Discovery, Synthesis and Evaluation of Novel Cholesterol Absorption Inhibitors

Article abstract of DOI:10.1111/j.1747-0285.2012.01421.x

Chemical-based common feature pharmacophore modelling of Niemann Pick C1 Like 1 inhibitors was performed to provide some insights on the important pharmacophore features essential for Niemann Pick C1 Like 1 inhibition using Discovery Studio V2.5. After in-house database screening, a new series of substituted oxazolidinones, selected from the top ranked hits, have been synthesized and evaluated as novel cholesterol absorption inhibitors. All compounds demonstrated effect of different degrees in lowering the total cholesterol in serum, especially compounds 1a, 2a and 2d, the potency of which was comparable to that of ezetimibe. It was also found that 1a, 1d and 2d could raise high-density lipoprotein cholesterol levels markedly. Interestingly, compounds 2a-2f appeared to have the moderate potential to lower triglyceride levels, which were superior to that of normal cholesterol absorption inhibitors including ezetimibe. Two classes of novel substituted oxazolidinones were identified by virtual screening as the candidates of cholesterol absorption inhibitors to treat coronary artery disease. All compounds demonstrated effect of different degrees in lowering the total cholesterol in serum. Interestingly, compounds 2a-2f appeared to have the moderate potential to lower triglyceride levels, which were superior to that of normal cholesterol absorption inhibitors including ezetimibe.

Full text of DOI:10.1111/j.1747-0285.2012.01421.x

ª 2012 John Wiley & Sons A/S
doi: 10.1111/j.1747-0285.2012.01421.x
Chem Biol Drug Des 2012; 80: 426–433
Research Article
Discovery, Synthesis and Evaluation of Novel
Cholesterol Absorption Inhibitors
Xiaoyun Zhu†, Jianfeng Ji†, Dandan Huang,
Clinically, as a monotherapy, ezetimibe achieves an average 18%
reduction in LDL. When used in combination with a statin, ezetim-
ibe affords an additional 15–18% reduction in LDL beyond that
achieved by the statin alone (4). Ezetimibe was discovered using an
empirical, in vivo screening protocol (3). However, researchers at
Schering-Plough ⁄ Merck recently demonstrated that it inhibited cho-
lesterol absorption via binding to Niemann Pick C1 Like 1 (NPC1L1)
which was highly expressed in the proximal intestine and was a
key mediator of cholesterol transport (5). Thus NPC1L1 lies in the
ezetimibe sensitive pathway for cholesterol absorption, making it a
likely candidate for the target of ezetimibe (6).
Yan Zhu, Chunlei Tang, Xuan Yang, Hai
Qian* and Wenlong Huang*
Centre of Drug Discovery, State Key Laboratory of Natural
Medicines, China Pharmaceutical University, 24 Tongjiaxiang,
Nanjing 210009, China
*Corresponding authors: Hai Qian, qianhai24@163.com; Wenlong
Huang, ydhuangwenlong@126.com
These authors contributed equally to this work.
Chemical-based common feature pharmacophore
modelling of Niemann Pick C1 Like 1 inhibitors was
performed to provide some insights on the impor-
tant pharmacophore features essential for Niemann
Pick C1 Like 1 inhibition using Discovery Studio
V2.5. After in-house database screening, a new ser-
ies of substituted oxazolidinones, selected from the
top ranked hits, have been synthesized and evalu-
ated as novel cholesterol absorption inhibitors. All
compounds demonstrated effect of different
degrees in lowering the total cholesterol in serum,
especially compounds 1a, 2a and 2d, the potency of
which was comparable to that of ezetimibe. It was
also found that 1a, 1d and 2d could raise high-den-
sity lipoprotein cholesterol levels markedly. Inter-
estingly, compounds 2a–2f appeared to have the
moderate potential to lower triglyceride levels,
which were superior to that of normal cholesterol
absorption inhibitors including ezetimibe.
The substituted azetidinone (7), ezetimibe, is currently commercially
available for the treatment of hypercholesterolemia. However, the
effectiveness of available antilipidemic therapies is limited, in part
because of poor patient compliance due to unacceptable side
effects and tolerability as well as minimal efficacy or potency (8).
Moreover, certain drug products may not be advantageous to all
patients because of genetic polymorphisms regarding cholesterol
biosynthesis. Furthermore, potential side effects of absorption of
certain azetidinones may be detrimental. For these reasons, there is
a continuing need for novel antilipidemic agents that may be used
alone or in combination with other agents that provide increased
efficacy and tolerability with decreased toxicity (9).
In the present investigation, we have generated Common Feature
Pharmacophore models using Discovery Studio V2.5 (DS) for NPC1L1
inhibitors. Then the best pharmacophore model was used as a 3D
query to screen in-house database for identifying new cholesterol
absorption inhibitors. The hit compounds were subsequently sub-
jected to filtering by Lipinski's rule of five. And several substituted
oxazolidinones (Figure 1), selected from the top ranked hits, have
been synthesized and evaluated as novel cholesterol absorption
inhibitors.
Key words: cholesterol absorption inhibitor, Niemann Pick C1 Like
1, oxazolidinones, pharmacophore
Abbreviations: NPC1L1, Niemann Pick C1 Like 1; LDL, low-density
lipoprotein; DS, Discovery Studio; TC, total cholesterol; HDL-C, high-den-
sity lipoprotein cholesterol; TG, triglyceride; HA, hydrogen bond accep-
tor; HD, hydrogen bond donor; HY, hydrophobic.
Methods and Materials
Received 20 February 2012, revised 18 April 2012 and accepted for pub-
lication 8 May 2012
Pharmacophore modelling studies
High serum cholesterol levels have been unequivocally implicated
as a cause of coronary artery disease, which is the major cause of
death and cardiovascular morbidity in the world (1). Along with sta-
tins, which lower low-density lipoprotein (LDL) cholesterol levels by
inhibition of cholesterol biosynthesis mediated by HMG-CoA reduc-
tase, the approval of the cholesterol absorption inhibitor, ezetimibe
(Figure 1), has been shown to be effective in decreasing blood LDL
cholesterol levels by decreasing uptake of dietary cholesterol (2,3).
Pharmacophore model generation
A training set of 12 reported NPC1L1 inhibitors with binding affinity,
as listed in Figure S1 in the Supporting Information (10,11), were
used to generate the pharmacophore model after added hydrogen.
Before starting the pharmacophore generation process, conforma-
tional analysis of the molecules was performed using the polling
algorithm. The conformers of training set were generated using the
BEST methods. The Common Feature Pharmacophore Generation
426

Novel Cholesterol Absorption Inhibitors
General procedure for the preparation of 5
Compound 4 (7.38 mmol) was dissolved in CH2Cl2 (50 mL) under
nitrogen followed by the addition of triethylamine (14.76 mmol) and
chilled to )20 ꢀC. A solution of triphosgene (7.38 mmol) dissolved
in CH2Cl2 (30 mL) was then added drop-wise over 15 min. The reac-
tion mixture was stirred overnight allowing it to warm to 25 ꢀC
over this period. The reaction was quenched by the addition of a
saturated solution of ammonium chloride (100 mL). The layers were
separated, organic layer dried with Na2SO4, filtered and evaporated
to dryness, yielding the brown oily residue which was purified by
column chromatography (PE ⁄ EtOAc 8:1, V ⁄ V) to provide 5.
2-Azetidinone
OH
Oxazolidinone
R1
OH
R2
O
N
O
N
F
O
F
O
Ezetimibe
F
1a-1g, 2a-2f
Figure 1: Structures of ezetimibe and target compounds
1a–1g, 2a–2f.
protocol implemented in DS V2.5 was applied to construct the
model.
General procedure for the preparation of 6
Compound 5 (13.2 mmol) was dissolved in anhydrous methanol
(80 mL) and cooled to )10 ꢀC. Sodium borohydride (6.6 mmol) was
then added portion-wise to the reaction and the mixture allowed to
warm to 25 ꢀC, stirred for 3 h. A 2 M HCl was then added to the
reaction and extracted with EtOAc, the resulting layers separated,
organic layer dried with Na2SO4, filtered and evaporated to dryness.
The resulting compound was purified using column chromatography
(PE ⁄ EtOAc 5:1, V ⁄ V) to provide 6.
Database screening
The best pharmacophore model was used as a 3D structural query
for retrieving potent molecules from in-house database comprised
of 530 compounds which were designed as cholesterol absorption
inhibitors. For each molecule in the database, 255 conformers were
generated with the fast conformer generation method, allowing a
maximum energy of 20 kcal ⁄ mol above that of the most stable con-
formation. The database screening was carried out using ligand
pharmacophore mapping protocol implemented in DS with
best ⁄ flexible search option (12). The hit compounds were subse-
quently subjected to filtering by Lipinski's rule of five. And two
classes of substituted oxazolidinones (Figure 1), selected from the
top ranked hits, were subjected to further studies.
General procedure for the preparation of 1d–1g,
2c–2f and 7–11
Compound 6 (3 mmol), corresponding acid (3 mmol) and 4-dimeth-
ylaminopyridine (DMAP) (0.2 equiv.) were dissolved anhydrous
CH2Cl2 (80 mL),
A solution of dicyclohexylcarbodiimide (DCC)
(4.5 mmol) dissolved in CH2Cl2 (20 mL) was then added drop-wise
over 10 min. The reaction mixture was stirred overnight at room
temperature. The mixture allowed to cool to )5 ꢀC, stand for 6 h.
The reaction mixture filtered and evaporated to dryness. The result-
ing compound was purified using column chromatography (PE ⁄ E-
tOAc) to provide 1d–1g, 2c–2f and 7–11.
Chemistry
All reagents were purchased from Shanghai Chemical Reagent Com-
pany (Shanghai, China). Column chromatography (CC): silica gel 60
(100–200 mesh). Thin-layer chromatography: silica gel 60 F254
plates (250 mm; Qingdao Ocean Chemical Company, Qingdao, China).
M.p.: capillary tube; uncorrected. IR spectra: Shimadzu FTIR-8400S
1
General procedure for the preparation of 1a–1c
and 2a–2b
spectrophotometer; per cm. H NMR spectra: Bruker ACF-300Q appa-
ratus at 300 MHz, in CDCl3 unless otherwise indicated; d in ppm rel.
to Me4Si, J in Hz. Mass spectrometry (MS): Hewlett-Packard 1100
LC ⁄ MSD spectrometer, in m ⁄ z; Elemental analyses: CHN-O-Rapid
instrument; %Purity of the target compounds (>97%) were deter-
mined by HPLC analysis (UV detector, wavelength: 230 nm, mobile
phase composed of ammonium acetate buffer (0.02 M, pH adjusted
to 7.0 with ammonium hydroxide) and acetonitrile).
Compounds 7–11 (3 mmol) dissolved in a mixture of EtOAc (10 mL)
and methanol (10 mL), and 10% Pd ⁄ C (10% weight) was then
added to the reaction. The reaction mixture was stirred at room
temperature for 8 h, filtered and evaporated to dryness. The result-
ing compound was purified using column chromatography (PE ⁄ E-
tOAc) to afford 1a–1c and 2a–2b.
General procedure for the preparation of 4
Hydroxyacetone (23.25 mmol), 4-fluoroaniline (15.5 mmol),
1-((4R,5S)-3-(4-fluorophenyl)-4-(4-
hydroxyphenyl)-2-oxooxazolidin-5-yl)ethyl
3-methylbenzoate (1a)
3
(14.1 mmol) and L-proline (2.82 mmol) were dissolved in DMSO
(150 mL) under nitrogen and stirred for 12 h at room temperature.
The reaction mixture was extracted with ammonium chloride solu-
tion (300 mL) and EtOAc (2 · 150 mL). The organic layers were
washed with saturated sodium chloride solution and dried with
Na2SO4, filtered and evaporated to dryness. The residue was then
purified using column chromatography with a (PE ⁄ EtOAc 6:1, V ⁄ V)
to afford 4.
1
Yield: 79.6%. Yellow powder, m.p.: 110–111 ꢀC. H-NMR (300 MHz,
CDCl3) d: 9.55–9.54 (s, 1H, -OH), 7.72–7.60 (m, 4H, Ar-H), 7.30–7.18
(m, 4H, Ar-H), 7.15–7.04 (m, 2H, Ar-H), 6.90–6.71 (m, 2H, Ar-H),
5.37–5.33 (m, 1H, -OCHCH3), 5.10–5.08 (d, 1H, -CHN-, J = 4.5 Hz),
4.42–4.40 (m, 1H, -OCH-), 2.21 (s, 3H, -ArCH3), 1.43–1.42 (m, 3H, -
CH3). IR (KBr, per cm) t: 3411, 2925, 2855, 1724, 1615, 1511, 1451,
1427, 1275, 1106, 1050, 929, 834, 785, 682. MS (70eV) m ⁄ z: 436.3
Chem Biol Drug Des 2012; 80: 426–433
427

Zhu et al.
([M + H]+). Anal. calcd. for C25H22FNO5: C, 68.96; H, 5.09; N, 3.22;
Found (%): C, 68.86; H, 5.12; N, 3.21%.
1-((4R,5S)-4-(2,3-dimethoxyphenyl)-3-(4-
fluorophenyl)-2-oxooxazolidin-5-yl)ethyl
3-methylbenzoate (1f)
1
Yield: 84.5%. Colourless oil. H-NMR (300 MHz, DMSO-d6) d: 7.69–
1-((4R,5S)-3-(4-fluorophenyl)-4-(4-
hydroxyphenyl)-2-oxooxazolidin-5-yl)ethyl
benzo[d][1,3]dioxole-5-carboxylate (1b)
7.42 (m, 4H, Ar-H) 7.37–7.21 (m, 4H, Ar-H), 6.91–6.88 (m, 3H, Ar-H),
5.88–5.60 (d, 1H, -CHN-, J = 3.3 Hz), 5.39–5.32 (m, 1H, -OCHCH3),
4.68–4.67 (m, 1H, -OCH-), 4.06 (s, 3H, -OCH3), 3.91 (s, 3H, -OCH3), 2.13
(s, 3H, -ArCH3), 1.46–1.44 (m, 3H, -CH3). IR (KBr, per cm) t: 3319,
2938, 1758, 1721, 1590, 1512, 1484, 1397, 1275, 1132, 1069, 834,
764. MS (70 eV) m ⁄ z: 480.4 ([M + H]+). Anal. calcd. for C27H26FNO6:
C, 67.63; H, 5.47; N, 2.92; Found: C, 67.68; H, 5.36; N,2.90%.
1
Yield: 78.8%. White powder, m.p.: 114–115 ꢀC. H-NMR (300 MHz,
DMSO-d6) d: 9.55–9.54 (s, 1H, -OH), 7.56–7.42 (m, 4H, Ar-H), 7.29–
7.10 (m, 5H, Ar-H), 6.90–6.71 (m, 2H, Ar-H), 6.11–6.10 (s, 2H, -
OCH2O-), 5.72–5.43 (d, 1H, -CHN-, J = 4.5 Hz), 5.35–5.32 (m, 1H, -
OCHCH3), 4.60–4.59 (m, 1H, -OCH-), 1.40–1.39 (m, 3H, -CH3). IR
(KBr, per cm) t: 3412, 2986, 1731, 1615, 1512, 1489, 1444, 1383,
1277, 1159, 1037, 927, 834, 760, 647. MS (70 eV) m ⁄ z: 466.3
([M + H]+). Anal. calcd. for C25H20FNO7: C, 64.51; H, 4.33; N, 3.01;
Found: C, 64.60; H, 4.29; N, 2.98%.
1-((4R,5S)-4-(2,3-dimethoxyphenyl)-3-
(4-fluorophenyl)-2-oxooxazolidin-5-yl)ethyl
4-nitrobenzoate (1g)
1
Yield: 83.0%. Yellow powder, m.p.: 157–159 ꢀC. H-NMR (300 MHz,
DMSO-d6) d: 8.84–8.81 (m, 2H, Ar-H), 7.69–7.32 (m, 4H, Ar-H),
7.29–7.01 (m, 2H, Ar-H), 6.73–6.70 (m, 3H, Ar-H), 5.88–5.72 (d, 1H,
-CHN-, J = 4.2 Hz), 5.53–5.47(m, 1H, -OCHCH3), 4.76–4.73 (m, 1H, -
OCH-), 4.04 (s, 3H, -OCH3), 3.95 (s, 3H, -OCH3), 1.51–1.49 (m, 3H, -
CH3). IR (KBr, per cm) t: 3462, 3101, 2935, 1760, 1628, 1589, 1547,
1512, 1484, 1462, 1397, 1345, 1167, 1070, 921, 833, 772. MS
(70 eV) m ⁄ z: 533.2 ([M + Na]+). Anal. calcd. for C26H23FN2O8: C,
61.17; H, 4.54; N, 5.49; Found: C, 61.21; H, 4.55; N, 5.53%.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-hydroxyphenyl)-2-oxooxazolidin-5-yl)ethyl
2-phenylacetate (1c)
Yield: 81.2%. Brown oil. 1H-NMR (300 MHz, DMSO-d6) d: 9.57 (s,
1H, -OH), 7.53–7.26 (m, 6H, Ar-H), 7.15–7.05 (m, 5H, Ar-H), 6.93–
6.88 (m, 2H, Ar-H), 5.40–5.28 (d, 1H, -CHN-, J = 5.1 Hz), 5.16–5.08
(m, 1H, -OCHCH3), 4.48–4.04 (m, 1H, -OCH-), 3.56–3.34 (s, 2H,
ArCH2CO-), 1.28–1.26 (m, 3H, -CH3). IR (KBr, per cm) t: 3319,
2935, 1736, 1614, 1598, 1512, 1402, 1307, 1228, 1066, 835, 764,
696. MS (70 eV) m ⁄ z: 458.3 ([M + Na]+). Anal. calcd. for
C25H22FNO5: C, 68.96; H, 5.09; N, 3.22; Found: C, 69.02; H, 5.07; N,
3.25%.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-hydroxyphenyl)-2-oxooxazolidin-5-yl)ethyl
2-phenoxyacetate (2a)
Yield: 73.8%. Yellow powder, m.p.: 77–78 ꢀC. 1H-NMR (300 MHz,
CDCl3) d: 7.30–7.19 (m, 4H, Ar-H), 7.14–6.91 (m, 4H, Ar-H), 6.89–
6.77 (m, 5H, Ar-H), 5.38–5.30 (m, 1H, -OCHCH3), 4.99–4.98 (d, 1H, -
CHN-, J = 5.4 Hz), 4.73 (s, 2H, -OCH2CO-), 4.40–4.38 (m, 1H, -OCH-),
1.44–1.35 (M, 3H, -CH3). IR (KBr, per cm) t: 3407, 2985, 2937,
1751, 1615, 1599, 1511, 1449, 1224, 1052, 868, 785, 690. MS
(70 eV) m ⁄ z: 474.2 ([M + Na]+). Anal. calcd. for C25H22FNO6: C,
66.51; H, 4.91; N, 3.10; Found: C, 66.55; H, 4.87; N, 3.13%.
1-((4R,5S)-3-(4-fluorophenyl)-4-(4-
methoxyphenyl)-2-oxooxazolidin-5-yl)ethyl
3-methylbenzoate (1d)
1
Yield: 75.4%. White powder, m.p.: 119–121 ꢀC. H-NMR (300 MHz,
DMSO-d6) d: 7.51–7.40 (m, 4H, Ar-H), 7.39–7.28 (m, 6H, Ar-H),
7.18–6.90 (m, 2H, Ar-H), 5.85–5.57 (d, 1H, -CHN-, J = 4.2 Hz), 5.38–
5.31 (m, 1H, -OCHCH3), 4.65–4.61 (m, 1H, -CHO-), 3.72 (s, 3H, -
OCH3), 2.16 (s, 3H, -ArCH3), 1.47–1.40 (m, 3H, -CH3). IR (KBr, per
cm) t: 3651, 3414, 3072, 2960, 2935, 1754, 1716, 1588, 1513,
1460, 1352, 1231, 1131, 1045, 942, 847, 760, 683. MS (70 eV) m ⁄ z:
472.3 ([M + Na]+). Anal. calcd. for C26H24FNO5: C, 69.48; H, 5.38; N,
3.12; Found: C, 69.54; H, 5.39; N, 3.14%.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-hydroxyphenyl)-2-oxooxazolidin-5-yl)ethyl
2-methyl-2-phenoxypropanoate (2b)
Yield: 74.9%. Yellow oil. 1H-NMR (300 MHz, DMSO-d6) d: 9.75 (s, 1H,
-OH), 7.43–7.31 (m, 4H, Ar-H), 7.30–7.06 (m, 4H, Ar-H), 7.02–6.70 (m,
5H, Ar-H), 5.41–5.23 (d, 1H, -CHN-, J = 4.5 Hz), 5.24–5.18 (m, 1H, -
OCHCH3), 4.53–4.48 (m, 1H, -OCH-), 1.45 (s, 6H, -C(CH3)2), 1.29–1.25
(m, 3H, -CH3). IR (KBr, per cm) t: 3330, 2990, 2925, 1836, 1615, 1597,
1512, 1453, 1385, 1275, 1173, 1054, 972, 834, 753, 696. MS (70 eV)
m ⁄ z: 502.2 ([M + Na]+). Anal. calcd. for C27H26FNO6: C, 67.63; H, 5.47;
N, 2.92; Found: C, 67.71; H, 5.45; N, 2.95%.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-methoxyphenyl)-2-oxooxazolidin-5-yl)ethyl
4-nitrobenzoate (1e)
1
Yield: 83.6%. White powder, m.p.: 120–122 ꢀC. H-NMR (300 MHz,
DMSO-d6) d: 8.83–8.80 (m, 2H, Ar-H), 7.52–7.47 (m, 2H, Ar-H),
7.44–7.29 (m, 4H, Ar-H), 7.18–7.08 (m, 2H, Ar-H), 6.94–6.89 (m, 2H,
Ar-H), 5.84–5.66 (d, 1H, -CHN-, J = 5.1 Hz), 5.48–5.45 (m, 1H, -OCH-
CH3), 4.74–4.71 (m, 1H, -OCH-), 3.70 (s, 3H, -OCH3), 1.51–1.46 (m,
3H, -CH3). IR (KBr, per cm) t: 3483, 3075, 2936, 1759, 1612, 1586,
1513, 1442, 1428, 1306, 1193, 971, 824, 752, 624. MS (70 eV) m ⁄ z:
503.2 ([M + Na]+). Anal. calcd. for C25H21FN2O7: C, 62.50; H, 4.41;
N, 5.83; Found: C, 62.56; H, 4.39; N, 5.86%.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-methoxyphenyl)-2-oxooxazolidin-5-yl)ethyl
2-phenoxyacetate (2c)
Yield: 68.7%. Yellow oil. 1H-NMR (300 MHz, DMSO-d6) d: 7.30–
7.10 (m, 6H, Ar-H), 6.98–6.96 (m, 2H, Ar-H), 6.92–6.86 (m, 5H, Ar-H),
428
Chem Biol Drug Des 2012; 80: 426–433

Novel Cholesterol Absorption Inhibitors
5.50–5.38 (d, 1H, -CHN-, J = 5.1 Hz), 5.26–5.18 (m, 1H, -OCHCH3),
4.87–4.69 (s, 2H, -OCH2CO-), 4.55–4.49 (m, 1H, -OCH-), 3.71 (s, 3H,
-OCH3), 1.33–1.31 (m, 3H, -CH3). IR (KBr, per cm) t: 3319, 2988,
1758, 1560, 1512, 1495, 1397, 1275, 1193, 1089, 835, 692. MS
(70 eV) m ⁄ z: 488.2 ([M + Na]+). Anal. calcd. for C26H24FNO6: C,
67.09; H, 5.20; N, 3.01; Found: C, 67.14; H, 5.18; N, 2.99%.
), 3.81 (s, 3H, -OCH3), 3.71 (s, 3H, -OCH3), 1.46 (s, 6H, -C(CH3)2),
1.26–1.23 (m, 3H, -CH3). IR (KBr, per cm) t: 3319, 2988, 1758, 1596,
1512, 1486, 1397, 1275, 1133, 1070, 834, 695. MS (70 eV) m ⁄ z:
524.3 ([M + H]+). Anal. calcd. for C29H30FNO7: C, 66.53; H, 5.78; N,
2.68; Found: C, 66.46; H, 5.79; N, 2.70%.
Evaluation of hypocholesterolemic effects
Cholesterol absorption inhibitory activities of new substituted oxazo-
lidinones 1a–1g and 2a–2f were assessed in orally dosed, choles-
terol-fed Sprague-Dawley rats as reported in literature (7,13).
Hypocholesterolemic Sprague-Dawley rats were used to test the
efficacy of these compounds. Male Sprague-Dawley rats, weighing
between 200 and 250 g, were maintained on rodent chow and pro-
vided with water. In order to induce a hypercholesterolemia in Spra-
gue-Dawley rats, their chow diet must be supplemented with 1%
cholesterol and 0.5% cholic acid. Treatment protocols consisted of
feeding this diet for 7 days. Test compounds, dissolved in 0.5 mL
corn oil, were administered to the animals by oral gavage daily
(mid-light cycle) during this time period. On the last day, the ani-
mals were sacrificed and blood sample was taken for lipid analy-
ses. Plasma cholesterol levels, triglyceride (TG) levels and high-
density lipoprotein cholesterol (HDL-C) levels were determined by a
commercial modification of the cholesterol oxidase method which
was available in a kit form.
1-((4R,5S)-3-(4-fluorophenyl)-4-
(4-methoxyphenyl)-2-oxooxazolidin-5-yl)ethyl
2-(4-chlorophenoxy)acetate (2d)
Yield: 73.9%. Yellow oil. 1H-NMR (300 MHz, DMSO-d6) d: 7.49–
7.40 (m, 2H, Ar-H), 7.39–7.28 (m, 4H, Ar-H), 7.15–7.10 (m, 2H, Ar-H),
6.93–6.90 (m, 4H, Ar-H), 5.48–5.37 (d, 1H, -CHN-, J = 5.4 Hz), 5.25–
5.18 (m, 1H, -OCHCH3), 4.87–4.73 (s, 2H, -OCH2CO-), 4.53–4.48 (m,
1H, -OCH-), 3.71 (s, 3H, -OCH3), 1.31–1.30 (m, 3H, -CH3). IR (KBr,
per cm) t: 3472, 1753, 1618, 1513, 1491, 1384, 1192, 824, 641. MS
(70 eV) m ⁄ z: 522.2 ([M + Na]+). Anal. calcd. for C26H23ClFNO6: C,
62.47; H, 4.64; Cl, 7.09; N, 2.80; Found: C, 62.38; H, 4.65; N,
2.81%.
1-((4R,5S)-4-(2,3-dimethoxyphenyl)-3-
(4-fluorophenyl)-2-oxooxazolidin-5-yl)ethyl
2-phenoxyacetate (2e)
Yield: 77.6%. Yellow oil. 1H-NMR (300 MHz, CDCl3) d: 7.40–7.21
(m, 5H, Ar-H), 7.05–6.76 (m, 7H, Ar-H), 5.44–5.32(d, 1H, -CHN-,
J = 4.2 Hz), 5.42–5.30 (m, 1H, -OCHCH3), 4.62 (s, 2H, -OCH2CO-),
4.44–4.31(m, H, -OCH-), 3.90 (s, 3H,-OCH3), 3.81 (s, 3H,-OCH3),
1.46–1.39 (m, 3H, -CH3). IR (KBr, per cm) t: 3319, 2936, 1758, 1599,
1512, 1485, 1398, 1273, 1192, 835, 756, 691. MS (70 eV) m ⁄ z:
496.3 ([M + H]+). Anal. calcd. for C27H26FNO7: C, 65.45; H, 5.29; N,
2.83; Found: C, 65.47; H, 5.31; N, 2.79%.
Results and Discussion
Pharmacophore modelling studies
By analysis of the chemical nature of the training set compounds,
important feature types, such as hydrogen bond acceptor (HA),
hydrogen bond donor (HD) and hydrophobic (HY), were selected for
each pharmacophore model generation. Finally, several pharmaco-
phore models were generated, of which the best model contains
one HD, two HAs and two HY features (Figure 2). The common fea-
tures model was used as a 3D query to screen in-house database,
as a result, 73 compounds were hit, then filtering by Lipinski's rule
of five. Two classes of representative oxazolidinones (13 com-
pounds) were selected, and subjected to synthesis and evaluation
(Table 1).
1-((4R,5S)-4-(2,3-dimethoxyphenyl)-3-
(4-fluorophenyl)-2-oxooxazolidin-5-yl)ethyl
2-methyl-2-phenoxypropanoate (2f)
1
Yield: 73.4%. Colourless oil. H-NMR (300 MHz, DMSO-d6) d: 7.47–
7.23 (m, 5H, Ar-H), 7.07–6.80 (m, 5H, Ar-H), 5.52–5.51 (d, 1H, -CHN-
, J = 4.2 Hz), 5.24–5.21 (m, 1H, -OCHCH3), 4.57–4.56 (m, 1H, -OCH-
A
B
Figure 2: The best pharmacophore model for Niemann Pick C1 Like 1(NPC1L1) inhibitors where HA, HD and HY are illustrated in green,
magenta and cyan, respectively. (A) The best pharmacophore model for NPC1L1 inhibitors. (B) 3D spatial relationship and distance constraints
of the model. HA, hydrogen bond acceptor; HD, hydrogen bond donor; HY, hydrophobic.
Chem Biol Drug Des 2012; 80: 426–433
429

Zhu et al.
Table 1: Structures of the synthesized oxazolidinones
Chemistry
moderate yield and excellent stereoselectivity. Treatment of 4 with
triphosgene and triethylamine resulted in the formation of oxazolidi-
none 5. Reducing reaction of the side chain ketone provided alco-
The structures of the substituted oxazolidinones 1a–1g and 2a–
2f are presented in Table 1. All the compounds were characterized
by spectroscopic methods. These compounds were prepared accord-
ing to the sequence of reactions, illustrated in Scheme 1. Utilizing
an asymmetric three-component Mannich reaction reported by List
and coworkers (10,11), hydroxyacetone (1) was subsequently
reacted with 4-fluoroaniline (2), substituted benzaldehyde 3 in the
presence of catalytic L-proline to afford 1,2-amino alcohol 4 in
hol
6 by treatment with NaBH4 in ethanol ⁄ ethyl acetate.
Compounds 7–11, 1d–1g and 2c–2f were prepared by coupling
alcohol 6 to corresponding acid in the presence of DCC ⁄ DMAP,
and 7–11 were subsequently debenzylated using transfer hydroge-
nation conditions to afford corresponding oxazolidinones 1a–1c
and 2a–2b.
430
Chem Biol Drug Des 2012; 80: 426–433

Novel Cholesterol Absorption Inhibitors
R1
R1
NH2
OH
O
R1
O
a
b
+
+
HO
O
HN
N
O
CHO
F
F
4
1
2
O
O
3
5
F
c
R1
R1
Scheme 1: Synthesis of the tar-
get compounds 1a–1g and 2a–2f
[Reagents and conditions: (a) L-pro-
line (20 mol%), DMSO, 25 ꢀC, 12 h;
(b) triphosgene, Et3N, CH2Cl2,
R2
HO
d
O
R
R
1 = 4-OBn
1 = 4-OH
7-11:
N
N
O
e
1a-1c, 2a-2b:
F
F
O
O
6
1d-1g, 2c-2f
)20 ꢀC–25 ꢀC, 12 h; (c) NaBH4,
CH3OH, )10 ꢀC–25 ꢀC; (d) 4-dimeth-
ylaminopyridine (DMAP) ⁄ dicyclohexyl-
carbodiimide (DCC), CH2Cl2, r.t.; (e)
H2, Pd ⁄ C (10%), EtOH ⁄ EtOAc, r.t.].
Reagents and conditions: (a) L-proline (20 mol%), DMSO, 25ºC, 12 h;
(b) triphosgene, Et
3N, CH2Cl2, –20ºC~25ºC, 12 h; (c) NaBH4, CH3OH, –10ºC~25ºC;
(d) DMAP/DCC, CH2Cl2, r.t.; (e) H2, Pd/C (10%), EtOH/EtOAc, r.t. .
In vivo cholesterol absorption inhibitory
activities
Table 2: Cholesterol absorption inhibition and mapping results
of new synthesized oxazolidinones and reference compound
Cholesterol absorption inhibition was assessed in orally dosed, cho-
lesterol-fed Sprafue-Dawley rats as previously described (13). The
results were presented in Table 2. As can be seen from the data,
all of the new compounds demonstrated effect of different degrees
in lowering the total cholesterol (TC) in serum, especially com-
pounds 1a, 2a and 2d, the potency of which was comparable to
that of ezetimibe. It was also found that 1a, 1d and 2d could
raise HDL-C levels markedly. More interestingly, compounds 2a–2f
appeared to have the moderate potential to lower the TG levels,
which was superior to that of normal cholesterol absorption inhibi-
tors including ezetimibe. The possible reason for this interesting
phenomenon is that the synthesized compounds 2a–2f were
designed by the introduction of derivatives of the phenoxyacetic
scaffold of clofibrate into side chain of oxazolidinones (Figure 3).
The fibric-acid derivatives–bezafibrate, ciprofibrate, clofibrate, fenof-
ibrate and gemfibrozil–primarily decrease TG by increasing lipopro-
tein lipase activity and by decreasing the release of free fatty acids
from peripheral adipose tissue (14).
Compounda
TCb
(% reduction)
TGc
(% reduction)
HDL-Cd
(% increase)
Best
Fit valuee
1a
49.91**
11.98
36.71*
22.69
16.34
30.38
16.61
48.72**
14.73
NE
NE
NE
NE
NE
NE
NE
5.67
10.19
5.06
12.23*
6.04
8.46
NE
29.69**
5.12
19.05
34.31**
25.25
12.64
21.03
23.86*
22.05
25.37*
27.13**
14.32
4.57
3.56
4.35
4.29
3.43
3.92
3.75
4.4
3.89
3.68
4.09
3.43
3.54
4.41
1b
1c
1d
1e
1f
1g
2a
2b
2c
29.01
2d
2e
42.26**
28.88*
27.04
2f
10.37
37.66**
Ezetimibe
48.3**
HDL-C, high-density lipoprotein cholesterol; TC, total cholesterol; TG, triglyc-
eride; NE, no effect.
**p < 0.01; *p < 0.05 as compared with the high-cholesterol diets group.
a6–8 Sprague-Dawley rats per group; dose: 50 mg ⁄ kg.
bReduction of total cholesterol comparing to the one in animals fed by high-
cholesterol diets.
Table 2 shows the Best Fit values of the selected compounds
together with their corresponding experimental TC values. The best
pharmacophore model, which aligned to ezetimibe and compound
1a that was the most potent derivative of this series as represen-
tative examples, is shown in Figure 4. As can be seen from Fig-
ure 4, the result suggests that the best pharmacophore model
perfectly fits ezetimibe. From the overall analysis, we conclude that
the best pharmacophore model truly reflects the features of NPC1L1
inhibitors.
cReduction of triglyceride comparing to the one in animals fed by high-cho-
lesterol diets.
dIncrease of HDL-C comparing to the one in animals fed by high-cholesterol
diets.
eBest Fit value indicates how well the features in the pharmacophore over-
lap the chemical features in the molecule.
NPC1L1 inhibition. After in-house database screening, a new series
of substituted oxazolidinones, selected from the top ranked hits,
have been synthesized and evaluated as novel cholesterol absorp-
tion inhibitors. Among all of these derivatives, compounds 1a, 2a
and 2d demonstrated high potencies to lower the TC in serum, the
potencies of which were comparable to that of the reference com-
pound, ezetimibe. Moreover, compounds 1a, 1d and 2d could raise
HDL-C levels markedly. Compounds 2a–2f appeared interestingly to
Conclusions and Future Directions
In this study, chemical based common feature pharmacophore mod-
elling of NPC1L1 inhibitors has been created using DS. The common
feature pharmacophore model was performed to provide some
insights on the important pharmacophore features essential for
Chem Biol Drug Des 2012; 80: 426–433
431

Zhu et al.
R1
O
O
O
N
O
F
2b, 2f
O
O
O
O
Cl
Clofibrate
R1
Figure 3: Structures of clofibrate
and target compounds 2a–2f. Syn-
thesized compounds 2a–2f were
designed by introduction of deriva-
tives of the phenoxyacetic scaffold
of clofibrate into side chain of oxa-
zolidinones.
R1
O
O
O
O
O
O
N
O
N
O
Cl
F
F
O
O
2a, 2c,2e
2d
A
B
Figure 4: The best pharmaco-
phore model aligned to ezetimibe
and compound 1a. (A) The best
model aligned to compound 1a. (B)
The best model aligned to ezetimibe.
HA, HD and HY are illustrated in
green, magenta and cyan, respec-
tively. HA, hydrogen bond acceptor;
HD, hydrogen bond donor; HY, hydro-
phobic.
have the moderate potential to lower TG levels. Compound 2d with
phenoxyacetic side chain showed the better potencies in raising
HDL-C levels, lowering TC levels, lowering TG levels and pharmaco-
phore feature mapping (Best Fit value: 4.09), respectively. It may be
considered as a promising candidate for the development of potent
agents for treatment of coronary artery disease.
putative native and recombinant rat sterol transporter Niemann-
Pick C1 Like 1 (NPC1L1) protein. Biochim Biophys Acta;1722:
282–292.
2. Clader J.W. (2004) The discovery of ezetimibe: a view from out-
side the receptor. J Med Chem;47:1–9.
3. McMasters D.R., Garcia-Calvo M., Maiorov V., McCann M.E.,
Meurer R.D., Bull H.G., Lisnock J., Howell K.L., Devita R.J. (2009)
Spiroimidazolidinone NPC1L1 inhibitors. 1: discovery by 3D-similar-
ity-based virtual screening. Bioorg Med Chem Lett;19:2965–2968.
4. Burnett J.R., Huff M.W. (2006) Cholesterol absorption inhibitors
as a therapeutic option for hypercholesterolaemia. Expert Opin
Investig Drugs;15:1337–1351.
Acknowledgments
This study was supported by the National Natural Science Founda-
tion of China (No. 81172932); the Project Program of State Key Lab-
oratory of Natural Medicines, China Pharmaceutical University (No.
JKGZ201103); and the Fundamental Research Funds for the Central
Universities of China (No. 2J10023 and JKY2011009).
5. Burnett D.A. (2004) Beta-lactam cholesterol absorption inhibitors.
Curr Med Chem;11:1873–1887.
6. Wang Y., Qian H., Huang W., Zhang H., Zhou J. (2011) Synthesis
and biological evaluation of ezetimibe analogs as possible
cholesterol absorption inhibitors. Lett Drug Des Discov;8:500–505.
7. Wang Y., Zhang H., Huang W., Kong J., Zhou J., Zhang B.
(2009) 2-Azetidinone derivatives: design, synthesis and evalua-
References
tion of cholesterol absorption inhibitors. Eur
Chem;44:1638–1643.
J
Med
1. Iyer S.P., Yao X., Crona J.H., Hoos L.M., Tetzloff G., Davis H.R.
Jr, Graziano M.P., Altmann S.W. (2005) Characterization of the
432
Chem Biol Drug Des 2012; 80: 426–433

Novel Cholesterol Absorption Inhibitors
8. Sampalis J.S., Bissonnette S., Boukas S. (2011) Effectiveness of
ezetimibe in reducing the estimated risk for fatal cardiovascular
events in hypercholesterolaemic patients with inadequate lipid
control while on statin monotherapy as measured by the SCORE
model. Adv Prev Med;2011:1–7.
9. Kashani A., Sallam T., Bheemreddy S., Mann D.L., Wang Y., Foo-
dy J.M. (2008) Review of side-effect profile of combination eze-
timibe and statin therapy in randomized clinical trials. Am J
Cardiol;101:1606–1613.
13. Salisbury B.G., Davis H.R., Burrier R.E., Burnett D.A., Bowkow
G., Caplen M.A., Clemmons A.L. et al. (1995) Hypocholesterolem-
ic activity of a novel inhibitor of cholesterol absorption, SCH
48461. Atherosclerosis;115:45–63.
14. Farmer J.A., Gotto A.M. Jr (1995) Currently available hypolipi-
daemic drugs and future therapeutic developments. Baillieres
Clin Endocrinol Metab;9:825–847.
Supporting Information
10. Pfefferkorn J.A., Larsen S.D., Van Huis C., Sorenson R., Barton
T., Winters T., Auerbach B. et al. (2008) Substituted oxazolidi-
nones as novel NPC1L1 ligands for the inhibition of cholesterol
absorption. Bioorg Med Chem Lett;18:546–553.
11. Ritter T., Kvaerno L., Werder M., Hauser H., Carreira E.M. (2005)
Heterocyclic ring scaffolds as small-molecule cholesterol absorp-
tion inhibitors. Org Biomol Chem;3:3514–3523.
12. Zhu X., Huang D., Lan X., Tang C., Zhu Y., Han J., Huang W.,
Qian H. (2011) The first pharmacophore model for potent G pro-
tein-coupled receptor 119 agonist. Eur J Med Chem;46:2901–
2907.
Additional Supporting Information may be found in the online ver-
sion of this article:
Figure S1. Structures and NPC1L1 binding affinity of training set.
Please note: Wiley-Blackwell is not responsible for the content or
functionality of any supporting materials supplied by the authors.
Any queries (other than missing material) should be directed to the
corresponding author for the article.
Chem Biol Drug Des 2012; 80: 426–433
433