A novel class of potent NF-κB signaling inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.08.022

A novel class of NF-κB pathway signaling inhibitors was discovered by virtual screening, medicinal chemistry, and QSAR analysis. An initial set of compounds inhibited NF-κB signaling in a whole cell reporter gene assay in the micro-molar range. Activity was improved step by step by medicinal chemistry to yield nano-molar signaling inhibitors.

Full text of DOI:10.1016/j.bmcl.2007.08.022

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5858–5862
A novel class of potent NF-jB signaling inhibitors
*
Johann Leban, Marcel Baierl, Jan Mies, Viola Trentinaglia, Sandra Rath,
Kerstin Kronthaler, Kristina Wolf, Astrid Gotschlich and Markus H. J. Seifert
4
SC AG, Am Klopferspitz 19a, 82152 Martinsried, Germany
Received 10 July 2007; revised 9 August 2007; accepted 10 August 2007
Available online 15 August 2007
Abstract—A novel class of NF-jB pathway signaling inhibitors was discovered by virtual screening, medicinal chemistry, and
QSAR analysis. An initial set of compounds inhibited NF-jB signaling in a whole cell reporter gene assay in the micro-molar range.
Activity was improved step by step by medicinal chemistry to yield nano-molar signaling inhibitors.
ꢀ
2007 Published by Elsevier Ltd.
The important role of the nuclear factor jB (NF-jB)
signaling pathway in cancer, immunology, and inflam-
N
S
HN
O
N
1
N
N
mation is well established. For example, in antigen
stimulated lymphocytes, NF-jB controls the expression
of multiple genes essential for survival, proliferation,
N
N
N
O
2
,3
and effector function. Additionally, some viruses take
Compound 1. Proteasome IC50 0.58 lM, NF-jB inhibi-
tion 30% at 20 lM, PBMC proliferation IC50 37 lM.
advantage of the NF-jB signaling pathway of the host
4
,5
cell for their own replication. The approach of targeting
host pathways for anti-viral therapy holds the promise
of a low potential for the development of resistant
strains. Finally, it has been reported that NF-jB down-
regulation can sensitize cells to the anti-viral effects of
1
1
Virtual screening for proteasome inhibitors resulted in
the discovery of compounds, for example, compound 1,
1
2
with inhibitory activity for the proteasome, NF-jB
signaling, and PBMC proliferation.
1
3
14
6
,7
interferon.
The mechanism of action of some known drugs or folk
medicines is presumed to function via the inhibition of
NF-jB signaling pathway. In general, the NF-jB sig-
Br
S
i,ii
O
Boc
N
8
O
NH2
O
naling pathway consists of a complex set of biochemical
principles, such as receptors, protein adapters, protein
kinases and phosphoprotein binding proteins, DNA
binding proteins, and transporter proteins. A series of
compounds targeting various enzymes in this pathway
S
N
R
N
iii,iv
Boc
N
H2N
OH
N
H
O
S
8
,9
have been developed. Other compounds, such as pro-
teasome inhibitors, were shown to interfere with the NF-
jB signaling pathway by inhibiting degradation of
H
N
HN
N
N
v
1
0
O
HN
IjB.
R
S
N
R1
H
N
N
N
O
HN
R
Keywords: NF-jB; Nuclear factor jB; Signaling; Inhibitor; Anti-viral;
Anti-inflammatory.
Scheme 1. Reagents and conditions: (i) DMF, heat; (ii) NaOH; (iii)
1
*
Corresponding author. Tel.: +49 89 700763 0; fax: +49 89 700763
8; e-mail: leban@4sc.com
HBTU, DIEPA, DMF, 75 ꢁC; HCl, Dioxan, DCM, 2 h; R -Cl,
2
DIEPA, 80 ꢁC, 3 h, DMF.
0
960-894X/$ - see front matter ꢀ 2007 Published by Elsevier Ltd.
doi:10.1016/j.bmcl.2007.08.022

J. Leban et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5858–5862
5859
Accordingly, the piperidinyl-thiazole scaffold was ex-
plored by medicinal chemistry. The synthesis proceeded
micro-molar activity with this series. The structure–
activity relationships seen in the cell-based NF-jB
reporter gene assay, however, could not be explained
consistently by proteasome inhibition. For example,
some compounds showed considerably higher activities
in the reporter gene assay compared to the proteasome
enzyme assay (e.g., compound 15: NF-jB EC50
0.3 lM, proteasome IC50 > 3 lM) while other com-
pounds were inactive in the NF-jB assay but active in
the proteasome assay (e.g., compound 9: NF-jB
1
5
as published in a patent (see Scheme 1). This proce-
dure allows the synthesis of analogs modified at the N-
1
6
and C-terminal of the molecule. All building blocks
were from commercial sources. Some of the 2-amino-
benzimidazoles were synthesized from the appropriate
commercial o-phenylenediamine by the reaction with
1
7
cyanogen-bromide. Heterocycles for the N-terminal
modification were from commercial source. The hetero-
cycle for compound 22 was synthesized according to
Ref. 18 and the heterocycle for compound 11 according
to Ref. 19.
EC50 ꢀ 20lM, proteasome IC 0.48 lM). Additionally,
50
even some compounds (e.g., compound 13) with activity
in the proteasome enzyme assay did not show activity in
a secondary proteasome assay, which uses native protea-
2
0
From a series of C-terminal modifications substituted
phenyl-piperazine analogs were found to have superior
activity (Table 1, compounds 2–10). In this set of com-
pounds low micro-molar activities in the NF-jB repor-
ter gene assay were obtained. But, in spite of an effort
with many analogs we were not able to reach lower than
some extracted from cells. Due to the particular
importance of the NF-jB signaling pathway further ef-
forts were directed to optimize the piperidinyl-thiazole
scaffold with respect to inhibition of this pathway.
In order to guide the optimization of the compounds
a QSAR model was used. The QSAR model of the
NF-jB activities was built using 28 MOE descriptors
and ridge regression (k = 1, 20 effective degrees of
freedom). This model was incrementally refined as
more experimental data became available and finally
Table 1. Phenyl-piperazine analogs
2
reached a correlation coefficient of R = 0.84, cross-
validated Q = 0.58, and a root mean square error of
S
N
CF3
2
R1
N
N
N
prediction (RMSEP) of 0.34 log units as validated
with a leave-multiple-out bootstrapping procedure.
This is close to the experimental error (SD) of 0.2
log units as estimated from multiple measurements
of two selected compounds (n = 6 and n = 7). Com-
pounds that showed a negative effect on cell viability
were not considered in this model (Fig. 1).
O
1
R
Compound
NF-jB EC50 (lM)
H
N
F
2
3.4
O
O
N
N
Selection of building blocks with respect to availability
and predicted NF-jB activity resulted in a series of com-
pounds, which is depicted in Table 2 with compounds
3
N
O
O
N
4.7
4
5
5.0
7.2
2.0
1.5
1.0
0.5
0.0
N
O
O
H
N
O
6
7
NC
8.5
O
S
O
S
ꢀ20
Cl
-0.5
-1.0
-1.5
O
S
O
8
9
ꢀ20
F
F3C
ꢀ20
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
O
log (EC50/μM) experimental
N
10
N
>20
N
Figure 1. QSAR of NF-jB activity of 71 compounds as determined in
the reporter gene assay.
N

5860
J. Leban et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5858–5862
Table 2. Amino-benzimidazole analogs
O
S
R1
N
N
H
N
N
Y =
R2
N
H
O
1
R
2
R
Compound
NF-jB EC50 (lM)
N
N
N
11
Y
0.1
N
1
1
1
2
3
4
Y
Y
Y
0.1
0.2
0.3
N
N
N
N
S
N
N
NH
O
N
O
N
1
1
1
5
6
7
Y
Y
Y
0.3
0.3
0.3
N
O
N
O
O
N
O
N
N
N
H
N
1
8
9
0.3
0.3
N
S
1
N
Y
O
N
O
O
O
N
H
N
2
0
1
0.4
0.5
N
N
S
S
N
H
N
N
N
2
S
N
N
22
N
N
Y
0.5
O
O

J. Leban et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5858–5862
5861
Table 2 (continued)
1
R
2
R
Compound
NF-jB EC50 (lM)
N
N
2
3
4
Y
Y
0.5
F3C
O
N
O
N
N
2
0.8
F3C
O
O
2
2
2
5
6
7
Y
Y
0.8
0.9
3.0
O
N
N
N
N
O
F3C
H
N
O
N
Table 3. Inhibition of human PBMC proliferation
Acknowledgments
Compound
IC50 (lM)
We thank Stefan Strobl and Daniel Vitt for helpful dis-
cussions and support, and Roland Baumgartner for
measuring the proteasome enzyme data.
3
15
1
1
2
2
3
2
3.6
2.2
2.5
References and notes
1
2
3
4
. Karin, M.; Greten, R. Nat. Rev. Immunol. 2005, 5, 749.
. Schulze-Luehrmann, J.; Ghosh, S. Immunity 2006, 25, 701.
. Thome, M. Nat. Rev. Immunol. 2004, 4, 348.
. Nimmerjahn, F.; Dudziak, D.; Dirmeier, U.; Hobom, G.;
Riedel, A.; Schlee, M.; Staudt, L. M.; Rosenwald, A.;
Behrends, U.; Bornkamm, G. W.; Mautner, J. J. Gen.
Virol. 2004, 85, 2347.
1
1–27. The replacement of the phenyl-piperazino moiety
by an amino-benzimidazole was tolerated but did not
improve activity considerably (compound 27). On the
average the substitution of the amino-benzimidazole
by a phenylketo residue resulted in a substantial
improvement. (e.g., compound 25 vs 27, compound 13
vs 18). At the N-terminus of the compounds different
aromatic substituents are tolerated without effecting
activity, however heterocycles similar to adenine mimics
did increase potency. Finally, potent inhibitors of the
NF-jB signaling pathway were identified.
5. Wurzer, W. J.; Ehrhardt, C.; Pleschka, S.; Berberich-
Siebelt, F.; Wolff, T.; Walczak, H.; Planz, O.; Ludwig, S.
J. Biol. Chem. 2004, 279, 30931.
6
. Wei, L.; Sandbulte, M. R.; Thomas, P. G.; Webby, R. J.;
Homayouni, R.; Pfeffer, L. M. J. Biol. Chem. 2006, 281,
1
1678.
7
. Pfeffer, L. M.; Kim, J. G.; Carrigan, D. J.; Baker, D. P.;
Wei, L.; Homayouni, R. J. Biol. Chem. 2004, 279, 31304.
. Merfort, I. Expert Opin. Ther. Pat. 2006, 16, 797.
To solidify the assumption that these compounds have
immunosuppressive activity we tested a subset of com-
pounds in a PBMC proliferation assay. Due to the
phytohemagglutinin stimulus used in the assay the
measured effect can be attributed mainly to T-cells.
A significant suppressive effect was detected (Table
8
9. http://people.bu.edu/gilmore/nf-kb/inhibitors/.
10. Adams, J.; Palombella, V. J.; Elliott, P. J. Invest. New
Drugs 2000, 18, 109.
1
1. Virtual screening: proteasome X-ray structure (PDB ID
1IRU) was used for virtual screening due to the sequence
identity within the chymotryptic subpocket of the bovine
proteasome located on the L and M chain and the human
proteasome X and HC5 chains. Protein–ligand docking
was used for virtual screening with the binding pocket
defined around the catalytically active threonine (THR1)
containing the subpockets S1, S2, and S3. Virtual high-
throughput screening was done using the docking engine
of 4SCan/ProPose (Seifert, M. H.; Wolf, K.; Vitt, D. DDT
Biosilico 2003, 1, 143.; Seifert, M. H.; Schmitt, F.; Herz,
1
4
3
).
In conclusion, we have discovered a novel class of po-
2
1
tent NF-jB signaling inhibitors. Such compounds
have the potential to be useful for the treatment of
inflammatory and autoimmune diseases, for anti-viral
therapy and interferon co-therapy, and could lead to a
novel concept of host cell pathway modulation toward
an anti-viral therapy.

5862
J. Leban et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5858–5862
T.; Kramer, B. J. Mol. Model. 2004, 10, 342) on 4SC’s
database of five million chemical compounds. The 5000
best docking hits were submitted to filtering: a distance
13. NF-jB reporter gene assay: The NF-jB reporter gene
assay was prepared with A549-NF-jB-SEAP cell line
(CCS cell culture service, Hamburg, Germany) according
to manufacturer’s instructions. In short, A549 cells stably
transfected with pNF-jB-SEAP reporter gene plasmid
˚
cutoff of 10 A to THR1 was applied, molecular properties
(
molecular weight, topological polar surface area, number
4
of rotational bonds) were filtered, unwanted functional
groups were omitted, and finally a visual inspection of the
docked structures yielded a ranking list. A representative
set of 236 compounds was selected for purchasing and
biochemical testing.
were plated at 2 · 10 /well and allowed to attach over-
night. The cells were subsequently incubated for 5 h with
described compounds at 100, 30, 10, 3, 1, 0.3, 0.1, and
0 lM and then stimulated with 10 ng/ml TNF-a for 22 h.
The supernatant of the cell was analyzed for SEAP activity
using a chemiluminescent SEAP reporter gene assay
(Roche, Mannheim, Germany), and a cell viability assay
was prepared using a CellTiter-BluTM Cell Viability
Assay (Promega, Mannheim, Germany). For each con-
centration of the compound four replicates were
measured.
1
2. 20S proteasome enzyme assay: (Stein, R. L.; Melandri, F.;
Dick, L. Biochemistry 1996, 35, 3899.) compounds were
characterized by monitoring the inhibitory effect on the
chymotryptic activity of the human 20S proteasome
(
Biomol). The assay was performed using a Tecan Ultra
plate reader and the fluorogenic substrate Suc-LLVY-
AMC (Bachem). In a black 96-well polypropylene plate,
14. Magaud, J. P.; Sargent, I.; Mason, D. Y. J. Immunol.
Methods 1988, 106, 95; Leban, J.; Kralik, M.; Mies, J.;
Gassen, M.; Tentschert, K.; Baumgartner, R. Bioorg.
Med. Chem. Lett. 2005, 15, 4854.
15. Know, P.; O’Sullivan, M.; Lentfer, H. Patent application
WO 2004/058750, 2004.
2
ll of compound, dissolved in 100% DMSO, was mixed
with 50 ll substrate solution (25 mM HEPES, pH 7.5,
.5 mM EDTA, and 75 lM Suc-LLVY-AMC). Reaction
was initiated by adding 150 ll proteasome solution (1.3 lg
0S proteasome in 25 mM HEPES, pH 7.5, 0.5 mM
0
2
EDTA, and 0.033% (w/v) SDS, pre-incubated for 10 min
at room temperature). Substrate hydrolysis was followed
by fluorescence spectroscopy (excitation wavelength:
16. All compounds were characterized by MS and NMR
(300 MHz) and exhibited satisfactory properties.
17. Lionel, J. J. Med. Chem. 1963, 6, 601.
3
3
60 nm; emission wavelength: 465 nm) for 20 min at
0 ꢁC and signals were calculated as relative fluorescence
18. Ohnishi, H.; Kosuzume, H.; Mizota M.; Suzuki, Y.;
Mochida, E. European patent EP 1984/0116421,
1984.
19. Fusaka, T.; Ujikawa, O.; Kajiwara, T.; Tanaka, Y, Patent
WO9711075, 1997.
TM
20. Proteasome-Glo Cell-Based Assay Technical Bulletin
(TB346), 2006 (Promega, Mannheim, Germany).
21. Leban, J.; Schmitt, H.; Wolf K, Pegoraro S, Wuzik A. US
Patent application US 2006/247253, 2006.
units (RFU) per minute. For the determination of the IC50
values (concentration of inhibitor required for 50%
inhibition) eight inhibitor concentrations were applied.
Data points were recorded in triplicate on single measure-
ment day. To obtain dose–response curves and IC50 values
data were fitted to a four parameter logistic function using
SigmaPlot.