Structure-based lead identification of ATP-competitive MK2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.04.018

MK2 kinase is a promising drug discovery target for the treatment of inflammatory diseases. Here, we describe the discovery of novel MK2 inhibitors using X-ray crystallography and structure-based drug design. The lead has in vivo efficacy in a short-term preclinical model.

Full text of DOI:10.1016/j.bmcl.2011.04.018

Bioorganic & Medicinal Chemistry Letters 21 (2011) 3818–3822
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure-based lead identification of ATP-competitive MK2 inhibitors
Tjeerd Barf, Allard Kaptein, Sander de Wilde, Ruud van der Heijden, Richard van Someren, Dennis Demont,
Carsten Schultz-Fademrecht, Judith Versteegh, Mario van Zeeland, Nicole Seegers, Bert Kazemier,
Bas van de Kar, Maaike van Hoek, Jeroen de Roos, Henri Klop, Ruben Smeets, Claudia Hofstra,
⇑
Jorrit Hornberg, Arthur Oubrie
Merck Research Laboratories, MSD, PO Box 20, 5340 BH Oss, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
MK2 kinase is a promising drug discovery target for the treatment of inflammatory diseases. Here, we
describe the discovery of novel MK2 inhibitors using X-ray crystallography and structure-based drug
design. The lead has in vivo efficacy in a short-term preclinical model.
Received 11 February 2011
Revised 29 March 2011
Accepted 3 April 2011
Available online 16 April 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
MAPKAPK2
MK2
Anti-TNF
SBDD
a
Rheumatoid arthritis
Inhibition of the action of the pro-inflammatory cytokines TNF
and IL-6 have shown both in animal models and in the clinic to be
efficacious in the control of autoimmune diseases like rheumatoid
a
data for the MK2 heterozygote mice in the CIA model.3 Since the
heterozygote mice also show a clear reduction in disease severity
and in LPS-induced production of TNF
a
and IL-6,5 full inhibition
arthritis or psoriasis. In addition to TNF
a
and IL-6 neutralization
of MK2 activity does not seem to be required to obtain a physiolog-
strategies via biologics, many pharmaceutical companies have
worked on the development of p38 inhibitors in order to block
the production of these pro-inflammatory cytokines. Although sev-
eral p38 inhibitors have been tested in phase 2 clinical trials, none
have progressed to phase 3 clinical trials.1 The lack of progress
seems due to two main findings. First, drug toxicity observed for
p38 inhibitors tested in the clinic may result in dose-limiting effi-
cacy. Secondly, p38 inhibition results in the activation of the ERK,
JNK and IKK signaling pathways via a feedback control of TAK1.2
MAPK activated protein kinase 2 (MK2) is downstream from
p38 in this pathway and a key mediator regulating the production
ically relevant effect.
In this letter, we report the discovery of a novel class of spiro-d-
lactam MK2 inhibitors using X-ray crystallography and structure
based drug design. The project was instigated by determination
of the X-ray structures of MK2 bound to two reference compounds,
1 and 2 (Figs. 1 and 2).6 This was done using a construct comprising
residues 41-364,7 as was also described elsewhere.8
The X-ray structure of MK2:1 revealed hydrogen bonding inter-
actions between (1) the compound’s pyridine nitrogen and MK2
backbone NH of Leu141; (2) cyclic amide carbonyl and catalytic
Lys93 (Fig. 2). The NH of the cyclic amide does not make a hydro-
gen bond with Asp207 as was observed in other structures of
MK2:1.8 This is due to the presence of malonate in the active site,
which was used as precipitant in the crystallization solutions.7 In
our structure, the NH of the cyclic amide makes an artificial polar
interaction with malonate rather than Asp207. The quinoline moi-
ety of the inhibitor is located in the front pocket, which is open,
of TNFa and IL-6. MK2 knockout mice show a strong reduction in
disease incidence and disease severity score in the arthritic CIA
model.3 Specific modulation of MK2 activity, without inhibition
of other downstream signaling events is hypothesized to avoid
p38-associated adverse events. In addition, MK2 does not partici-
pate in the feedback control of TAK1 and may therefore result in
sustained inhibition of the inflammatory response mediated via
the p38/MK2 signaling pathway.4 Furthermore, MK2 may be an
interesting target for pharmacological modulation based on the
O
O
OH
NH
N
N
N
N
N
H
⇑
Corresponding author.
E-mail address: arthur.oubrie@merck.com (A. Oubrie).
NH2
N
N
1
2
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.04.018
Figure 1. Pyrrolopyridinone 1 and pyrazole derivative 2.

T. Barf et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3818–3822
3819
case the methylamine group ends up alpha to the amide in the cy-
clized product.6a In choosing the scaffold, the pyrazole was favored
over the pyrrole owing to its greater synthetic accessibility, even
though the pyrrole scaffold is normally superior to the pyrazole
when comparing identical substructures. For instance, EC50 values
of 8.5 and 216 nM were reported for pyrrole 1 and the otherwise
identical pyrazole twin, respectively.6a With an EC50 of 151 nM in
our IMAP assay,9 hydrid compound 3 is not as potent as the refer-
ence compounds (EC50À1 = 49 nM; EC50À2 = 28 nM; Table 1), yet it
demonstrated sufficient activity to be considered as a good starting
point. In the ribose pocket, the crystal structure of MK2:3 shows a
similar binding mode to the previous protein:ligand structures
with the cyclic amide forming a hydrogen bonding network with
Lys93 and Asp207 (Fig. 3). The primary amine of 3 is hydrogen
bonded to the main chain carbonyl of Glu190 but the distance to
Asn191 appears too long for proper formation of a hydrogen bond.
The position of the amine of 3 thus appears less optimal for the for-
mation of polar interactions as compared to 2 and this may in part
explain the relatively low inhibitory activity of 3. Also, the reduced
potency of 3 can be attributed to the absence of the ionic interac-
tion as demonstrated in the MK2:2 complex.
In the next step, we envisaged that incorporation of the amino
function in a rigid aliphatic spiro ring would reduce the number of
rotatable bonds and thus improve biochemical potency. However,
the synthesis of those examples initially posed a challenge, since
the optimal ring size and exact position of the basic amine was
not known. To circumvent trial and error, we chose to examine
the space in the ribose pocket first in terms of tolerability towards
cycloalkyl rings in combination with the more potent pyrrolopy-
ridinone scaffold. Another potential advantage of aliphatic ring
systems was to overcome the lack of cellular activity observed
for des-spiro derivative 3 (data not shown). Cyclopentyl analogues
were the first derivatives we decided to explore owing to availabil-
ity of starting materials and their relatively small size. A conver-
gent synthesis route towards the cyclopentyl derivatives was
employed (Scheme 1). For the generation of the key pyrimidine
building block C, we relied on the Stille coupling reaction with
an ethylvinyl tin reagent on 2,4-dichloropyrimidine. Direct bro-
mination on the resulting enol ether B in ethanol was achieved
to give C with high yield. Preparation of the cyclopentyl piperidin-
edione started with acylation of commercially available 1-(amino-
methyl)cyclopentane carboxylate D with ethyl malonyl chloride.
Cyclization of E with a Dieckmann condensation and subsequent
decarboxylation rendered intermediate F in good yields. Generally,
the Paal–Knorr condensation reaction with chloropyrimidine
derivative C worked reasonably well. The generation of final com-
pound 4 was part of a small library effort, in which aryl moieties
were introduced in varying yields on chloropyrimidine G via Suzu-
ki coupling reactions. A number of analogues were prepared but no
significant potency improvements were made within this chemical
series (not shown).
Figure 2. X-ray structures of MK2-41-364 bound to compounds 1 (yellow) and 2
(pink).7 Specific amino acids are labeled.
hydrophobic and has a flat surface, well suited to accommodate
planar substitutions. The quinoline nitrogen is not involved in a
hydrogen bond with protein or water, and the thiol of Cys140 is
too far away for a direct interaction. However, the aryl ring stacks
on top of the main chain carbonyl of Leu141, and a consistently
short distance of ꢀ3.2 Å was observed in all quinoline-containing
structures. The structure of MK2:2 showed that the quinoline
and pyridine groups of the ligand make the same interactions with
MK2 as in 1. The carboxylic acid of 2 makes polar interactions with
Lys93 and the main chain NH of Asp207, while the primary amine
is hydrogen bonded to the carbonyl groups of the main chain of
Glu190 and side chain of Asn191 (Fig. 2).
Analysis of these structures led to synthesis of a hybrid pyrazo-
lopyridinone 3, which combines the rigid lactam moiety of 1 with
the primary amine of 2 (Fig. 3). A similar approach was followed by
Pfizer researchers using the propylamine analogue of 2, but in this
O
NH
*
N
H
O
O
NH
NH
O
*
*
N
N
N
H
OH
*
N
NH2
N
Cyclopentyl
Hybrid, 3
NH2
Quinoline analogue 4 was among the most potent cyclopentyl
pyrimidine MK2 inhibitors and exhibited a moderate biochemical
potency (EC50 = 234 nM), comparable to methylamino derivative
3 (Table 1). A crystal structure of MK2 bound to a close analogue
of 4 was solved (Fig. 3). This structure indicated that the cyclopen-
tyl ring fitted snugly in the ribose pocket, apparently located with-
in hydrogen-bond distance of the main chain of Glu190 and side
chain of Asn191. The cyclopentyl ring with its hydrophobic nature
thus appears to be tolerated in the polar ribose pocket owing to the
formation of weak polar interactions. Unfortunately, properties
including compound polarity, membrane permeability, and micro-
somal stability varied significantly within this chemical series and
optimization of all these parameters simultaneously appeared
problematic. For instance, the kinetic solubility for this subseries
was consistently low (Table 1). Introduction of polar aryl
Figure 3. X-ray structures of MK2-41-264 complexed with pyrazole
cyclopentyl analogue.
3 and a

3820
T. Barf et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3818–3822
Table 1
Biochemical and cell activity, and in vitro ADME properties
a
b
Compound MK2 IMAP EC50 (nM) TNF-THP1 EC50
(lM) PAMPA nm/s Caco-2 nm/s Kinetic solubility (pH 7.4, mg/L) HLM/RLM (t½ min) PPB (%) c Log P
1
2
4
5a
5b
49
28
234
6
38%@10
l
M
M
6
5
0
33
0
23
37
35/40
99.7
NT
NT
NT
97.0
2.7
0.1
3.5
1.6
1.7
7%@10
50%@10
4.8
l
NTc
NT
0
NT
NT
NT
1
NT/NT
NT/NT
NT/NT
91/>120
lM
4
0.98
0
a
b
c
Values are means of two independent experiments with duplicates for each dilution per experiment.9
THP1 cells are pre-incubated for 30 min with MK2 inhibitors prior to stimulation with LPS. Values are means of duplicates for each dilution.14
NT means Not tested.
logue 4. The prerequisite Boc-protected oxolactam intermediate
H was essentially prepared as described earlier,10 and in analogy
with cyclopentyl oxolactam F as outlined in Scheme 1. After the
Suzuki coupling with 3,4-methylenedioxophenyl- or quinolin-3-
ylboronic acid, the final envisaged products could be isolated fol-
lowing acidic removal of the Boc protective group.
The introduction of an amine function as in 5a and 5b resulted
in improvement in terms of both potency and physicochemical
parameters as compared to fully aliphatic rings (Table 1). Activity
was improved to single digit nanomolar inhibition in biochemical
assays for these, and for many other analogues.11 Compound 5b
displayed the best overall profile since it inhibited LPS-induced
OEt
O
a
b
N
Cl
N
N
N
N
N
O
Br
Cl
Cl
Cl
A
B
C
O
O
d
OEt
O
NH2
O
HN
NH
c
MeO
MeO
O
and e
D
E
F
TNF
peripheral blood mononuclear cells (PBMCs, EC50 = 0.26
cellular inhibition was confirmed in the Hsp27 phosphorylation as-
say in THP1 cells (EC50 = 0.62 M), a target engagement marker for
MK2 activity. Also, 5b displayed a high degree of selectivity and ki-
nases that are directly associated with the TNF pathway were not
inhibited at 1
M.11
a
production/release in THP1 cells (EC50 = 0.98
l
l
M) and
M). The
O
O
NH
NH
g
f
N
N
N
l
N
N
H
N
N
H
Cl
a
G
4
l
The molecular basis for the MK2 inhibition is well understood
on the basis of crystal structures of 5b with both MK2 and MK3,
a close analog of MK2. MK3 was recently described as a suitable
surrogate structural model for MK2.12 In our hands, crystals of
MK3 could indeed be grown more readily than those of MK2. Fur-
thermore, MK3 crystals contained only one rather than 12 mole-
cules per asymmetric unit and diffracted to higher resolution.13
The structures of MK2 and MK3 bound to 5b are shown in Fig. 4.
The active sites of both proteins are indeed very similar, both in
terms of amino acid composition and general architecture of the
pocket. Only one amino acid is different, Leu141 (MK2) vs
Met121 (MK3). This difference does not influence the shape, size,
or physicochemical properties of the ATP pocket. Compound 5b
binds nearly identical to MK2 and MK3. In both structures and
Scheme 1. Reagents and conditions: (a) 1-Ethoxyvinyltri-N-butyltin, PdCl2(Ph3P)2,
DMF, 80 °C; KF workup (79%); (b) NBS, THF/H2O (5/2), rt (89%); (c) ClCOCH2COOEt,
Et3N, DCM, 0 °C–rt (72%); (d) NaOMe, MeOH, 65 °C; (e) ACN/H2O (1/1), 80 °C (79%,
two steps); (f) Cmpd. C, NH4OAc, EtOH, rt (59%); (g) library: arylboronic acids,
Pd(Ph3P)4, K2CO3, toluene/EtOH (4/1), 140 °C, MW (5–30%).
substitutions in the front pocket that were necessary to rescue sol-
ubility proved to be detrimental for membrane permeability, and
thus cellular potency. Nevertheless, the cycloalkyl ring derivatives
provided confidence that aza–spiro groups would be worthwhile
pursuing, and this prompted us to investigate a third scaffold
modification.
Compound 5a and 5b represent the combined effort (Scheme 2).
The synthetic route towards these 4-piperidylpyridinone deriva-
tives was largely identical to the preparation of cyclopentyl ana-
O
O
NH
NH
N
N
H
N
O
a
b
Cl
and c
N
N
H
J
Boc
Boc
O
*
*
NH
N
*
O
N
N
N
H
O
N
H
5a
5b
Scheme 2. Reagents and conditions: (a) Cmpd. C, NH4OAc, EtOH, rt (29%); (b)
arylboronic acid, Pd(Ph3P)4, K2CO3, toluene/EtOH (4/1), 140 °C, MW (c) 4 N HCl,
dioxane, rt (33–44% in two steps).
Figure 4. Crystal structures of 5b in human MK2 (grey) and MK3 (cyan).12 Specific
amino acids are labeled.

T. Barf et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3818–3822
3821
50000
40000
30000
20000
10000
**
***
0
VEH s.c.
30
10
3
1
5b (s.c.)
treatment in mg/kg
Figure 5. Dose response curve of LPS-induced release of TNF
a
in rats after subcutaneous dosing of 5b.15 LPS was administered 1.5 h before extraction of blood. Statistically
significant differences are indicated as ⁄⁄(p <0.05) and ⁄⁄⁄(p <0.01).
consistent with other MK2:inhibitor structures, one pyrimidine N
forms a hydrogen bond with the main chain amide of Leu141/
Met121 (MK2/MK3 numbering), and the cyclic amide of 5b is in
contact with Lys93/Lys73 and Asp207/187. In the structures, the
basic amine is not involved in direct polar interactions with either
Glu145 or Glu190, as was shown for 2. The MK3:5b structure re-
veals an elaborate hydrogen-bonding network involving the pyr-
5b for further development was its lack of oral bioavailability
and this will be addressed in an accompanying paper.
Acknowledgments
The authors wish to acknowledge Gerard Vogel for generating
the PK data of 5b and Joost Uitdehaag for stimulating discussions.
role-NH, the basic amine, Glu145, and
a number of water
molecules. Due to the limited diffraction quality of MK2 crystals,
such interactions could not be observed in our MK2 structures.
However, on the basis of the otherwise high similarity between
the MK2 and MK3 structures these interactions most likely also ex-
ist in MK2. In addition to this water-mediated hydrogen-bond net-
work, the surface of both proteins is relatively acidic in the area of
the ribose and phosphate sub-pockets. This suggests that part of
the improvement in potency as a result of introducing a basic
amine in our compounds can be attributed to the formation of ben-
eficial electrostatic interactions.
In addition to improving potency, the introduction of a basic
amine in 5b also improved other properties including kinetic solu-
bility, lipophilicity, plasma protein binding, and stability in human
and rat liver microsomes (Table 1). On the other hand, absorption
properties were compromised as a result of the presence of basic
amines, see for example Caco-2 values (Table 1). This was also re-
flected by PK data in rats for 5 following per oral (p.o.), subcutane-
References and notes
1. (a) Sweeney, S. E. Nat. Rev. Rheumatol. 2009, 5, 475–477; (b) Hammaker, D.;
Firestein, G. S. Ann. Rheum. Dis. 2010, 69, i77.
2. Cheung, P. C. F.; Campbell, D. G.; Nebreda, A. R.; Cohen, P. EMBO J. 2003, 22,
5793.
3. Hegen, M.; Gaestel, M.; Nickerson-Nutter, C. L.; Lin, L.-L.; Telliez, J.-B. J.
Immunol. 2006, 177, 1913.
4. Ronkina, N.; Menon, M. B.; Schwermann, J.; Tiedje, C.; Hitti, E.; Kotlyarov, A.;
Gaestel, M. Biochem. Pharmacol. 2010, 80, 1915.
5. Kotlyarov, A.; Neininger, A.; Schubert, C.; Eckert, R.; Birchmeier, C.; Volk, H. D.;
Gaestel, M. Nat. Cell. Biol. 1999, 1, 94.
6. (a) Hanau, C. E. et al. WO patent 2004058176; (b) Anderson, D. R. et al. WO
patent 2004058762; (c) Anderson, D. R.; Meyers, M. J.; Vernier, W. F.; Mahoney,
M. W.; Kurumbail, R. G.; Caspers, N.; Poda, G. I.; Schindler, J. F.; Reitz, D. B.;
Mourey, R. J. J. Med. Chem. 2007, 50, 2647.
7. Structures of MK2:1 and MK2:2 were determined in collaboration with Sareum
and obtained using a construct comprising residues 41-364 with a C-terminal
His-tag. This construct was expressed in Escherichia coli and purified using
immobilized metal affinity, size exclusion, and ion exchange chromatography.
The protein was crystallized in space group
precipitant. Data sets with and were collected at 3.0 and 3.2 Å,
P 4132 using malonate as
1
2
respectively. The coordinates have been deposited with the Protein Data
Bank at Rutgers, the respective codes are 3R2Y and 3R3O.
ous (s.c.) and intravenous (i.v.) dosing. The rat AUC was 0.016
after p.o. dosing at 10 mol/kg, 1.26 M h after i.v. dosing at
mol/kg, and 2.71 M h after s.c. dosing at 10 mol/kg. These
data showed that the compound does not get absorbed via the oral
route, whereas availability is moderate following subcutaneous
administration. Oral bioavailability thus became one of the main
issues for this chemical series in lead optimization, which will be
addressed in an accompanying paper.11
lM h
l
l
8. (a) Underwood, K. W.; Parris, K. D.; Federico, E.; Mosyak, L.; Czerwinski, R. M.;
Shane, T.; Taylor, M.; Svenson, K.; Liu, Y.; Hsiao, C. L.; Wolfrom, S.; Maguire, M.;
Malakian, K.; Telliez, J. B.; Lin, L. L.; Kriz, R. W.; Seehra, J.; Somers, W. S.; Stahl,
M. L. Structure 2003, 11, 627; (b) Hillig, R. C.; Eberspaecher, U.; Monteclaro, F.;
Huber, M.; Nguyen, D.; Mengel, A.; Muller-Tiemann, B.; Egner, U. J. Mol. Biol.
2007, 369, 735.
9. MK2 enzyme activity is measured using an IMAP assay (Immobilized Metal
Assay for Phosphochemicals-based coupled assay). Test compounds in DMSO
(final concentration is assay is 0.1% DMSO) and MK2 enzyme (peptide 46-end
(Millipore) 25 mU/mL final concentration in assay) are pre-incubated 30 min at
room temperature, before adding Fluorescin labeled substrate peptide (Fluo-
betaA-11A NeoMPS, final substrate peptide concentration is 50 nM). Kinase
4
l
l
l
The potential of 5b to inhibit production of TNF
in rats following s.c. and i.v. administration and an LPS challenge.15
Compound 5b inhibited 85% and 92% of the TNF release in rats
a was analyzed
a
assay is started by adding ATP (final ATP concentration is 1 lM (=Km ATP)).
following s.c. (30 mg/kg) and i.v. (12.5 mg/kg) administration,
respectively. A dose response curve was also conducted after s.c.
Following incubation for 2 h at room temperature the enzyme reaction is
stopped by adding IMAP Progressive Binding Solution (according to suppliers
(Molecular Devices) protocol). After 60 min incubation at room temperature in
the dark the FP signal is read. Each concentration is tested in two independent
experiments with duplicates in each experiment.
dosing. Fig. 5 shows a dose-dependent reduction of TNF
a release
with statistically significant inhibition at 10 and 30 mg/kg, of 54%
and 89% respectively.
10. Revesz, L.; Schlapbach, A.; Aichholz, R.; Dawson, J.; Feifel, R.; Hawtin, S.;
Littlewood-Evans, A.; Koch, G.; Kroemer, M.; Möbitz, H.; Scheufler, C.; Velcicky,
J.; Huppertz, C. Bioorg. Med. Chem. Lett. 2011, 20, 4719.
11. Kaptein, A.; Oubrie, A.; Zwart, E. de; Hoogenboom, N.; Wit, J. de; Kar, B. van de;
Hoek, M. van; Vogel, G.; Kimpe, V. de; Schultz-Fademrecht, C.; Borsboom, J.;
Zeeland, M. van; Versteegh, J.; Kazemier, B.; Roos, J. de; Wijnands, F.; Dulos, J.;
Jaeger, M.; Leandro-Garcia, P.; Barf, T. Bioorg. Med. Chem. Lett. 2011.
doi:10.1016/j.bmcl.2011.04.016.
In conclusion, MK2 inhibitor 5b was identified using structure-
based drug design. The compound exhibits good potency in bio-
chemical and in cell-based assays, including primary human cells
(human PBMCs). It appeared clean in many toxicity assays and
showed efficacy in short-term, preclinical in vivo models such as
LPS-induced release of the cytokine TNFa. The major issue with
12. Cheng, R.; Felicetti, B.; Palan, S.; Toogood-Johnson, I.; Scheich, C.; Barker, J.;
Whittaker, M.; Hesterkamp, T. Protein Sci. 2010, 19, 168.

3822
T. Barf et al. / Bioorg. Med. Chem. Lett. 21 (2011) 3818–3822
13. In house structures of MK2 and MK3 bound to 5b: The MK2 structure was
obtained using a construct comprising residues 41-364 with a cleavable N-
terminal His-tag. MK2 protein was purified using immobilized metal affinity,
size exclusion, ion exchange chromatography and removal of the His-tag with
(Fetal Bovine Serum) in 96 well plate. Cells are left to rest for at least 18 h in a
humidified atmosphere at 5–7% CO2 at 37 °C, prior to incubation with test
compounds. Dilutions of test compounds in DMSO (0.1% DMSO final
concentration in assay) are added to the cells. Following a 30 min incubation
Thrombin. This construct crystallized in space group
molecules per asymmetric unit. The structure of MK2:5b was determined at
2.9 Å resolution. MK3 protein was purchased from Evotec and crystallized with
P
212121 with 12
LPS is added (final LPS concentration 10
5–7% CO2 at 37 °C in humidified atmosphere, culture medium is collected for
the measurement of TNF . The amount of TNF in the culture medium is
quantified by ELISA using HuTNF assay from Cytoset according to protocol.
lg/mL). Following a 4 h incubation at
a
a
1
using protocols described in Ref. 12. MK3:1 crystals were soaked and
a
exchanged with 5b. The structure of MK3:5b was determined at 2.1 Å
resolution. The coordinates have been deposited with the Rutgers Protein
Data Bank, with accession codes 3R2B and 3R1N.
15. Rats are treated with different compound doses. Control animals are treated
with placebo. 1 h after compound treatment, rats are intraperitoneally injected
with LPS (0.5 mg/kg). 1.5 h after LPS injection, serum is collected for the
14. THP1 cells are seeded at 1.5 Â 105 cells per well in DMEM F12 (Gibco)
measurement of TNFa level. The amount of TNFa in the serum is quantified by
supplemented with Penicilin/Streptomycin (80 U/mL; 80 lg/mL) and 10% FBS
ELISA.