Synthesis and Biological Evaluation of a Potent E-Selectin Antagonist

Article abstract of DOI:10.1021/jm990422n

An early step of the inflammatory response - the rolling of leukocytes on activated endothelial cells - is mediated by selectin/carbohydrate interactions. The tetrasaccharide sialyl Lewis^x (sLe^x) 1 is a ligand for E-, P-, and L-selectin and, therefore, serves as a lead structure to develop analogues which allow the control of acute and chronic inflammation. Here we describe the efficient synthesis (10 linear steps) of the potent sLe^x mimetic 2. Compared to sLe^x, compound 2 showed a 30-fold improved affinity in a static, cell-free E-selectin-ligand binding assay (IC50 = 36 μM). These data were confirmed by a marked inhibition in an in vitro cell-cell rolling assay which simulates in vivo conditions (IC50 ca. 40 μM). The assays are predictive for the in vivo efficacy of test compounds as indicated by a marked inhibitory effect of 2 in a thioglycollate induced peritonitis model of acute inflammation in mice (ED50 ca. 15 mg/kg).

Full text of DOI:10.1021/jm990422n

J . Med. Chem. 1999, 42, 4909-4913
4909
Syn th esis a n d Biologica l Eva lu a tion of a P oten t E-Selectin An ta gon ist
Gebhard Thoma,*,† Willy Kinzy,† Christian Bruns,† J ohn T. Patton,‡ J ohn L. Magnani,‡ and Rolf Ba¨nteli†
Novartis Pharma AG, P.O. Box, CH-4002 Basel, Switzerland, and GlycoTech Corporation, 14915 Broschart Road,
Rockville, Maryland 10850
Received August 20, 1999
An early step of the inflammatory responsesthe rolling of leukocytes on activated endothelial
cellssis mediated by selectin/carbohydrate interactions. The tetrasaccharide sialyl Lewisx (sLex)
1 is a ligand for E-, P-, and L-selectin and, therefore, serves as a lead structure to develop
analogues which allow the control of acute and chronic inflammation. Here we describe the
efficient synthesis (10 linear steps) of the potent sLex mimetic 2. Compared to sLex, compound
2 showed a 30-fold improved affinity in a static, cell-free E-selectin-ligand binding assay (IC50
) 36 µM). These data were confirmed by a marked inhibition in an in vitro cell-cell rolling
assay which simulates in vivo conditions (IC50 ≈ 40 µM). The assays are predictive for the in
vivo efficacy of test compounds as indicated by a marked inhibitory effect of 2 in a thioglycollate
induced peritonitis model of acute inflammation in mice (ED50 ≈ 15 mg/kg).
Ch a r t 1
In tr od u ction
Leukocyte influx from blood vessels into the sur-
rounding tissues causing inflammation is a beneficial
response of the body to control infections and injuries.1
Excessive leukocyte influx, however, may result in acute
or chronic reactions as observed in reperfusion injuries,
stroke, psoriasis, rheumatoid arthritis, or respiratory
diseases.2 An early step in the cascade of events which
finally leads to the extravasation of leukocytes is their
rolling on the endothelial cells which cover the inner
surface of the blood vessel wall.3 It has been shown that
a set of inducible adhesion molecules, the so-called
selectins, are involved in this step.4 A possible strategy
for preventing the adverse effects of an excessive leu-
kocyte influx is the inhibition of the leukocyte/selectin
interactions. The sialyl Lewisx tetrasaccharide 1 (Chart
1) appears to be a common epitope of the physiological
selectin ligands and is recognized by the three known
selectins (E-, P-, and L-selectin).4 Therefore, it became
a lead structure to identify simplified but more potent
inhibitors. Extensive work elucidated that the essential
pharmacophores required for binding to E-selectin are
the three hydroxyl groups of the L-fucose, the 4- and
the 6-hydroxyl group of the D-galactose, and the car-
boxylic acid function. On the basis of this knowledge, a
multitude of E-selectin antagonists have been pre-
pared.5
P r ep a r a tion of E-Selectin An ta gon ist 2
The sialyl Lewisx mimic 2 contains D-galactose and
L-fucose to provide all the essential hydroxyl groups
required for E-selectin binding. Instead of neuraminic
acid we incorporated S-cyclohexyllactic acid, a modifica-
tion which has already been shown to improve selectin
inhibition.6 N-Acetyl-D-glucosamine was replaced by a
glucal-derived building block.7
Commercially available triacetyl-D-glucal 3 was used
as starting material (Scheme 1). Hydrogenation in THF8
followed by deacetylation of the crude intermediate and
benzylidene protection of the 4- and the 6-hydroxyl
groups gave the required building block 4 which could
be isolated either by chromatography (88%) or, alter-
natively, by precipitation with hexane (78%). D-Galac-
tose was introduced using thioglycoside 59 which was
activated with NIS/trifluorosulfonic acid. As expected
the glycosylation was completely stereoselective leading
to â-linked disaccharide 6 in 71% yield. The stereo-
chemistry was deduced from the large coupling constant
of 8.0 Hz between H-1 Gal and H-2 Gal. No R-anomer
could be detected.10 Reductive opening of the ben-
zylidene acetal of 6 with NaCNBH3/HCl gave regiose-
lectively compound 7 (72%) with an unprotected hy-
droxyl group in the 4-position. Fucosylation of 7 was
In this paper we report the synthesis of the novel,
potent E-selectin antagonist 2 which can be prepared
from commercially available starting materials in 10
linear steps in an overall yield of 18%. Sialic acid and
N-acetyl glucosamine of the lead structure sialyl Lewisx
1 have been replaced by S-cyclohexyllactic acid and a
D-glucal-derived building block, respectively. The com-
bination of these two modifications led to a significantly
improved E-selectin inhibition of 2 compared to sLex.
* Author to whom correspondence should be addressed. E-mail:
gebhard.thoma@pharma.novartis.com.
† Novartis Pharma Inc.
‡ GlycoTech Corporation.
10.1021/jm990422n CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/28/1999

4910 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
Thoma et al.
Sch em e 1
Ta ble 1. E-Selectin Inhibition by Compound 2 Compared to
Sch em e 2
sLex (Static, Cell-Free Ligand Binding Assay)
compound 2
sLex (1)
IC50 [µM]
a
IC50 [µM]
rel IC50
33
24
51
0.028
0.027
0.039
1179
889
1308
36 ( 13
0.031 ( 0.007
1125 ( 214
a
Rel IC50 ) IC50(compound 2)/IC50(sLex).
achieved using thioglycoside 8.11 In situ generation of
the corresponding bromide followed by treatment with
Et4NBr gave R-glycoside 9 in 76% yield. The small
coupling constant of 4.0 Hz between H-1 Fuc and H-2
Fuc proved the stereochemistry. The corresponding
â-epimer could not be isolated. Removal of the acetates
in the presence of catalytic amounts of sodium methy-
late furnished tetraol 10 in 98% yield. Activation of the
3-hydroxyl group of the galactose with dibutyltinoxide12
followed by treatment with triflate 11 in the presence
of CsF gave trisaccharide 12 in 68% yield. Compound
12 is not very stable (probably due to lactone formation
involving the 2- or the 4-hydroxyl group of galactose).
The isolated material contained small quantities (∼5%)
of impuries and was immediately used in the final
hydrogenation. The desired selectin inhibitor was trans-
formed into its sodium salt 2 by passage through a
column with ion exchange resin and further purified by
reverse phase chromatography (82%).13
The triflate 11 was prepared from commercially
available D(+)-3-phenyllactic acid 13 (Scheme 2). Hy-
drogenation followed by benzyl ester formation using
the crude intermediate gave R-hydroxy ester 14 in 80%
yield. Treatment with trifluoromethanesulfonic anhy-
dride in the presence of di-tert-butyl pyridine furnished
triflate 11 in 75% yield. Compound 11 is surprisingly
stable and can be purified by chromatography on silica
gel. It can be stored in the deep freezer for months
without decomposition.
The measurement of the dose-dependent displacement
of the polymer by potential inhibitors allows the deter-
mination of their IC50 values (concentration to achieve
an inhibitory effect of 50%). Test compounds are always
compared to sialyl Lewisx 1 tetrasaccharide (Chart 1)
which is assayed as a reference on each test plate. To
compare the test results for compounds which have been
tested on different plates, we calculated the relative IC50
value which is defined by IC50(test compound)/IC50-
(sLex). In three separate assays analogue 2 (IC50 ) 36
µM, rel IC50 ) 0.03; Table 1) showed approximately 30-
fold improvement of activity compared to sLex (IC50
)
1125 µM, rel IC50 ) 1.00).
Advantages of this static cell free assay are the good
reproducibility and the high throughput. On the other
hand it has to be noted that rolling of leukocytes on
activated endothelium mediated by selectin/oligosac-
charide interactions is a nonequilibrium process which
occurs under hydrodynamic flow. Therefore, static in
vitro assays which measure inhibition under equilibri-
um conditions could generate data which are not predic-
tive of the in vivo behavior of test compounds under flow
conditions. To overcome this, a cell-based in vitro flow
assay has been developed which mimics the nonequi-
librium in vivo conditions and allows one to monitor
E-selectin-dependent rolling under shear stress.15
A
solution containing human polymorphonuclear neutro-
phils (PMNs) is pumped through a parallel plate flow
chamber containing a monolayer of human umbilical
vein endothelial cells (HUVECs) which have been
stimulated with TNF-R. Rolling is recorded by video
microscopy. The reduction of the number of interacting
cells (NIC) in the presence of a selectin antagonist is a
measure for its inhibitory effect. Sialyl Lewisx was
inactive in this assay up to 1000 µM. Antagonist 2,
however, led to 94% inhibition of the NIC at 200 µM,
Biologica l Eva lu a tion of 2
As a primary screen for E-selectin activity a cell-free,
competitive ligand binding assay has been developed
using a polylysine/sialyl Lewisa conjugate as a multi-
valent ligand.14 Immobilized E-selectin is incubated
with the polymer and then treated with test compounds.

Synthesis of a Potent E-Selectin Antagonist
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23 4911
68.4 (t), 69.6 (t), 70.2 (t), 72.3 (t), 74.6 (t), 75.1 (t), 76.0 (t),
79.6 (t), 80.5 (t), 83.1 (t), 99.1 (t), 99.5 (t), 183.0 (q). MS/HR
calcd for C27H45O15Na (M - Na)- 609.2758, found 609.2753.
60% at 50 µM, and 10% (not significant) at 10 µM. The
IC50 value of 2 for the flow assay can be estimated to
30-40 µM. The closely related E-selectin inhibitor 15
(Chart 1)6 which showed 10-fold improved E-selectin
inhibition compared to sLex in the static ligand binding
assay was used as a reference compound in the flow
assay. Compound 15 reduced the NIC by only 45% when
directly compared at 200 µM with 2 (94% reduction) on
the same test day using the same cells. Thus, the
inhibitory effects of E-selectin antagonists found in the
static equilibrium assay were confirmed by the non-
equilibrium flow assay. Our in vitro E-selectin assays
seem to be predictive for the in vivo efficacy of test
compounds. Thus we observed a significant inhibitory
effect of 2 (ED50 ) 15 mg/kg) in the thioglycollate
induced peritonitis reaction in mouse whereas sLex was
inactive in this model of acute inflammation at doses
up to 100 mg/kg.16
4: A suspension of 3 (11.0 g, 40.3 mmol) and Pd (10%) on
charcoal in THF (200 mL) was hydrogenated at 20 °C using a
balloon filled with H2 for 16 h. The catalyst was filtered off
and the solvent removed. The remaining oil was dissolved in
methanol (75 mL) and freshly prepared NaOMe (2.2 mmol)
was added and stirred at 20 °C for 6 h. Then, 1 g of acidic ion
exchange resin (Dowex Wx8) was added, the resin filtered off,
and the solvent removed. The crude material was evaporated
with CH3CN (2 × 50 mL) and suspended in CH3CN (300 mL).
Benzaldehyde dimethylacetal (12.25 g, 80.6 mmol) and PTSA
(0.43 g, 2.5 mmol) were added and stirred for 2 h at 20 °C.
NaHCO3 (1.65 g, 20 mmol) was added to the clear solution.
After 10 min the solids were filtered off, the solvent was
removed, and the residue was subjected to flash chromatog-
raphy (silica gel, toluene/ethyl acetate 2:1 f 1:1). Compound
4 (8.52 g, 89%) was isolated as a colorless solid. 1H NMR (400
MHz, CDCl3) δ ) 1.80 (1 H, tdd, 13.0/11.0/5.0 Hz, H-2ax), 2.02
(1 H, ddt, 13.0/5.0/1.5 Hz, H-2eq), 3.33 (1 H, td, 9.5/5.0 Hz,
H-5), 3.43 (1 H, t, 9.0 Hz, H-4), 3.56 (1 H, td, 12.5/2.0 Hz,
H-1ax), 3.70 (1 H, t, 10.0 Hz, H-6a), 3.88 (1 H, dddd, 11.0/9.0/
5.0/1.5, H-3), 3.99 (1 H, ddd, 12.0/5.0/1.5 Hz, H-1eq), 4.28 (1
H, dd, 10.0/5.0 Hz, H-6b), 5.56 (1 H, s, Ar-CH-), 7.35-7.50 (5
H, m, Ar-H). MS/HR calcd for C13H17O4 (M + H)+ 237.1127,
found 237.1124.
6: To a solution of 4 (4.50 g, 19.11 mmol), 59 (5.00 g, 12.74
mmol), and NIS (3.58 g, 15.93 mmol) in CH2Cl2 (50 mL) was
added droppwise at 0 °C a solution of trifluoromethanesulfonic
acid (0.15 M). When the color changed from dark red to dark
brown the addition was stopped. After complete consumption
of 4 (TLC, toluene/ethyl acetate 2:1) saturated NaHCO3
solution was added (30 mL) and the organic phase separated.
The aqueous phase was extracted with CH2Cl2 (50 mL), and
the combined organic phases were washed with Na2S2O3
solution (2 × 25 mL) and dried with Na2SO4. Following
filtration the solvent was removed and the residue subjected
to chromatography (silica gel, hexane/ether 1:1 f 1:3. Com-
pound 6 was isolated as a colorless solid (5.23 g, 72%). 1H NMR
(400 MHz, CDCl3) δ ) 1.85 (1 H, m, H-2ax), 1.96 (3 H, s, -C(O)-
CH3), 1.97 (1 H, m, H-2eq), 1.98 (3 H, s, -C(O)CH3), 2.03 (3 H,
s, -C(O)CH3), 2.13 (3 H, s, -C(O)CH3), 3.34 (1 H, td, 9.5/5.0
Hz, H-5), 3.56 (1 H, td, 12.5/2.5 Hz, H-1ax), 3.64 (1 H, t, 9.0
Hz, H-4), 3.71 (1 H, ddd, 8.0/5.5/1.0 Hz, H-5 Gal), 3.74 (1 H, t,
10.0 Hz, H-6a), 3.91 (1 H, m, H-3), 3.94 (1 H, dd, 11.0/5.5 Hz,
H-6a Gal), 4.01 (1 H, ddd, 11.5/5.0/2.0 Hz, H-1eq), 4.09 (1 H,
dd, 11.0/8.0 Hz, H-6b Gal), 4.29 (1 H, dd, 10.5/5.0 Hz, H-6b),
4.69 (1 H, d, 8.0 Hz, H-1 Gal), 4.98 (1 H, dd, 10.5/3.5 Hz, H-3
Gal), 5.22 (1 H, dd, 10.5/8.0 Hz, H-2 Gal), 5.33 (1 H, dd, 3.5/
1.0 Hz, H-4 Gal), 5.59 (1 H, s, Ar-CH-), 7.35-7.52 (5 H, m,
Ar-H). MS/HR calcd for C27H34O13 (M + Na)+ 589.1897, found
589.1896.
Con clu sion
The E-selectin antagonist 2 is a simplified analogue
of sLex which can be prepared in 10 linear steps in an
overall yield of 18%. The substantially increased inhibi-
tory effect of compound 2 compared to sLex was dem-
onstrated in both static and dynamic in vitro assays.
The in vitro data were confirmed in an in vivo mouse
model of acute inflammation. An explanation for the
improved activity of 2 compared to 15 could be an
additional interaction between E-selectin and the extra
hydroxymethylene group in 2. An alternative explana-
tion for the good activity of 2 could be an improved
preorganization of the bioactive conformation in solution
due to beneficial conformational restraints in 2 com-
pared to 15. We are currently investigating these
hypotheses in our laboratory.
Exp er im en ta l Section
Gen er a l. All reactions were carried out under an atmo-
sphere of dry argon. Commercially available absolute solvents
were used. The NMR spectra were recorded on a Bruker
Avance DPX 400 spectrometer. The signal assignments are
based on two-dimensional 1H/1H-correlation (COSY) and 1H/
13C-correlation spectroscopy (HSQC). The MS spectra were
obtained on a Finnigan MAT 90 mass spectrometer.
2: A suspension of 12 (700 mg, 0.660 mmol) and Pd (10%)
on charcoal (250 mg) in dioxane/water (18 mL, 5:1) was
hydrogenated at 20 °C for 56 h using a balloon filled with H2.
After filtration the solvent was removed and the residue
subjected to chromatography (silica gel, CHCl3/MeOH/H2O 5:5:
0.25 f 2:5:2). The product was dissolved in water and passed
through a column with Na ion exchange resin to give com-
pound 12 which was further purified by reverse phase chro-
matography (RP18, water/methanol 1:0 f 1:0.3). Lyophiliza-
7: At 0 °C a freshly prepared saturated solution of HCl in
diethyl ether is added dropwise to a suspension of 6 (2.00 g,
3.53 mmol) and NaCNBH3 (2.22 g, 35.3 mmol) in THF (60 mL).
After complete consumption of 6 (TLC ethyl acetate/hexane
3:1), solid NaHCO3 (1.5 g) was added and stirred for 5 min.
The mixture was diluted with ethyl acetate (300 mL), extracted
with NaHCO3 solution (2 × 150 mL) and brine (2 × 150 mL),
and dried with Na2SO4. Following filtration the solvent was
removed and the residue subjected to chromatography (silica
gel, toluene/acetone 3:1 f 2:1). Compound 7 was isolated as a
1
tion gave compound 12 as a colorless solid (340 mg, 81%). H
NMR (400 MHz, D2O) δ ) 0.78-1.75 (14 H, m, H-2ax, -CH2-
cC6H11), 1.12 (3 H, d, 6.5 Hz, H-6 Fuc), 2.17 (1 H, dd (br), 12.0/
5.0 Hz, H-2eq), 3.32 (1 H, m, H-5), 3.33 (1 H, dd, 9.5/3.0 Hz,
H-3 Gal), 3.42 (1 H, t (br), 11.5 Hz, H-1ax), 3.51 (1 H, t, 9.0
Hz, H-4), 3.55 (1 H, t, 6.0 Hz, H-5 Gal), 3.56 (1 H, t, 9.0 Hz,
H-2 Gal), 3.67 (2 H, d, 6.0 Hz, H-6a Gal, H-6b Gal), 3.71 (1 H,
dd, 10.5/4.0 Hz, H-2 Fuc), 3.75 (1 H, d (br), 3.0 Hz, H-4 Fuc),
3.76 (1 H, dd, 12.0/5.0 Hz, H-6a), 3.81 (1 H, dd, 10.5/3.0 Hz,
H-3 Fuc), 3.84 (1 H, d, 3.0 Hz, H-4 Gal), 3.86 (1 H, dd, 12.0/
2.5 Hz, H-6b), 3.88-4.01 (3 H, m, O-CH-CO2Na, H-1eq, H-3),
4.45 (1 H, d, 8.0 Hz, H-1 Gal), 4.71 (1 H, q, 6.5 Hz, H-5 Fuc),
4.89 (1 H, d, 4.0 Hz, H-1 Fuc). 13C NMR (100 MHz, D2O) δ )
15.9 (p), 26.1 (s), 26.3 (s), 26.5 (s), 30.6 (s), 32.2 (s), 33.7 (t),
34.0 (s), 41.6 (s), 60.4 (s), 62.0 (s), 65.5 (s), 66.7 (t), 67.3 (t),
1
colorless foam (1.45 g, 72%). H NMR (400 MHz, CDCl3) δ )
1.72-1.86 (2 H, m, H-2ax, H-2eq), 2.00 (3 H, s, -C(O)CH3), 2.01
(3 H, s, -C(O)CH3), 2.07 (3 H, s, -C(O)CH3), 2.17 (3 H, s, -C(O)-
CH3), 3.32 (1 H, ddd, 9.0/6.0/2.0 Hz, H-5), 3.40-3.47 (2 H, m,
H-1ax, H-4), 3.52 (1 H, ddd, 14.0/8.0/5.5 Hz, H-3), 3.63 (1 H,
dd, 10.5/6.0 Hz, H-6a), 3.75 (1 H, d, 1.5 Hz, OH), 3.82 (1 H,
dd, 10.5/2.0 Hz, H-6b), 4.00 (1 H, td, 6.5/1.0 Hz, H-5 Gal), 4.02
(1 H, m, H-1eq), 4.14 (2 H, d, 6.5 Hz, H-6a Gal, H-6b Gal),
4.56 (1 H, d, 8.0 Hz, H-1 Gal), 4.58 (1 H, d, 12.0 Hz, OBn),
4.62 (1 H, d, 12.0 Hz, OBn), 5.03 (1 H, dd, 10.5/3.5 Hz, H-3
Gal), 5.24 (1 H, dd, 10.5/8.0 Hz, H-2 Gal), 5.40 (1 H, dd, 3.5/

4912 J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23
1.0 Hz, H-4 Gal), 7.16-7.37 (5 H, m, Ar-H). MS/HR calcd for
Thoma et al.
68.2 (s), 81.9 (t), 128.6 (2 × t), 128.7 (2 × t), 128.9 (t), 134.4
(q), 167.6 (q). MS/CI 394 (M)+.
C
27H36O13 (M + Na)+ 591.1054, found 591.2051.
12: A suspension of 10 (800 mg, 0.980 mmol) and Bu2SnO
(268 mg, 1.08 mmol) in methanol (15 mL) was heated under
reflux for 2 h. The solvent was removed and the resulting
colorless foam dried in a vacuum for 60 h. The residue was
dissolved in DME (15 mL) followed by the addition of 11 (778
mg, 1.96 mmol) and CsF (179 mg, 1.18 mmol). The resulting
suspension was stirred at 20 °C for 8 h. Water (150 mL) was
added followed by extraction with ethyl acetate (3 × 50 mL).
The organic phase was washed with brine (2 × 50 mL) and
dried with Na2SO4, the solvent was removed, and the residue
was subjected to chromatography (silica gel, toluene/acetone
2.5:1). Compound 12 was isolated as a colorless foam (710 mg,
68%). 1H NMR (400 MHz, CDCl3) δ ) 0.84-1.86 (14 H, m,
H-2ax, -CH2-cC6H11), 1.17 (3 H, d, 6.5 Hz, H-6 Fuc), 2.09 (1 H,
dd (br), 13.0/5.0 Hz, H-2eq), 2.63 (1 H, m, Gal-6 OH), 2.73 (1
H, m (br), Gal-2 OH), 3.25 (1 H, s, Gal-4 OH), 3.26 (1 H, dd,
10.0/3.0 Hz, H-3 Gal), 3.32-3.42 (3 H, m, H-1ax, H-5, H-5 Gal),
3.67 (1 H, dd, 11.0/1.5 Hz, H-6a), 3.68-4.00 (8 H, m, H-1eq,
H-3, H-4, H-6b, H-2 Gal, H-4 Gal, H-6a Gal, H-6b Gal), 3.78
(1 H, s (br), H-4 Fuc), 4.01 (1 H, dd, 10.0/2.5 Hz, H-3 Fuc),
4.05 (1 H, dd, 10.0/3.0 Hz, H-2 Fuc), 4.32 (1 H, d, 8.0 Hz, H-1
Gal), 4.35 (1 H, m, O-CH-CH2-cC6H11), 4.37 (1 H, d, 12.0 Hz,
OBn), 4.45 (1 H, d, 12.0 Hz, OBn), 4.61 (1 H, d, 11.5 Hz, OBn),
4.63 (1 H, q, 6.5 Hz, H-5 Fuc), 4.64 (1 H, d, 11.5 Hz, OBn),
4.78 (2 H, m, 2 × OBn), 4.83 (1 H, d, 11.5 Hz, OBn), 4.97 (1 H,
d, 11.5 Hz, OBn), 5.06 (1 H, d, 3.0 Hz, H-1 Fuc), 5.17 (1 H, d,
12.0 Hz, OBn), 5.21 (1 H, d, 12.0 Hz, OBn), 7.20-7.42 (25 H,
m, Ar-H). MS (ESI) 1083 (M + Na)+.
14: A suspension of 13 (200 g, 1.20 mol) and Rh (5%) on
activated Al2O3 (6.0 g) in THF/H2O (2000 mL, 1:1) was
hydrogenated at 20 °C for 6 h. The catalyst was filtered off,
and the solvents were removed in a vacuum. The residue was
suspended in CH3OH/H2O (1250 mL, 9:1) and the pH adjusted
to 8.1 by the addition of an aqueous solution of cesium
carbonate. The solvents were removed, and the residue was
repeatedly evaporated with ethanol and then with hexane. The
gray material was dissolved in DMF (1300 mL). Benzylbromide
(219.3 g, 1.20 mol) was added at 20 °C within 20 min. After
the mixture was stirred for 16 h, CH2Cl2 (200 mL) was added,
the layers separated, and the aqueous layer was extracted with
CH2Cl2 (3 × 250 mL). The solvents were removed in a vacuum,
and the residue was subjected to chromatography (silica gel,
hexane/ethyl acetate 7:1). Compound 14 was isolated as a
colorless solid (265 g, 80%). 1H NMR (400 MHz, CDCl3) δ )
0.85-1.86 (13 H, m, CH-CH2-cC6H11), 2.70 (1 H, M, 3.5 Hz,
OH), 4.27 (1 H, m, CH-CH2-cC6H11), 5.20 (2 H, M, OBn), 7.34-
7.43 (5 H, m, Ar-H). 13C NMR (100 MHz, CDCl3) δ ) 27.0 (s),
27.2 (s), 27.4 (s), 33.3 (s), 34.6 (t), 34.8 (s), 43.0 (s), 68.1 (s),
69.6 (t), 129.2 (2 × t), 129.4 (t), 129,5 (2 × t), 136.2 (q), 176.7
(q). MS/EI 262 (M)+.
9: A solution of Br2 (298 mg, 1.86 mmol) in CH2Cl2 (2 mL)
was added dropwise at 0 °C to a solution of 811 (808 mg, 1.69
mmol) in CH2Cl2 (2 mL). After stirring for 30 min at 0 °C,
cyclohexene (0.2 mL) was added to consume excessive Br2. The
mixture was added within 10 min to a solution of 7 (800 mg,
1.41 mmol) and Et4NBr (353 mg, 1.69 mmol) in DMF/CH2Cl2
(10 mL, 1:1). The mixture was stirred for 44 h at 20 °C. Ethyl
acetate (100 mL) was added followed by washings with NaS2O3
solution (2 × 50 mL), water (2 × 50 mL), and brine (2 × 50
mL). The resulting solution was dried with Na2SO4 and
filtered, the solvent removed, and the residue subjected to
chromatography (silica gel, toluene/ethyl acetate 1.5:1). Com-
pound 9 was isolated as a colorless solid (1.05 g, 76%). In
addition a small amount of starting material 7 (100 mg, 12%)
1
was recovered. H NMR (400 MHz, CDCl3) δ ) 1.23 (3 H, d,
6.5 Hz, H-6 Fuc), 1.60 (1 H, m, H-2ax), 1.87 (3 H, s, -C(O)-
CH3), 1.97 (3 H, s, -C(O)CH3), 2.01 (1 H, m, H-2eq), 2.04 (3 H,
s, -C(O)CH3), 2.05 (3 H, s, -C(O)CH3), 3.32 (1 H, m, H-5), 3.36
(td, 11.5/1.5 Hz, H-1ax), 3.63 (1 H, dd, 10.5/2.0 Hz, H-6a), 3.68
(1 H, d (br), 2.5 Hz, H-4 Fuc), 3.78-3.85 (3 H, m, H-3, H-4,
H-6b), 3.86 (1 H, td, 7.0/1.0 Hz, H-5 Gal), 3.94 (1 H, dd, 10.0/
2.5 Hz, H-3 Fuc), 3.99 (1 H, m, H-1eq), 4.02 (2 H, d, 7.0 Hz,
H-6a Gal, H-6b Gal), 4.09 (1 H, dd, 10.0/4.0 Hz, H-2 Fuc), 4.39
(1 H, d, 12.0 Hz, OBn), 4.45 (1 H, d, 12.0 Hz, OBn), 4.55 (1 H,
d, 8.0 Hz, H-1 Gal), 4.61 (1 H, d, 11.5 Hz, OBn), 4.66 (1 H, q
(br), 6.5 Hz, H-5 Fuc), 4.73 (1 H, d, 12.0 Hz, OBn), 4.78 (2 H,
s, 2 × OBn), 4.82 (1 H, d, 11.5 Hz, OBn), 4.97 (1 H, dd, 10.5/
4.0 Hz, H-3 Gal), 4.98 (1 H, d, 12.0 Hz, OBn), 5.04 (1 H, d, 4.0
Hz, H-1 Fuc), 5.11 (1 H, dd, 10.5/8.0 Hz, H-2 Gal), 5.35 (1 H,
dd, 3.5/1.0 Hz, H-4 Gal), 7.24-7.44 (20 H, m, Ar-H). MS/HR
calcd for C54H64O17 (M + Na)+ 1007.4041, found 1007.4048.
10: A solution of 9 (1.00 g, 1.02 mmol) and freshly prepared
NaOMe (0.10 mmol) in methanol (10 mL) was stirred at 20
°C for 3 h. Then, 2 drops of acetic acid was added, the solvent
removed, and the residue subjected to chromatography (silica
gel, ethyl acetate/2-propanol 9:1 f 5:1). Compound 10 was
isolated as a colorless solid (820 mg, 98%). 1H NMR (400 MHz,
CD3OD) δ ) 1.25 (3 H, d, 6.5 Hz, H-6 Fuc), 1.68 (1 H, qd, 12.5/
5.0 Hz, H-2ax), 2.13 (1 H, dd (br), 12.5/5.0 Hz, H-2eq), 3.27 (1
H, ddd, 9.0/4.4/1.5 Hz, H-5), 3.41 (t (br), 12.5 Hz, H-1ax), 3.48
(1 H, m, H-5 Gal), 3.48 (1 H, dd, 9.5/3.0 Hz, H-3 Gal), 3.57 (1
H, dd, 9.5/7.5 Hz, H-2 Gal), 3.62 (1 H, dd, 11.0/1.5 Hz, H-6a),
3.70 (1 H, dd, 11.0/4.5 Hz, H-6a Gal,), 3.74 (1 H, t, 9.0 Hz,
H-4), 3.83 (1 H, dd, 11.0/7.0 Hz, H-6b Gal), 3.84 (1 H, d (br),
3.0 Hz, H-4 Gal), 3.89 (1 H, dd, 11.0/3.0 Hz, H-6b), 3.90-3.96
(3 H, m, H-1eq, H-3, H-4 Fuc), 3.95 (1 H, dd, 10.5/3.5 Hz, H-2
Fuc), 3.94 (1 H, dd, 10.5/2.5 Hz, H-3 Fuc), 4.35 (1 H, d, 7.5
Hz, H-1 Gal), 4.36 (1 H, d, 12.0 Hz, OBn), 4.50 (1 H, d, 11.5
Hz, OBn), 4.53 (1 H, d, 12.0 Hz, OBn), 4.64 (1 H, d, 11.5 Hz,
OBn), 4.73 (2 H, d, 11.5 Hz, 2 × OBn), 4.82 (1 H, d, 11.5 Hz,
OBn), 4.87 (1 H, m, H-5 Fuc), 4.93 (1 H, d, 11.5 Hz, OBn),
5.04 (1 H, d, 3.5 Hz, H-1 Fuc), 7.23-7.43 (20 H, m, Ar-H). MS/
HR calcd for C46H56O13 (M + Na)+ 839.3619, found 839.3624.
Ackn owledgm en t. We thank Mr. Franz Schwarzen-
bach and Mrs. Christine Bourquin for excellent techni-
cal support.
11: Compound 14 (265 g, 1.01 mol) was dissolved in CH2-
Cl2 (3000 mL) followed by the addition of 2,6-di-tert-butyl
pyridine (251.3 g, 1.27 mol). The solution was cooled to -20
°C, trifluoromethane sulfonic acid anhydride (334.9 g, 1.16 mol)
was added within 10 min, and the mixture was stirred for 4
h. The mixture was diluted with CH2Cl2 (1000 mL) and added
to a 1 M solution of K2HPO4 (5400 mL). The layers were
separated, and the aqueous layer was extracted with CH2Cl2
(3 × 1000 mL). After the mixture was dried with Na2SO4, the
solvent was removed (bath temp < 35 °C) and the residue
subjected to chromatography (silica gel, hexane/ethyl acetate
10:1). Compound 11 was isolated as a colorless oil (298 g, 75%).
Refer en ces
(1) Lobb, R. R. In Adhesion. Its Role in Inflammatory Disease;
Harlan, J . M., Liu, D. Y., Eds.; W. H. Freeman: New York, 1992;
Chapters 1-18, p 1. (b) Paulson, J . C. In Adhesion. Its Role in
Inflammatory Disease; Harlan, J . M., Liu, D. Y., Eds.; W. H.
Freeman: New York, 1992; Chapter 2, p 19-35.
(2) Mousa, S. A. Recent advances in cell adhesion molecule (CAM)
research and development: Targeting CAMs for therapeutic and
diagnostic applications. Drugs Future 1996, 21, 283-289.
(3) Spertini, O.; Luscinskas, F. W.; Gimbrone, M. A.; Tedder, T. F.
Monocyte attachment to activated human vascular endothelium
in vitro is mediated by leukocyte adhesion molecule-1 (L-selectin)
under nonstatic conditions. J . Exp. Med. 1992, 175, 1789-1792.
(4) Kansas, G. S. Selectins and their ligands: current concepts and
controversies. Blood 1996, 88, 3259-3287.
1
[R]D ) 38.6 ( 0.4° (CHCl3, 20 °C). H NMR (400 MHz, CDCl3)
δ ) 0.87-1.77 (11 H, m, CH-CH2-cC6H11), 1.80 (1 H, ddd, 15.0/
9.0/4.0 Hz, CH-CHaHb-cC6H11), 1.91 (1 H, ddd, 15.0/9.0/5.0 Hz,
CH-CHaHb-cC6H11), 5.21 (1 H, dd, 9.0/4.0 Hz, CH-CH2-cC6H11),
5.23 (1 H, d, 12.0 Hz, OBn), 5.27 (1 H, d, 12.0 Hz, OBn), 7.35-
7.37 (5 H, m, Ar-H). 13C NMR (100 MHz, CDCl3) δ ) 25.7 (s),
26.0 (s), 26.1 (s), 30.1 (q), 32.0 (s), 33.2 (s), 33.4 (s), 39.3 (t),
(5) The efforts of the various research groups involved have been
recently been summarized in an excellent review. Simanek, E.
E.; McGarvey, G. J .; J ablonowski, J . A.; Wong, C.-H. Selectin-
Carbohydrate Interactions: From Natural Ligands to Designed
Mimics. Chem. Rev. 1998, 98, 833-862.

Synthesis of a Potent E-Selectin Antagonist
J ournal of Medicinal Chemistry, 1999, Vol. 42, No. 23 4913
(6) Kolb, H. C.; Ernst, B. Development of tools for the design of
selectin antagonists. Chem. Eur. J . 1997, 1571-1578. (b)
J ahnke, W.; Kolb, H. C.; Blommers, M. J . J .; Magnani, J . L.;
Ernst, B. Comparison of the bioactive conformations of sialyl
Lewisx and a potent sialyl Lewisx mimic. Angew. Chem., Intl.
Ed. 1997, 36, 2603-2607. (c) Kolb, H. C.; Ernst, B. Recent
progress in the glycodrug area. Pure Appl. Chem. 1997, 69,
1879-1884.
(7) A glucal-derived building block has been incorporated into
selectin inhibitors before, but these compounds showed no
improved E-selectin inhibition compared to sLex. Toepfer, A.;
Kretzschmar, G.; Schuth, S.; Sonnentag, M. Malonic acid deriva-
tives as sialyl Lewis X mimetics. Bioorg. Med. Chem. Lett. 1997,
7, 1317-1322.
(8) Hydrogenation in methanol gave less satisfactory results, most
probably due to partial decomposition of the sensitive starting
material.
(13) It has been reported that compounds contaminated with trace
amounts of polyanions released from ion exchange resins can
give false positive test results in selectin assays. To exclude such
an effect, a batch of the free carboxylic acid of compound 2 has
been prepared without applying ion exchange resin. The biologi-
cal results in our assays were identical for both compounds.
Kretzschmar, G.; Toepfer, A.; Huels, C.; Krause, M. Pitfalls in
the preparation and biological evaluation of sialyl Lewisx
mimetics as potential selectin antagonists. Tetrahedron 1997,
53, 2485-2494.
(14) Thoma, G.; Magnani, J . L.; Oehrlein, R.; Ernst, B.; Schwarzen-
bach, F.; Duthaler, R. O. Synthesis of Oligosaccharide-Polylysine
Conjugates: A Well Characterized Sialyl Lewisa Polymer for
ELISA. J . Am. Chem. Soc. 1997, 119, 7414-15.
(15) Thoma, G.; Duthaler, R. O.; Ernst, B.; Magnani, J . L.; Patton,
J . T. Preparation of multivalent polymers as agents for preparing
biologically active compounds. PCT Int. Appl., WO 97/19105,
1997. (b) Thoma, G.; Patton, J . T.; Magnani, J . L.; Ernst, B.;
Oehrlein, R.; Duthaler, R. O. Versatile functionalization of
polylysine: synthesis, characterization, and use of neoglycocon-
jugates. J . Am. Chem. Soc. 1999, 121, 5919-5929.
(16) This model of acute inflammation has been described earlier.
Bosse, R.; Vestweber, D. Only simultaneous blocking of the L-
and P-selectin completely inhibits neutrophil migration into
mouse peritoneum. Eur. J . Immunol. 1994, 24, 3019-3024. We
demonstrated that the model is dependent on E-selectin by using
an anti-mouse E-sel mAb (10E9.6; 10 µg/animal) which showed
a marked inhibitory effect.
(9) Contour, M. O.; Defaye, J .; Little, M.; Wong, E. Stereoselective
thioglycoside synthesis. Part XI. Zirconium(IV) chloride-cata-
lyzed synthesis of 1,2-trans-1-thioglycopyranosides. Carbohydr.
Res. 1989, 193, 283-287.
(10) Very surprisingly, under identical reaction conditions but using
galactosyl donors with an ether linkage in the 3-position and
ester protecting groups in the 2-, 4-, and 6-position we isolated
10-15% of the corresponding R-anomer in addition to the desired
â-linked product (∼60%). The formation of the side product could
not be suppressed when NIS/TfOH was replaced by DMTST.
(11) Yamazaki, F.; Kitajima, T.; Nukada, T.; Ito, T.; Ogawa, T.
Synthetic studies on cell surface glycans. 66. Synthesis of an
appropriately protected core glycotetraoside, a key intermediate
for the synthesis of “bisected” complex-type glycans of a glyco-
protein. Carbohydr. Res. 1990, 201, 15-30.
(12) David, S.; Hanessian, S. Regioselective manipulation of hydroxyl
groups via organotin derivatives. Tetrahedron 1985, 41, 643-
663.
J M990422N