2,3-Diarylimidazo[1,2-a]pyridines as potential inhibitors of UV-induced keratinocytes apoptosis: Synthesis, pharmacological properties and interactions with model membranes and oligonucleotides by NMR

Article abstract of DOI:10.1016/j.ejps.2004.10.009

Four 2,3-diarylimidazo[1,2-a]pyridines (I, 1a-c) were synthesized as inhibitors of UV-induced apoptosis and showed quite different properties. First, only the pyridinyl derivative I showed protection in molt cells. From the supposed intracellular target,

Full text of DOI:10.1016/j.ejps.2004.10.009

European Journal of Pharmaceutical Sciences 24 (2005) 219–227
2,3-Diarylimidazo[1,2-a]pyridines as potential inhibitors of UV-induced
keratinocytes apoptosis: synthesis, pharmacological properties and
interactions with model membranes and oligonucleotides by NMR
a
b
b
b
´
´
Cecile Enguehard-Gueiffier , Florence Fauvelle , Jean-Claude Debouzy , Andre Peinnequin ,
Isabelle Therya, Vincent Dabouisb, Alain Gueiffiera,∗
a
UPRES-EA 3247, Laboratoire de Chimie The´rapeutique, Faculte´ de Pharmacie, 31 Avenue Monge, 37200 Tours, France
b
CRSSA/BCM, 24 Avenue des maquis du Gre´sivaudan, 38702 La Tronche Cedex, France
Received 26 May 2004; received in revised form 15 October 2004; accepted 20 October 2004
Available online 13 December 2004
Abstract
Four 2,3-diarylimidazo[1,2-a]pyridines (I, 1a–c) were synthesized as inhibitors of UV-induced apoptosis and showed quite different
properties. First, only the pyridinyl derivative I showed protection in molt cells. From the supposed intracellular target, phospholipid membrane
models were studied by 1H, 2H and 31P NMR spectroscopy. All these molecules can incorporate the membrane bilayer of small unilamellar
vesicles of lecithin (SUV). However, I is clearly closed to the external polar head of the lipids, and is relatively mobile in the layer. Conversely,
the other molecules are strongly immobilized in the deep part of the external layer. 31P solid-state NMR spectra recorded on phospholipid
dispersions (multilayers vesicles (MLV)) completely excluded any detergent effect or any modification of temperature transition. The only
structural or dynamic effect observed was a homogeneous, but limited, reduction in the chemical shift anisotropy in the presence of I, in
2
agreement with its superficial location. H NMR experiment performed on the same model using perdeuterated phospholipids showed no
significant fluidity reduction at the level of terminal CD3 groups in the presence of 1a–c, according to their deep location. Finaly, their
interactions with synthetic oligonucleotide, d(CGATCG)2 was studied showing non specific interactions of 1a on the external GC pair, while
no interaction was observed with the other derivatives.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Imidazo[1,2-a]pyridine; Apoptosis; NMR; Model membrane; Oligonucleotide
1. Introduction
ported (Astles et al., 1996) that diarylimidazoles exemplified
by III were potent inhibitors of Acyl-CoA:cholesterol
O-acyltransferase (ACAT), while the pyridyl analog IV was
a potent inhibitor of p38 kinase demonstrating that a nitrogen
atom in the position four of one aryl group was a requisite
for the p38 inhibition properties (McLay et al., 2001).
Moreover, it is now established that TNF-␣ plays a role in
UV-induced keratinocytes apoptosis (Zhuang et al., 2000).
In order to demonstrate that I may interact with the apopto-
sis signaling pathway possibly through its inhibition of TNF-
␣ production, the study of the protection against UV-induced
apoptosis in human cell lines in the presence of I was then
undertaken. This protection was estimated by comparing I
with three nitrogen deprived derivatives 1a–c. The pyridinyl
In 1991, the 2-(4-fluorophenyl)-3-(pyridin-4-yl)imidazo-
[1,2-a]pyridine I (Fig. 1) was shown to inhibit the interleukin
1 (IL-1) and tumor necrosis factor-␣ (TNF-␣) productions
(Badger et al., 1991). From the multiple IL-1 receptor
signaling pathways (Auron, 1998), a possible underlined
mechanism for I is the inhibition of p38 kinases. Indeed,
this compound showed a strong structural analogy with
pyridyl-imidazole p38 inhibitors, exemplified by SB203580
II (Wilson et al., 1997). Furthermore, it was recently re-
∗
Corresponding author. Tel.: +33 2 47 36 71 38; fax: +33 2 47 36 72 88.
E-mailaddress: gueiffier@univ-tours.fr (A. Gueiffier).
0928-0987/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejps.2004.10.009

220
C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
Fig. 1. Structures of compounds I, II, III and IV.
derivative I was obtained as reported (Badger et al., 1991).
The 3-phenyl compounds 1a–c were synthesized according
to the pathway recently published by our group (Enguehard
et al., 2000) (Scheme 1). Furthermore, as the target is intra-
cellular, it was of interest to evaluate the interactions with the
cellular barriers, i.e. the membranes. Phospholipidic mem-
branes model are a convenient tool for such studies, espe-
cially in combination with NMR technics. Finaly in order to
avoid a possible mechanistic pathway by direct interaction
of the compound with DNA, the interaction of I and 1a–c
with an oligonucleotide was also investigated. Hence, three
different aspects of the pharmacological properties of these
molecules were studied in terms of:
2001) that the cell nucleus, more precisely the DNA, ap-
pears as possible molecular target. After the attribution of
the 1H NMR resonances, their interactions with synthetic
oligonucleotides as model of the supposed molecular tar-
get – nucleic acids – were then studied using the same
methods in water.
2. Materials and methods
2.1. Chemistry
The melting points were determined on a Kofler hot-
stage and are uncorrected. The NMR spectra were recorded
on spectrometer Bruker DPX 200 or AM 400 WB instru-
ments and chemical shifts are expressed in ppm relative
to tetramethylsilane (TMS) from residual CHCl3 at δ 7.3
(1H) and central resonance of CDCl3 at δ 77.1 (13C). Pos-
sible inversion of two values in the NMR spectra is ex-
pressed by an asterisk. Known compounds were obtained
according to described procedures: 2-(4-fluorophenyl)-3-
(pyridin-4-yl)imidazo[1,2-a]pyridine I (Badger et al., 1991),
tetrakis(triphenylphosphine)palladium(0) (Coulson, 1971),
2-(4-fluorophenyl)imidazo[1,2-a]pyridine 2a (Tomoda et al.,
1999), 2-(4-fluorophenyl)-3-iodoimidazo[1,2-a]pyridine 3a
• In vitro protection against UV-induced apoptosis, using
Molt human cell lines as described-above (Chen et al.,
1996).
• Interactions of these drugs with membranes: their interac-
tions with synthetic small unilamellar vesicles (SUV) and
their location in the lipid bilayer were studied by 1H NMR,
while the membrane structure and dynamics in the pres-
ence of these compounds were investigated by 31P- and 2H
NMR experiments.
• Possible interactions of these drugs with DNA from clas-
sically admitted biological considerations (Debouzy et al.,
Scheme 1.

C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
221
and 2-(4-fluorophenyl)-3-phenyl-imidazo[1,2-a]pyridine 1a
2.1.2.1. 3-Iodo-2-(4-trifluoromethylphenyl)imidazo[1,2-
(Enguehard et al., 2000).
a]pyridine (3b). Eighty-seven percent yield; mp = 186 ◦C;
1H NMR (CDCl3, 200 MHz) δ: 8.30 (m, 3H, H-5, CF3-
Ph-2,6), 7.78 (d, 1H, J = 8.2 Hz, CF3-Ph-3,5), 7.68 (d, 1H,
J = 9.2 Hz, H-8), 7.34 (m, 1H, H-7), 7.01 (t, 1H, J = 6.8 Hz,
H-6); 13C NMR (CDCl3, 50 MHz) δ: 148.64 (C-2*), 146.93
(C-8a*), 137.39 (CF3-Ph-1), 130.60 (J = 32 Hz, CF3-Ph-4),
129.21 (CF3-Ph-2,6), 127.11 (C-5), 126.63 (C-7), 125.73
(J = 3.5 Hz, CF3-Ph-3,5), 124.62 (J = 270 Hz, CF3), 118.25
(C-8), 114.09 (C-6), 60.67 (C-3).
2.1.1. Generalprocedure for cyclization
A solution of 2-aminopyridine (0.1 mol) and the suit-
able phenacyl bromide (0.12 mol) was refluxed in ethanol
(250 mL) for 4 h. After cooling, the solution was concentrated
in vacuo and the residue diluted in water. The solution was
made basic with Na2CO3 and extracted with CH2Cl2. The
dried organic layers were evaporated to dryness. The residue
was chromatographed on silica gel, eluting with CH2Cl2.
2.1.2.2. 2-(3,4-Difluorophenyl)-3-iodoimidazo[1,2-
a]pyridine (3c). Fifty-five percent yield; mp = 173 ◦C;
1H NMR (CDCl3, 200 MHz) ␦: 8.26 (dt, 1H, J = 6.8–1.1 Hz,
H-5), 7.97 (ddd, 1H, J = 11.7–7.8–2.2 Hz, di-F-Ph-2),
7.87 (m, 1H, di-F-Ph-6), 7.65 (dt, 1H, J = 9–1.1 Hz,
H-8), 7.37–7.23 (m, 2H, H-7, di-F-Ph-5), 6.99 (td, 1H,
J = 6.8–1.1 Hz, H-6); 13C NMR (CDCl3, 50 MHz) ␦: 150.81
(J = 252 Hz, di-F-Ph-3*), 150.62 (J = 248.5 Hz, di-F-Ph-4*),
148.53 (C-2*), 146.38 (C-8a*), 131.10 (J = 6.5–4 Hz, di-
F-Ph-1), 127.01 (C-5), 126.45 (C-7), 125.03 (J = 6.5–4 Hz,
di-F-Ph-6), 118.07 (C-8), 117.95 (J = 11.5–1.5 Hz, di-F-Ph-
2), 117.60 (J = 12–1 Hz, di-F-Ph-5), 113.93 (C-6), 59.93
(C-3).
2.1.1.1. 2-(4-Fluorophenyl)imidazo[1,2-a]pyridine (2a).
Sixty-five percent yield; mp = 164 ◦C (Lit. 169 ◦C); 1H
NMR (CDCl3, 200 MHz) δ: 8.14 (dt, 1H, J = 6.8–1.3 Hz,
H-5), 7.96 (dd, 2H, J = 9–5.4 Hz, F-Ph-2,6), 7.84 (s, 1H,
H-3), 7.65 (dt, 1H, J = 9.1–1.2 Hz, H-8), 7.21 (ddd, 1H,
J = 9.1–6.8–1.3 Hz, H-7), 7.16 (t, 2H, J = 9 Hz, F-Ph-3,5),
6.81 (td, 1H, J = 6.8–1.2 Hz, H-6); 13C NMR (CDCl3,
50 MHz) δ: 163.1 (J = 246.5 Hz, F-Ph-4), 146.1 (C-8a*),
145.3 (C-2*), 130.4 (J = 3.1 Hz, F-Ph-1), 128.1 (J = 8.1 Hz,
F-Ph-2,6), 126.0 (C-7), 125.2 (C-5), 117.8 (C-8), 116.0
(J = 21.6 Hz, F-Ph-3,5), 112.9 (C-6), 108.2 (C-3).
2.1.1.2. 2-(4-Trifluoromethylphenyl)imidazo[1,2-
a]pyridine (2b). Sixty-three percent yield; mp > 260 ◦C;
1H NMR (CDCl3, 200 MHz) δ: 8.17 (d, 1H, J = 6.8 Hz,
H-5), 8.09 (d, 2H, J = 8.2 Hz, CF3-Ph-2,6), 7.96 (s, 1H,
H-3), 7.70 (m, 3H, CF3-Ph-3,5, H-8), 7.24 (ddd, 1H,
J = 9.2–6.8–1.3 Hz, H-7), 6.85 (td, 1H, J = 6.8–1.3 Hz, H-6);
13C NMR (CDCl3, 50 MHz) ␦: 146.22 (C-8a*), 144.56
(C-2*), 137.63 (CF3-Ph-1), 130.08 (J = 32 Hz, CF3-Ph-4),
126.51 (CF3-Ph-2,6), 126.12 (J = 4 Hz, CF3-Ph-3,5), 126.03
(C-5), 125.66 (C-7), 124.68 (J = 272 Hz, CF3), 118.10 (C-8),
113.25 (C-6), 109.44 (C-3).
2.1.3. Generalprocedure for Suzuki Reaction
In a three necked round bottom flask was introduced
under nitrogen the 3-iodoimidazo[1,2-a]pyridine derivative
(300 mg) in DME (6.5 mL). We then added under stirring
tetrakis(triphenylphosphine)palladium (0) (5 mol%), ben-
zeneboronic acid (1.1 eq) and sodium hydroxyde (2 eq in
3.2 mL of H2O). The resulting mixture was warmed at 75 ◦C
and after cooling, diluted in H2O. The aqueous layer was
extracted with CH2Cl2. The combined organic layers were
dried over CaCl2, filtered and evaporated to dryness. The
residue was chromatographed twice on silica gel, eluting with
CH2Cl2.
2.1.1.3. 2-(3,4-Difluorophenyl)imidazo[1,2-a]pyridine (2c).
Forty-five percent yield; mp = 143 ◦C; 1H NMR (CDCl3,
400 MHz) ␦: 8.16 (d, 1H, J = 6.6 Hz, H-5), 7.84 (s,
1H, H-3), 7.80 (d, 1H, J = 8 Hz, diF-Ph-2), 7.68 (m,
1H, H-8), 7.67 (d, 1H, J = 8.6 Hz, diF-Ph-6), 7.26 (dd,
1H, J = 9–6.6 Hz, H-7), 7.24 (t, 1H, J = 8.6 Hz, diF-Ph-
5), 6.85 (t, 1H, J = 6.6 Hz, H-6); 13C NMR (CDCl3,
50 MHz) ␦: 150.92 (J = 246.1–25.2 Hz, diF-Ph-3*), 150.68
(J = 246.6–25.2 Hz, diF-Ph-4*), 146.11 (C-8a*), 144.20 (C-
2*), 131.46 (J = 6–3.5 Hz, diF-Ph-1), 126.06 (C-5), 125.52
(C-7), 122.34 (J = 6–3.5 Hz, diF-Ph-6), 118.00 (C-8), 117.90
(J = 17.6 Hz, diF-Ph-5), 115.37 (J = 18.6 Hz, diF-Ph-2),
113.12 (C-6), 108.59 (C-3).
2.1.3.1. 3-Phenyl-2-(4-trifluoromethylphenyl)imidazo[1,2-
a]pyridine (1b). The mixture was warmed at 75 ◦C for 8 h
(56% yield); mp = 111 ◦C; 1H NMR (CDCl3, 200 MHz)
δ: 7.99 (dt, 1H, J = 7–1.2 Hz, H-5), 7.82 (d, 2H, J = 9 Hz,
CF3-Ph-2,6), 7.74 (dt, 1H, J = 9–1.2 Hz, H-8), 7.64–7.47
(m, 7H, Ph’, CF3-Ph-3,5), 7.28 (ddd, 1H, J = 9–7–1.2 Hz,
H-7), 6.81 (td, 1H, J = 7–1.2 Hz, H-6); 13C NMR (CDCl3,
100 MHz) δ: 145.34 (C-8a), 141.20 (C-2), 138.18 (CF3-
Ph-1), 131.10 (Ph’-2,6), 131.09 (Ph’-1), 130.18 (Ph’-3,5),
129.73 (Ph’-4), 129.58 (J = 30.3 Hz, CF3-Ph-4), 128.47
(CF3-Ph-2,6), 125.62 (C-7, CF3-Ph-3,5), 124.18 (J = 270 Hz,
CF3), 123.86 (C-5), 122.45 (C-3), 118.10 (C-8), 113.07
(C-6).
2.1.2. Generalprocedure for iodination
To a solution of imidazo[1,2-a]pyridine (4 mmol) in dry
acetonitrile (5 mL) was added N-iodosuccinimide (0.95 g, 4
mmol). The reaction mixture was stirred at RT for 30 min
and the resulting white solid was filtered off, dried and used
without further purification.
2.1.3.2. 2-(3,4-Difluorophenyl)-3-phenylimidazo[1,2-
a]pyridine (1c). The resulting mixture was warmed at 75 ◦C
for 1 h (14% yield, not optimized); mp = 102 ◦C; H NMR
1
(CDCl3, 200 MHz) δ: 7.96 (dt, 1H, J = 6.8–1.2 Hz, H-5), 7.70

222
C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
(dt, 1H, J = 9–1.2 Hz, H-8), 7.61–7.55 (m, 4H, Ph-3ꢀ,4ꢀ,5ꢀ,
diF-Ph-2), 7.50–7.37 (m, 3H, Ph-2ꢀ,6ꢀ, diF-Ph-6), 7.26
(ddd, 1H, J = 9–6.8–1.2 Hz, H-7), 7.08 (m, 1H, diF-Ph-5),
6.79 (td, 1H, J = 6.8–1.2 Hz, H-6); 13C NMR (CDCl3,
100 MHz) δ: 150.54 (J = 245.2–30.7 Hz, diF-Ph-3*), 150.41
(J = 245.3–33 Hz, diF-Ph-4*), 146.03 (C-8a*), 140.51
(C-2*), 131.09 (Ph’-2,6), 131.08 (Ph’-1) 130.22 (Ph’-3,5),
129.79 (Ph’-4), 129.57 (diF-Ph-1), 125.71 (C-7), 124.41
(diF-Ph-6), 123.83 (C-5), 121.66 (C-3), 117.88 (C-8), 117.51
(J = 17 Hz, diF-Ph-5), 117.24 (J = 18.6 Hz, diF-Ph-2), 113.09
(C-6).
foetal bovine serum (Sigma), l-glutamine (2 mM), penicillin
(100 units/mL), and streptomycin (1 mg/mL), at 37 ◦C in a
5% CO2 humidified atmosphere. Cell numbers were deter-
mined with a hemocytometer. Cell viability was assessed us-
ing the Trypan dye exclusion method. Drug experiments were
led in 2% DMSO.
2.3.2. UV-induced apoptosis and drug toxicity assays
Actively growing cells were plated 18 h before irradiation
in matched pairs (sham and irradiated), into 96-well culture
plates, at an initial density of 1.5 × 105 cells/mL. Apoptosis
was induced by 254 nm UV-irradiation using a Stratalinker
XL1000 oven (StratageneTM) (Chen et al., 1996). Exposures
at 100 J/m2 were performed in order to decrease cell prolif-
eration to approximately 70–80% of non irradiated control
cells. The alamarBlue Assay (Interchim®, France) was used
to assess cellular proliferation 16 h after UV-irradiation as de-
scribed previously (Peinnequin et al., 2000). This method in-
corporates an oxidation–reduction indicator that fluoresces in
response to chemical reduction of the medium resulting from
cell growth. Measurements were performed in 96-well cul-
ture platesusinga Cytofluor 2350fluorometer(Millipore®) at
excitation 530 nm and emission 590 nm. The reproductibility
of the assay defined as the ratio of standard deviation to mean
is 2.8% (data not shown). Statistical significance is assessed
using student t-test with a 5% significance level. Drug toxi-
city assays were led in the same condition as UV-irradiated
cells using control and treated matched pairs.
2.2. NMR experiments
2.2.1. Generalpoints
NMR resonances in membranes and in complexed forms
were identified by spectra comparisons. l-α-Phosphatidyl
choline from egg yolk (EPC) and synthetic dimyristoylphos-
phatidylchoilne (DMPC) were obtained from Sigma, La
`
Verpillere, France. DMPC deuterated on both chains was
purchased from Interchim, Montluc¸on, France. The self-
complementary hexamer d(CGATCG)2 was purchased from
Institut Pasteur, Paris and was used without further purifica-
tion dissolved in a buffer containing 160 mM KCl, 0.15 mM
EDTA, 20 mM sodium phosphate, pH 7 in either pure
D2O or D2O/H2O 10/90 for the observation of exchange-
able protons. In pure D2O experiment, the solution was
lyophilized, resuspendedin99.98%D2Oanddegassedbefore
experiment.
2.3.3. NMR experiments
1H NMR experiments were recorded on Bruker AM-400
or AVANCE-DMX-400 spectrometers and referenced rela-
tive to internal tetramethylsilane. Standard COSY NMR ex-
periments were used for protons assignement (Plateau and
Gueron, 1982). One-dimensional spectra in H2O/D2O, 90/10
and D2O were recorded at 277 and 296 K, respectively, us-
ing a linewidth of 4000 Hz and 32,000 complex points. A
«jump-and-return» sequence (Sklenar and Bax, 1987) with-
out water presaturation was used to observe exchangeable
protons in H2O/D2O. NOESY spectra were obtained with
the TPPI method (Kumar et al., 1980; Marion and Wuethrich,
1983) and 2048 points in both dimensions. Typical spectral
width was 3800 Hz. Sinebell shifted filters were also used
for Fourier transform in both dimensions. Mixing time were
ranged from 70 ms (for nOe intensities estimation) up to 0.7 s.
Double Quantum filter COSY (Kumar et al., 1980) and HO-
HAHA (Bax and Davis, 1985; Clore and Gronenborn, 1985)
experiments (spin lock of 2.5 kHz) were also performed to
confirm proton attribution as described in other works. 31P
NMR spectra were recorded on the same samples used for
proton recording using standard acquisition and composite
pulseprotondecouplingandapresaturationofthesolventres-
onances was used for all the experiments in aqueous medium.
2.2.2. Vesicle preparation
2.2.2.1. Small unilamellar vesicles (SUV). EPC in its chlo-
roformic solution (100 mg/mL) was evaporated to a film and
resuspended in pure D2O to a final lipid concentration of
10 mM. SUV were formed by 1 h bath sonication. For drug
containing SUV, the respective chloroformic solutions were
mixed before lyophylization to a molar ratio of imidazopy-
ridines/phospholipids = 1:10.
2.2.2.2. Multilayers vesicles (multibilayers, MLV). DMPC
liposomes for 31P and 2H NMR experiments were prepared
following Roux (Roux, 1987). The suspensions were de-
gassed under nitrogen gas then introduced into NMR tubes
and sealed. The liposomes were formed by fast freezing and
thawing cycles. The final lipid concentration was 50 mM and
drug/lipid molar ratio were 1:10. For 2H NMR experiments,
lyophilized DMPC containing 20% perdeuterated phospho-
lipids was resuspended in deuterium-depleted water.
2.3. Biological studies
2.3.1. Cell cultures
31P and H NMR experiments were performed at 162
2
Human T-leukemia cell lines Molt 4 were obtained
from European Collection of Cell Cultures. Both were
routinely cultured in RPMI 1640 supplemented with 10%
and 61 MHz, respectively. Phosphorus spectra were recorded
using a dipolar echo sequence (π/2–t–π–t) with a t value

C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
223
Fig. 2. Effects of compound I, 1a–c on cell survival in human T cell line Molt4. Cell proliferation was measured 16 h after addition of the evaluated compound
(2% DMSO) using alamarBlue assay (Interchim). Results are expressed as the percentage of alamarBlue fluorescence of control cells cultured in 2% DMSO.
Each value is the average of 10 measures. The error margins represent S.E.M.
of 12 ␮s and a broadband two level proton decoupling
(Gorenstein, 1984). Phosphoric acid was used as external ref-
erence. Deuterium spectra were recorded using a quadrupolar
echo sequence (π/2–t–π/2–t) with a t value of 20 ␮s. The free
induction decay was shifted by fractions of the dwelling time
to ensure that its effective time for the Fourier transform cor-
responds to the top of the echo (Davis et al., 1976).
until 50 ␮M while compounds 1a and 1c led to a slight effect
on cell survival from 25 ␮M.
3.1.2. Drug effects on UV-induced apoptosis
Immediately after irradiation UV Molt cells were exposed
to compounds I, 1a–c in the 0–25 ␮M range. Cell prolifer-
ation was assessed 16 h after treatment (Fig. 3). Compound
I led to a slight but statistically significant increased of cell
survival after UV-irradiation in the 15–25 ␮M range (t = 2.16
and t = 3.23; p < 0.05 and p < 0.01). There was no difference
in cell survival between treated and control cells up to 10 ␮M.
Compound 1c did not interact with survival after UV-induced
apoptosis up to 25 ␮M. Conversely, compound 1a and 1b
markedly decreased cell survival after UV-induced apoptosis
in the 15–25 ␮M range.
3. Results and discussion
3.1. UV-induced apoptosis
3.1.1. Drug toxicity in Molt cell line
Actively growing Molt cells were plated with compound I,
1a–c in up to 100 ␮M and cultured 16 h before cell prolifera-
tion assay. As shown in Fig. 2, the compounds exhibited quite
different interactions with cultured cells. Compound 1b led to
dramatic cellular damage from 25 ␮M. Microscopic diagno-
sis is compatible with a disruption of the plasma membrane.
Compound I had no significant effect on cell proliferation
3.2. Interactions with membranes
3.2.1. Drug localization in small unilamellar vesicles
1
Fig. 4 shows the aromatic part of H-NMR spectra for
the different drugs I (left) and 1c (right), 1a and 1b giving
Fig. 3. Effects of compound I, 1a–c on cell survival after UV-irradiation in human T cell lines Molt4. Apoptosis was induced in actively growing cells with
a 100 J/m2 UV-irradiation in a strata linker (stratagene), before addition of compound I, 1a–c in 2% DMSO. Then, cell proliferation was measured 16 h after
addition of compounds using alamarBlue assay. Results are expressed as the percentage of alamarBlue fluorescence of untreated irradiated control cell. The
error margins represent S.E.M. Statistical significance; /p > 0.05; *p < 0.05; **p < 0.01.

224
C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
Fig. 4. Aromatic part of 1H NMR spectra recorded on I (left column) and
1c (right column), (A) pure in CDCl3 solution, 2 mM, 297 K; (B) in the
presence of SUV of lecithin (10 mM, D2O), drug/lipid ratio of 1/10, M/M;
(C) same as in (B), after external addition of 4 ␮L MnCl2, 0.1 M.
quite similar results as 1c. As expected from the chemical
structure, all these compounds present a very good solubil-
ity in CDCl3 giving typical linewidthes narrower than 1 Hz.
By the way of contrast, all these compounds are completely
insoluble in water, and thus give no detectable NMR sig-
nal. Conversely, SUV that are small membrane systems give
relatively resolved lines in 1H NMR (see the projections on
Fig. 5). Such a system, thus, constitutes a good tool to observe
drug–membrane interactions, since both lipid and drug reso-
nances can be detected on the same spectrum (Debouzy et al.,
1989; Bossanne et al., 1994). Hence, by comparison with 1H
NMR spectra observed on SUV, the lines attributed to lipid
protons (the 0–5.5 ppm region) are quite superimposable with
those recorded in the presence of the I, 1a–c, indicating the
complete absence of detergent effect on the membranes (not
shown; molar ratio up to 1/1 M/M drug/lipid). Conversely,
NMR lines are detected in the aromatic region (6.5–9 ppm).
As those molecules are insoluble in water, these resonances
are related to molecules in hydrophobic environement, i.e.
the membrane. However, the typical linewithes observed are
clearly different: I exhibits still relatively resolved lines of
less that 20 Hz, consistent with a good mobility and solubil-
ity in the membrane (Fig. 4, left trace, B). Conversely, the
other drugs show poorly resolved resonances of 30–80 Hz
width, revelative of relatively immobilized structures.
Fig. 5. nOeSY experiments in SUV (mixing time of 250 ms) under the same
conditions as in Fig. 4B recorded in the presence of I (top) and 1b (bot-
tom). The projection of top trace was expanded to show phospholipid res-
onances: CH3 (0.8 ppm); CH2 CH2CH2 (1.3 ppm); CH2CH2CO (1.6 ppm);
CH2CH=(2 ppm); CH2CO (2.4 ppm); =CHCH2CH= (2.8 ppm); ␥-choline
(3.3 ppm, indicated by a vertical bar); CH=CH (5.5 ppm).
I resonances (top trace, Fig. 4), while the paramagnetic effect
is relatively slight for the other drugs. This feature indicates
that I is both mobile and located close to the superficial part
of the layer, while other drugs are more deeply embedded in
the membrane that is less sensible to paramagnetic effects.
Definitive argument for drug location was given by record-
ing NOESY experiments on the same systems (mixing time
of 250 ms). NOE connectivities are clearly detected be-
tween I and the superficial choline ␥ protons (N(CH3)3:
δ = 3.23 ppm), related with the location of the drug at this
superficial level (Fig. 5). Other connectivities are also found
with methylenic groups (1.3 ppm), but cannot be interpre-
tated since these groups are themselves located all along the
acyl chain of phospholipids.
A supplementary argument is given by the addition of
4 ␮L of MnCl2 (paramagnetic molecule) that broadens the
resonances of geometrically accessible protons (e.g. the po-
lar cholin headgroup) of the external layer of the vesicles
(Elhakmaoui, 1992). Under these conditions, the spectra
recorded on the same systems show a dramatic broadening of
Conversely, the other drugs give correlations with the pro-
ton of the terminal methyl group of the acyl chains (CH3,
0.9 ppm) and the methylenic groups, confirming a deep loca-
tion of these drugs in the layer.

C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
225
tion around 297 K, and CSA values in the 294–297 K range
very close to those of pure DMPC. It clearly appears that
the temperature of transition gel–liquid crystal (297 K) is not
modified in the presence of the drugs. Conversely, I induces
a significant reduction of the CSA on all the temperature do-
main explored (drug/lipid 1/10, w/w), revealing a fluidizing
effect of the superficial part of the layer, where phospho-
rus nuclei are located, all in agreement with the location of
I. Moreover, the spectra of I containing systems are clearly
those of phospholipids over transition temperature, on all the
295–313 K temperature range. Besides, no difference was
2
found for the other molecules. H NMR experiments (not
shown) gave no supplementary informations except a lim-
ited increase in the quadrupolar splitting of the terminal CD3
groups of perdeuterated DMPC in the presence of the drugs
except I.
At this point, all the observations are consistant with a dif-
ferent location in the membrane of I in one hand, and of the
three other derivatives on the other hand. Hence, I is located
close to the surface of the layer, where an increase of the
membrane fluidity is observed as consequence. Conversely,
the other isosteric derivatives that are more hydrophobic
(log P = 3.515 (I), against log P = 4.948 (1a); log P = 5.690
(1b); log P = 5.022(1c)) are logically found in the depth of
the membrane at the acyl chain level where only non signifi-
cant structural modifications can be evoqued. The following
step to undertake was then to observe the interactions of the
compounds with synthetic DNA d(CGATCG)2, as described
in the next section.
Fig. 6. Plot of the chemical shift anisotropy (ppm) as a function of tempera-
ture, measured on 31P NMR spectra of pure DMPC(ꢀ), and in the presence
of I (ꢁ); 1a or 1b (᭹) and 1c (ꢂ). A typical spectrum recorded at 298 K on
pure DMPC is shown in insert.
Such differences in the membrane location are also con-
sistent with the less markedly hydrophobic properties of I.
3.2.2. Influence of heterocycles presence on membrane
structure: 31P, 2H NMR experiments in phospholipid
dispersions
3.3. Interactions with oligonucleotide d(CGATCG)2
In the next section, we investigate the consequences of the
presence of I, 1a–c on the bilayer structure by using phospho-
lipid dispersions (multibilayers, MLV) as membrane models
and 31P and 2H NMR spectroscopies (Seelig, 1977).
TheinsertofFig. 6showsatypicalaxiallysymmetricpow-
derphosphorusspectrumcorrespondingtotypicalofstandard
bilayer structures of DMPC over phase transition (Debouzy
et al., 1989). The temperature dependance of 31P chemical
shift anisotropy (CSA, figured by an horizontal bar under
the insert) was measured on the phosphorus NMR spectra
of multilayers (Peinnequin et al., 2000) of pure DMPC or in
the presence of I, 1a–c (see legend of Fig. 6). As classically
admitted (Duchene, 1992), CSA value can be considered as
representative of the fluidity of the superficial part of a mem-
brane, where phosphorus nuclei are located. By comparison
with DMPC, no significant modification in the line shape was
detected on the spectra of drug-containing MLV (under the
same experimental conditions) indicating that the main bi-
layer structure is not clearly modified. Hence, no resolved
resonance indicative of detergent effect was observed at the
isotropic position. As neither the intensity ratio of σꢁ σ⊥ nor
typical linewidth of σ⊥ resonance were modified, membrane
(re)orientation or change in local motion were also refuted.
Furthermore, in the presence of 1a–c, the plots presented on
Fig. 6 are consistent with a gel/liquid crystal phase transi-
The interactions of I, 1a–c with d(CGATCG)2 were ob-
served on the different building blocks of the DNA fragment,
1
i.e. by recording H NMR spectra of the iminoprotons en-
gaged in hydrogen bonds in the duplex, the non labile protons
of the bases and the sugar groups and also by recording 31
NMR spectra of the inter residues phosphate groups.
P
3.3.1. Observation of the exchangeable protons of
d(CGATCG)2 in H2O
The downfield part of the 1H NMR spectrum of
d(CGATCG)2 recorded at 277 K is shown in Fig. 7. As pre-
viously described (Delepiere et al., 1989), the three expected
resonances of bonded imino protons were identified as fol-
lows.
The central broad resonance (13 ppm) that vanishes when
increasing temperature was attributed to external GC pairs. Ir-
radiation of the resonance located around 13.5 ppm induces a
nOe with the H2 proton of 3A (7.5 ppm) and was identified as
the proton imino A3T/T4A. The other resonance (12.5 ppm)
was attributed to G2C/C5G. As described before, intercalat-
ing drugs, groove binders and non specific outside binders
can be easily identified by the observation of the chemical
shift and linewidth variations of imino protons. Successive
spectra of d(CGATCG)2 were then recorded in the presence

226
C. Enguehard-Gueiffier et al. / European Journalof PharmaceuticalSciences 24 (2005) 219–227
of increasing amounts of the drugs, as shown from bottom
to top of Fig. 7. The only difference was observed in the
presence of 1a; however, the slight highfield chemical shift
and line broadening at very high molar ratios (drug/DNA of
1/2 M/M) cannot be related to a specific interaction such as
true intercallation or any external binding. Hence, non spe-
cific interactions or environmental factors would be more
reasonably proposed (Feigon et al., 1984).
The absence of true intercalation was ascertained by the
complete absence of effect of the drugs on the aromatic
1
(7–9 ppm) and sugar (6.5–1 ppm) H NMR resonances in
D2O, and also by the almost complete absence of difference
in the 31P NMR spectra in the presence of the drugs.
4. Conclusion
From the present study different observations can be
drawn:
• only I, was shown to protect Molt cell against UV-induced
apoptosis;
• I does not exhibit any interaction with synthetic
oligodeoxynucleotide, excluding nucleic acids as possi-
ble molecular target. The slight effect could be due to non
specific interactions or influence on environmental factors,
but is not clearly due to a specific interaction.
These results are in good agreement with the previous
observations that I is an inhibitor of the TNF-␣ production
and that TNF-␣ plays a role in the UV-induced keratinocytes
apoptosis. Then I interacts with the apoptosis signaling path-
way probably through the inhibition of TNF-␣ production.
Furthermore, the incapacity of compounds 1a–c to protect
against UV-induced apoptosis demonstrates the importance
of the nitrogen atom present in position four of one of the
two aryl groups of I. This result is evocative of the requisite
previously determined for p38 inhibition which could be the
cellular target for compound I of whom structure presents a
strong analogy with pyridyl-imidazole p38 inhibitors. More-
over, it is well known that stress-activated protein kinases
SAPK/JNK and p38 are rapidly activated and involved in
apoptosis in UV-irradiated cells, notably in human leukemia
cells.
In the drug design process, the choice of chemical modi-
fications starting from a lead can be significantly guided by
the nature of the expected target as well as the pathway that
will be used to reach this target and overcome natural barri-
ers. As the assumed target is intracellular, it was of interest
to evaluate the interactions with the cellular barriers, i.e. the
membranes.
From the membrane model studied, compound I appeared
as the most amphiphilic of the tested compounds, as shown
by its superficial location in the layer, resulting in a local in-
crease of the membrane fluidity. Both higher solubility and
superficial interaction with membranes, would allow a trans-
membrane transport towards intracellular medium. By the
Fig. 7. Imino part of 1H NMR spectra of pure d(CGATCG)2 (4 mM, 277 K,
H2O/D2O, 90/10, bottom trace) and from bottom to top in the presence of
1c, 1b, 1a and I with drug/d(CGATCG)2 molar ratios of 0.5.

C. Enguehard-Gueiffier et a.l / European Journalof PharmaceuticalSciences 24 (2005) 219–227
227
way of contrast, 1a–c are extremely hydrophobic molecules
that integrate phospholipid bilayers in their deep hydropho-
bic core. This results in an overall unability to cross this bar-
rier, and to reach the intracellular medium by the classical
«recycling pathway» of natural membranes (Berkaloff et al.,
1981).
Delepiere, M., Maroun, R., Garbay-Jaureguiberry, C., Igolen, J., Roques,
B., 1989. Proton and phosphorus-31 nuclear magnetic resonance
studies of the differences in DNA deformation induced by anti-
tumoral 7H-pyrido[4,3-c]carbazole dimers. J. Mol. Biol. 210, 211–
228.
Duchene, D., 1992. New Trends in Cyclodextrins and Derivatives. Edi-
tions de la Sant
e´, Paris.
Elhakmaoui, A., 1992. Synthe`se, e´tude structurale et passage transmem-
branaire de C-nucle´osides acycliques en se´rie imidazo[1,2-a]azines.
Ph.D. Thesis. Montpellier.
In conclusion, I in one hand, and the other molecules
(1a–c) in the other hand, follow radically different direc-
tions, demonstrating the crucial role of the nitrogen atom
of the pyridin ring present in compound I. This nitrogen was
shown to be a requisite for p38 inhibition activity, but we also
demonstrate its determinant influence in the bioavailability of
the diaryl series of compounds.
Enguehard, C., Renou, J.L., Collot, V., Hervet, M., Rault, S., Gueiffier,
A., 2000. Reactivity of 3-iodoimidazo[1,2-a]pyridines using a Suzuki-
type cross-coupling reaction. J. Org. Chem. 65, 6572–6575.
Feigon, J., Denny, W.A., Leupin, W., Kearns, D.R., 1984. Interactions of
antitumor drugs with natural DNA: 1H NMR study of binding mode
and kinetics. J. Med. Chem. 27, 450–465.
Gorenstein, D.G., 1984. Phosporus 31NMR: Principles and Applications.
Academic Press, London.
References
Kumar, A., Ernst, R.R., Wuethrich, K., 1980. A two-dimensional nuclear
Overhauser enhancement (2D NOE) experiment for the elucidation of
complete proton-proton cross-relaxation networks in biological macro-
molecules. Biochem. Biophys. Res. Commun. 95, 1–6.
Marion, D., Wuethrich, K., 1983. Application of phase sensitive two-
dimensional correlated spectroscopy (COSY) for measurements of
1H–1H spin–spin coupling constants in proteins. Biochem. Biophys.
Res. Commun. 113, 967–974.
McLay, I.M., Halley, F., Souness, J.E., McKenna, J., Benning, V., Bir-
rel, M., Burton, B., Belvisi, M., Collis, A., Constan, A., Foster, M.,
Hele, D., Jayyosi, Z., Kelley, M., Maslen, C., Miller, G., Ouldelhkim,
M.C., Page, K., Phipps, S., Pollock, K., Porter, B., Ratcliffe, A.J.,
Redford, E.J., Webber, S., Slater, B., Thybaud, V., Wilsher, N., 2001.
The discovery of RPR 200765A, a p38 MAP kinase inhibitor display-
ing a good oral anti-arthritic efficacy. Bioorg. Med. Chem. 9, 537–
554.
Peinnequin, A., Piriou, A., Mathieu, J., Dabouis, V., Sebbah, C., Malabiau,
R., Debouzy, J.C., 2000. Non-thermal effects of continuous 2.45 GHz
microwaves on Fas-induced apoptosis in human Jurkat T-cell line.
Bioelectrochemistry 51, 157–161.
Plateau, P., Gueron, M., 1982. Exchangeable proton NMR without base-
line distorsion, using new strong-pulse sequences. J. Am. Chem. Soc.
104, 7310–7311.
Astles, P.C., Ashton, M.J., Bridge, A.W., Harris, N.V., Hart, T.W., Parrott,
D.P., Porter, B., Riddel, D., Smith, C., Williams, R.J., 1996. Acyl-
CoA: cholesterol O-acyltransferase (ACAT) inhibitors. 2-(1,3-dioxan-
2-yl)-4,5-diphenyl-1H-imidazoles as potent inhibitors of ACAT. J.
Med. Chem. 39, 1423–1432.
Auron, P.E., 1998. The interleukin I receptor: ligand interactions and
signal transduction. Cytokine Growth Factor. Rev. 9, 221–237.
Badger, A.M., Bender, P.E., Esser, K.M., Griswold, D.E., Nabil, H., Lee,
J.C., Votta, B.J., Simon, P.L., 1991. Inhibition of interleukin-1 and
tumor necrosis factor production by monocytes and/or macrophages.
WO 91/00092 [CAN 115: 64756].
Bax, A., Davis, D.G., 1985. MLEV-17 based 2D-homonuclear magneti-
zation transfer spectroscopy. J. Magn. Res. 65, 355–360.
Berkaloff, A., Bourguet, J., Favard, P., Lacroix, J.-C., 1981. Biologie and
physiologie. In: Membranes, vol. 1. Hermann Ed, Paris.
Bossanne, P., Heintz, O., Fauvelle, F., Baron, D., Mariotte, A.M., De-
bouzy, J.C., 1994. Interaction of a flavonoid extracted fro the plane-
tree buds with membranes. Boll. Chim. Pharm. 133, 401–406.
Chen, Y.R., Wang, X., Templeton, D., Davis, R.J., Tan, T.H., 1996. The
role of c-Jun N-terminal kinase (JNK) in apoptosis induced by ultra-
violet C and gamma radiation. Duration of JNK activation may deter-
mine cell death and proliferation. J. Biol. Chem. 271, 31929–31936.
Clore, G.M., Gronenborn, A.M., 1985. Probing the three-dimensional str-
cutures of DNA and RNA oligonucleotides in solution by nuclear
Overhauser enhancement measurements. FEBS Lett. 179, 187–198.
Coulson, D.R., 1971. Tetrakis(triphenylphosphine)palladium(0). Inorg.
Chem. 13, 121–124.
Davis, J.H., Jeffrey, K.R., Bloom, M., Valic, M.I., Higgs, T., 1976.
Quadrupolar echo deuteron magnetic resonance spectroscopy in or-
dered hydrocarbon chains. Chem. Phys. Lett. 44, 390–394.
Debouzy, J.C., Neumann, J.M., Herve, M., Daveloose, D., Viret, J., Apitz-
Castro, R., 1989. Interaction of antiagregant molecule ajoene with
membranes: an ESR and 1H, 2H, 31P NMR study. Eur. Biophys. J.
17, 211–216.
Roux, M. 1987. Etude des interactions lipides-prote´ines par RMN. Ph.D.
Thesis. Paris.
Seelig, J., 1977. Deuterium magnetic resonance: theory and application
to lipid membranes. Quart. Rev. Biophys. 10, 353–418.
Sklenar, V., Bax, A., 1987. Spin-echo water suppression for the generation
of pure-phase two-dimensional NMR spectra. J. Magn. Reson. 74,
469–479.
Tomoda, H., Hirano, T., Saito, S., Mutai, T., Araki, K., 1999. Substituent
effects on fluorescent properties of imidazo[1,2-a]pyridine-based com-
pounds. Bull. Chem. Soc. Jpn. 72, 1327–1334.
Wilson, K.P., McCaffrey, P.G., Hsiao, K., Pazhanisamy, S., Galullo, V.,
Bemis, G.W., Fitzgibbon, M.J., Caron, P.R., Murcko, M.A., Su, M.S.,
1997. The structural basis for the specificity of pyridinylimidazole
inhibitors of p38 MAP kinase. Chem. Biol. 4, 423–431.
Zhuang, L., Wang, B., Sauder, D.N., 2000. Molecular mechanism of
ultraviolet-induced keratinocyte apoptosis. J. Interf. Cytok. Res. 20,
445–454.
Debouzy, J.C., Dabouis, V., Crouzy, S., Bachelet, C., Favier, A., Pein-
nequin, A., Gueiffier, A., 2001. Modulation of intercalating properties
of pyrido[1,2-e]purines via side chain modifications: NMR and MD
studies. Pharmazie 56, 125–132.