysis. Reversed-phase analysis was routinely performed on a
Waters 600 system using a Vydac C18 silica (5 µm, 4.6 mm ×
250 mm) HPLC column. The elution was performed with the
following conditions: eluent A, 0.05% TFA (v/v) in water; eluent
B, 0.05% TFA (v/v) in acetonitrile; gradient 10-80% B over 25
min; UV detection at 254 nm; flow rate of 1 mL/min. The HPLC
analysis gave the following results: 3 at tR ) 14.34 min and 4 at
tR ) 10.25 min.
Mouse Paw Edema. Mice (18-20 g) were separated in groups
and lightly anesthetized with enflurane. Each group received
subplantar injection of 50 µL of carrageenan, 1 wt %/vol, into the
left footpad. Paw volume was measured by using a hydroplethis-
mometer modified for small volume (Ugo Basile, Italy) immediately
before the subplantar injection and 2, 4, 6, 24, 48, and 72 h
thereafter. The increase in paw volume was evaluated as the
difference between the paw volume at each time point and the basal
paw volume. Mice were treated 1 h prior to the injection of
carrageenan with 3-6.
closer than 7 Å to indomethacin (meaning any atom of indometha-
cin) were allowed to move without artificial constraints to allow
for an induced fit. Thus, a flexible binding site for COX-2 consisting
of Val-89, His-90, Leu-93, Thr-94, Met-113, Val-116, Leu-117,
Ser-119, Arg-120, Gln-192, Phe-198, Phe-205, Thr-206, Val-344,
Ile-345, Tyr-348 to Tyr-355, Phe-357, Leu-359, Phe-381, Leu-384,
Tyr-385, Trp-387, Val-434, Leu-507, Arg-513, Ala-516 to Lys-
532, and Leu-534 was constructed. 3-6 were placed inside the
enzyme by superimposition on heavy atoms that are common to
indomethacin and the new inhibitor molecules. For this purpose
the piperdine of 3-6 was oriented in the same way as the phenyl
ring of indomethacin. The protonation states were positively charged
for the piperadine moiety and negatively charged for the 3-acetic
acid substituent occurring only in 4 and 5. For the generation of
the COX-1/indomethacin complex models, two crystallographically
determined complex structures comprising COX-1 (PDB code
1DIY) and COX-2 (PDB code 1CVU) in complex with the same
ligand (i.e., arachidonic acid) were compared. Both structures were
found to be very similar in the general fold and in the specific
details of the binding site, although crystallographic B factors that
are a measure of local flexibility or heterogeneity are generally
larger for the COX-1 than for the COX-2 complex. Several other
COX/inhibitor complex structures ((COX-1) 1EBV, 1PGG; (COX-
2) 1PTH, 1CX2, 6COX) also exhibited similarities. Therefore, the
coordinates of COX-1 of the 1DIY structure were superimposed
on those of COX-2 of the COX-2/3-6 starting structure, thus
generating the COX-1/3-6 starting model. The COX-1/indometha-
cin complex was built in the same way as all calculations were
carried out for 3-6 and indomethacin as ligands, with indomethacin
serving as a reference. The flexible binding site of COX-1
comprised the same part of the sequence as for COX-2 but with a
few changes in amino acid residue types due to the differences in
sequence between the two COX isoforms. The following changes
occur within the 7 Å binding area when moving from COX-1 to
COX-2: I89V, V119S, Q351H, L357F, I434V, H513R, S516A,
S521T, I523V, M525L. In the first step of the calculations the model
complex consisting of the COX-1 or COX-2 enzyme and the
respective ligand was energy-minimized to relax steric clashes
generated by the placement of the inhibitor. MD simulated annealing
of the complex was started at 10 K (for 10 ps). Coupling to a
temperature bath (time constant of 1 ps) with stepwise increasing
target temperatures served to heat the simulation system gradually.
After 50 ps at 100 K the most relevant part of 100 ps at 300 K was
followed by an additional 50 ps at 500 K to test for the stability of
the complex established at 300 K.
Acetic Acid Induced Writhing. Male Swiss mice (25-30 g)
were injected intraperitoneally with acetic acid (0.6%) in a final
volume of 500 µL. The mice were placed in individual cages for
observation. After the first writhing movement appeared, the animals
were kept under observation for 15 min and the number of writhings
was counted during this period.
Formalin Paw Licking. Mice (18-20 g) were treated 1 h prior
to the injection of formalin with gabapentin or 3-6. Injection of
10 µL of formalin (5%) in the plantar area was performed under
light anesthesia with enflurane. Mouse paw licking behavior was
evaluated for two periods by an observer unaware of the treat-
ment: first phase, 0-15 min; second phase, 15-45 min.
Acute Gastric Damage. Rats (120-140 g) were deprived of
food but not water for 18 h and were then given 3-6 orally at a
dose of 100 mg/kg. Another group was treated with an equal volume
of the vehicle. Each group consisted of six rats. After 5 h, the rats
were anesthetized with sodium pentobarbital and the stomach was
excised and opened by an incision along the greater curvature. The
extent of macroscopic damage was determined by an observer
unaware of the treatments that the rats had received, as previously
described.18 The method involved measuring the lengths of the
lesions in millimeters and then adding the lengths of all lesions
observed in each stomach. Data were analyzed by ANOVA for
the nonparametric Kruskall-Wallis test followed by Dunn’s post-
test.
Statistical Analysis of Pharmacological Data. Triplicate wells
were used for the various treatment conditions. Results are expressed
as the mean of three experiments for the percent of inhibition of
PGE2 production by reference and test compounds with respect to
control samples. Data (nanograms of PGE2 produced) were
analyzed by using one-way ANOVA followed by a Bonferroni post
hoc test for multiple comparisons. A p value less than 0.05 was
considered statistically significant.
Molecular Modeling. All calculations were performed with the
DISCOVER program (Accelrys, San Diego, CA) on Silicon
Graphics O2 R5000 computers (SGI, Mountain View, CA). The
force field CVFF with a time step of 1 fs and a cutoff of 12 Å for
the nonbonded interactions was employed for all calculations. A
distance-dependent dielectric constant ꢀ ) 4r (r in Å) was used
for mimicking the dielectric properties inside the protein molecule.
The published X-ray structure of the COX-2 enzyme in complex
with indomethacin11 (PDB code 4COX) was used as the starting
point for the molecular modeling. Indomethacin and our new
compounds 3-6 are almost completely buried within the receptor,
and we have therefore not included explicitly the solvent to speed
up the calculations. A generous binding site around indomethacin
was allowed to move freely during the MD simulations, and no
artificial constraints were imposed on the ligands. To enhance the
stability of the protein structure during the simulation, especially
at higher temperatures, all parts of the enzyme that are far from
the binding site were fixed in space. Amino acid residues of the
COX-2 protein that had in the X-ray structure at least one atom
Acknowledgment. The NMR spectral data were provided
by Centro di Ricerca Interdipartimentale di Analisi Strumentale,
Universita` degli Studi di Napoli “Federico II”. The assistance
of the staff is gratefully appreciated.
Supporting Information Available: NMR, MS, and elemental
analysis data. This material is available free of charge via the
Internet at http://pubs.acs.org.
References
(1) Roller, A.; Bahr, O. R.; Streffer, J.; Winter, S.; Heneka, M.; Deininger,
M.; Meyermann, R.; Naumann, U.; Gulbins, E; Weller, M. Selective
potentiation of drug cytotoxicity by NSAID in human glioma cells:
the role of COX-1 and MRP. Biochem. Biophys. Res. Commun. 1999,
259, 600-605.
(2) Lu, J. X.; Shen, Qi.; Jiang, J. H.; Shen, G. L.; Yu, R. Q. QSAR
analysis of cyclooxygenase inhibitor using particle swarm optimiza-
tion and multiple linear regression. J. Pharm. Biomed. Anal. 2004,
35, 679-687.
(3) Selvam, C.; Jachak, S. M.; Thilagavathi, R.; Chakraborti, A. K.
Design, synthesis, biological evaluation and molecular docking of
curcumin analogues as antioxidant, cyclooxygenase inhibitory and
anti-inflammatory agents. Bioorg. Med. Chem. Lett. 2005, 15, 1793-
1797.
(4) Chen, Q. H.; Rao, P. N. P.; Knaus, E. E. Design, synthesis, and
biological evaluation of N-acetyl-2-carboxybenzenesulfonamides: a
novel class of cyclooxygenase-2 (COX-2) inhibitors. Bioorg. Med.
Chem. 2005, 13, 2459-2468.

7780 Journal of Medicinal Chemistry, 2006, Vol. 49, No. 26
Perissutti et al.
(5) Rao, P. N. P.; Uddin, Md. J.; Knaus, E. E. Design, synthesis, and
structure-activity relationship studies of 3,4,6-triphenylpyran-2-ones
as selective cyclooxygenase-2 inhibitors. J. Med. Chem. 2004, 47,
3972-3990.
(6) Grosser, T.; Fries, S.; FitzGerald, A. G. Biological basis for the
cardiovascular consequences of COX 2 inhibition: therapeutic
challenges and opportunities. J. Clin. InVest. 2006, 116, 4-15.
(7) Cullen, L.; Kelly, L.; Connor, S. O.; Fitzgerald, D. J. Selective
cyclooxygenase-2 inhibition by nimesulide in man. J. Pharmacol.
Exp. Ther. 1998, 287 (2), 578-582.
(13) Brenneis, C.; Maier, T. J.; Schmidt, R.; Hofacker, A.; Zulauf, L.;
Jakobsson, P. J.; Scholich, K.; Geisslinger, G. Inhibition of prostag-
landin E2 synthesis by SC-560 is independent of cyclooxygenase 1
inhibition. FASEB J. 2006, 20, 1352-1360.
(14) Martinez, R. V.; Reval, M.; Campos, M. D.; Terron, J. A.;
Dominguez, R.; Lopez-Munoz, F. J. Involvement of peripheral
cyclooxygenase-1 and cyclooxygenase-2 in inflammatory pain. J.
Pharm. Pharmacol. 2002, 54, 405-412.
(15) Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.; Mcdonald, J. J.;
Stegeman, R. A.; Pak, J. Y.; Gildehaus, D.; Miyashiro, J. M.; Penning,
T. D.; Seibert, K.; Isakson, P. C.; Stallings, W. C. Structural basis
for selective inhibition of cyclooxygenase-2 by anti-inflammatory
agents. Nature 1996, 384, 644-648.
(8) Futaki, N.; Takahashi, S.; Kitagawa, T.; Yamakawa, Y.; Tanaka, M.;
Higuchi, S. Selective inhibition of cyclooxygenase-2 by NS-398 in
endotoxin shock rats in vivo. Inflammation Res. 1997, 46 (12), 496-
502.
(9) Kirchner, T.; Argentieri, D. C.; Barbone, A. G.; Singer, M.; Steber,
M.; Ansell, J.; Beers, S. A.; Wachter, M. P.; Wu, W.; Malloy, E.;
Stewart, A.; Ritchie, D. M. Evaluation of the anti-inflammatory
activity of a dual cyclooxygenase 2 selective/5-lipoxygenase inhibitor,
RWJ 63556, in a canine model of inflammation. J. Pharmacol. Exp.
Ther. 1997, 282 (2), 1094-1101.
(16) Zingarelli, B.; Southan, G. J.; Gilad, E.; O’Connor, M.; Salzman, A.
L.; Szabo`, C. The inhibitory effects of mercaptoalkylguanidines on
cyclooxygenase activity. Br. J. Pharmacol. 1997, 120, 357-366.
(17) Sautebin, L.; Lanaro, A.; Rombola`, L.; Ialenti, A.; Sala, A.; Di Rosa,
M. Cyclooxygenase-2-dependent generation of 8-epiprostaglandin p2a
by lipopolysaccharide-activated J774 macrophages. Inflammation Res.
1999, 48, 503-508.
(18) Davies, N. M.; Roseth, A. G.; Aplleyard, C. B.; McKnight, W.; Del
Soldato, P.; Calignano, A.; Cirino, G.; Wallace, J. L. NO-naproxen
vs. naproxen: ulcerogenic, analgesic and anti-inflammatory effects.
Aliment. Pharmacol. Ther. 1997, 11, 69-79.
(10) Borch, R. F.; Berstein, M. D.; Durst, H. D. Cyanohydridoborate, anion
as a selective reducing agent. J. Am. Chem. Soc. 1971, 93, 2897-
2904.
(11) Westheimer, F. H.; Taguchi, K. Catalysis by molecular sieves in the
preparation of ketimines and enamines. J. Org. Chem. 1971, 36,
1570-1572.
(12) Wallace, J. L.; Mcnight, W.; Reuter, B. K. NSAID-induced gastric
damage in rats: requirement for inhibition of both cyclooxygenase
1 and 2. Gastroenterology 2000, 119 (3), 706-714.
JM0608199