Thiazolidine-2,4-dione-based irreversible allosteric IKK-β kinase inhibitors: Optimization into in vivo active anti-inflammatory agents

Article abstract of DOI:10.1016/j.ejmech.2019.111955

Selective kinase inhibitors development is a cumbersome task because of ATP binding sites similarities across kinases. On contrast, irreversible allosteric covalent inhibition offers opportunity to develop novel selective kinase inhibitors. Previously, we

Full text of DOI:10.1016/j.ejmech.2019.111955

Journal Pre-proof
Thiazolidine-2,4-dione-based irreversible allosteric IKK-β kinase inhibitors:
Optimization into in vivo active anti-inflammatory agents
Ahmed Elkamhawy, Nam youn Kim, Ahmed H.E. Hassan, Jung-eun Park, Sora Paik,
Jeong-Eun Yang, Kwang-Seok Oh, Byung Ho Lee, Mi Young Lee, Kye Jung Shin, Ae
Nim Pae, Kyung-Tae Lee, Eun Joo Roh
PII:
S0223-5234(19)31107-9
DOI:
https://doi.org/10.1016/j.ejmech.2019.111955
EJMECH 111955
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 6 August 2019
Revised Date: 5 December 2019
Accepted Date: 5 December 2019
Please cite this article as: A. Elkamhawy, N.y. Kim, A.H.E. Hassan, J.-e. Park, S. Paik, J.-E. Yang,
K.-S. Oh, B.H. Lee, M.Y. Lee, K.J. Shin, A.N. Pae, K.-T. Lee, E.J. Roh, Thiazolidine-2,4-dione-based
irreversible allosteric IKK-β kinase inhibitors: Optimization into in vivo active anti-inflammatory agents,
European Journal of Medicinal Chemistry (2020), doi: https://doi.org/10.1016/j.ejmech.2019.111955.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2019 Published by Elsevier Masson SAS.

Graphical Abstract
7a
Irreversible allosteric IKK-
IC50 = 200 nM
β
inhibitor
Kinact = 0.013 min-1
TNF- IC50 = 6.27
PGE2 IC50 = 9.83
α
µM
µ
M
Plasma stability = 73.6% after two hours
Microsomal stability t1/2= 1 hour
In vivo protection against septic shock = 80% survival at 36 hours post injection

Thiazolidine-2,4-dione-based irreversible allosteric IKK-β kinase inhibitors:
Optimization into in vivo active anti-inflammatory agents
Ahmed Elkamhawya,b,1, Nam youn Kima,1, Ahmed H.E. Hassanc,1, Jung-eun Parka, Sora Paika,d, Jeong-
Eun Yanga, Kwang-Seok Ohe, Byung Ho Leee, Mi Young Leee, Kye Jung Shinf, Ae Nim Paeg,h, Kyung-Tae
Leei, Eun Joo Roha,h*
a Chemical Kinomics Research Center, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea.
b Department of Pharmaceutical Organic Chemistry, Faculty of Pharmacy, Mansoura University, Mansoura 35516, Egypt.
c Department of Medicinal Chemistry, Faculty of Pharmacy, Mansoura University, Mansoura 35516, Egypt.
d Department of Fundamental Pharmaceutical Sciences, College of Pharmacy, Kyung Hee University, Seoul, 02447, Republic of Korea.
e Therapeutics & Biotechnology Division, Korea Research Institute of Chemical Technology, 141 Gajeong-ro, Yuseong, Daejeon 34114,
Republic of Korea.
f Integrated Research Institute of Pharmaceutical Sciences, College of Pharmacy, The Catholic University of Korea, Bucheon-si, Gyeonggi-
do 14662, Republic of Korea.
g Convergence Research Center for Diagnosis, Treatment and Care System of Dementia, Korea Institute of Science and Technology (KIST),
Seoul 02792, Republic of Korea.
h Division of Bio-Medical Science & Technology, KIST School, Korea University of Science and Technology, Seoul 02792, Republic of Korea
i Department of Life and Nanopharmaceutical Sciences, Kyung Hee University, Seoul 02447, Republic of Korea.
1 These authors contributed equally to this work
*Corresponding author:
Eun Joo Roh: r8636@kist.re.kr

Abstract
Selective kinase inhibitors development is a cumbersome task because of ATP binding sites
similarities across kinases. On contrast, irreversible allosteric covalent inhibition offers
opportunity to develop novel selective kinase inhibitors. Previously, we reported thiazolidine-
2,4-dione lead compounds eliciting in vitro irreversible allosteric inhibition of IKK-β. Herein,
we address optimization into in vivo active anti-inflammatory agents. We successfully
developed potent IKK-β inhibitors with the most potent compound eliciting IC50 = 0.20 µM.
Cellular assay of a set of active compounds using bacterial endotoxin lipopolysaccharide
(LPS)-stimulated macrophages elucidated significant in vitro anti-inflammatory activity. In
vitro evaluation of microsomal and plasma stabilities showed that the promising compound
7a is more stable than compound 7p. Finally, in vivo evaluation of 7a, which has been
conducted in a model of LPS-induced septic shock in mice, showed its ability to protect mice
against septic shock induced mortality. Accordingly, this study presents compound 7a as a
novel potential irreversible allosteric covalent inhibitor of IKK-β with verified in vitro and in
vivo anti-inflammatory activity.
Keywords:
IKK-β modulators; NF-κB signaling pathway; Thiazolidine-2,4-diones; Allosteric modulation;
Anti-inflammatory.

1. Introduction
Encompassing more than five hundred members, kinases are the third largest class of
enzymes [1]. More than 30% of the entire set of proteins expressed in humans estimated to
undergo phosphorylation kinases [2]. Irregular hyperactivity of certain kinases has been
characterized in some pathological conditions including tumors, neurological and
inflammatory diseases. Therefore, development of the specific kinase inhibitors has become a
hot topic in drug discovery programs [3-8]. As most of the devoted efforts have focused on
targeting the ATP binding pocket, which bears similarities across different kinases,
development of a selective inhibitor is a challenge. Poorly selective ATP competitive
molecules could face limited clinical application [9]. Lately, development of allosteric kinase
modulators gained traction among researchers in the field as targeting the more variable sites
of kinases might afford more efficiently selective ligands [10-12]. Also recently, more
evidences have accumulated documenting the possibility of developing covalent inhibitors
that could be more efficient than the classical reversible inhibitors, yet, are devoid from the
previously doubted toxicities [13-18]. In fact, the concept of covalent inhibition has been
extended to the development of covalent allosteric inhibitors that has been claimed to offer
higher level of selectivity and potency [19-21].
The “nuclear factor kappa-light-chain-enhancer of activated B cells” (NF-κB) is a
transcription factor that induces the expression of more than 500 genes encompassing several
pro-inflammatory genes, including cytokines (TNF-α, IL-1β and IL-6), chemokines (CCL5,
CCL20, CXCL8 and CXCL10) in addition to proinflammatory enzymes including inducible
nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX2) [5, 22-25]. In addition, NF-κB-
triggered genes expression invokes immunological responses, cells proliferation, and
apoptosis [26, 27]. The canonical NF-κB signaling pathway is activated by the kinase IKK-β

(inhibitor of nuclear factor kappa-B kinase subunit beta) which phosphorylates IκB (inhibitor
of NF-κB) resulting in release and translocation of NF-κB into the nucleus [28]. Because
IKK-β is a bottleneck for activation of NF-κB signaling pathway, its inhibition might be a
successful strategy for treatment of acute and chronic inflammatory diseases, as well as,
cancer and autoimmune diseases [28-31]. In addition to IκB, IKK-β phosphorylates other
substrates such as TSC1, SNAP23 and Foxo3a [32, 33]. Consequently, it elicits NF-κB-
independent signals in inflammation mediated tumor angiogenesis, mast cell degranulation
and cell cycle arrest [34-36].
Several literatures report development of inhibitors of NF-κB signaling pathway including
IKK-β inhibitors of diverse chemical scaffolds [26, 31]. As far as we know, most of the
developed IKK-β inhibitors are reversible competitive inhibitors; for example, compound 1
(Fig. 1) developed by the Bayer pharmaceutical company [37]. Although irreversible
allosteric inhibitors offer advantages over the reversible classical ATP competitive inhibitors
as mentioned previously, reports of even reversible allosteric IKK-β inhibitors are limited
(e.g. BMS-345541 (2); a non-selective allosteric inhibitor of both IKK-α and IKK-β [24]).
There is an obvious need to develop small molecules as covalent allosteric inhibitors of
IKK-β. A recent report revealed that the natural product (+)-ainsliadimer (3) covalently
inhibits IKK-β at an allosteric site [38]. In another report, covalent binding of the natural
product isoliquiritigenin (4) was also found to suppress the activity of T lymphocytes [39].
Recently, Park et al. reported 1,3,5-triazine-2,4-diamine derivative (5) as an IKK-β inhibitor
hit compound and used it as starting point to develop orthosteric IKK-β inhibitors [40]. Their
report presented also a 2-imino-4-thiazolidin-4-one derivative as a hit compound possessing
high IC50 value of 97.4 µM. In a recent report, we have undertaken the task of development
of the latter hit compound into thiazolidine-2,4-dione derivative 6 which elicited a potential

in vitro covalent allosteric inhibition of IKK-β but still its IC50 value is in the micromolar
range [41]. Despite further attempt to optimize the kinase inhibitory activity, the latter
compounds were not advanced for further cell-based or in vivo evaluation [42]. In this report,
we address optimization of the lead compound 6 into more potent irreversible allosteric IKK-
β inhibitor with potential in vivo anti-inflammatory activity.
O
NH
OH
O
O
H
H
HO
OH
OH
O
H
CN
H
N
OH
N
N
N
NH2
NH2
O
H
HO
O
N
H
H
2
3
1
4
O
NO2
NH2
O
O
S
N
NH2
N
N
N
O
N
4 O
NH2
N
5
6
Fig. 1. Reported inhibitors compounds for IKK-β
In this study, as shown in Fig. 2, we explored variable lengths of the alkyloxy chain linker in
addition to studying the outcome of variation of the linker attachment point between para and
meta positions of the phenyl ring attached to the nitrogen atom of the thiazolidine-2,4-dione
scaffold. Congruent with reported influence of cycloalkylamines modification such as
piperazines on bioactivity [43], we found in our previous report that the terminal
N-substituent on piperazine ring elicits a crucial impact on the elicited activity with the all
carbon substituents, such as N-cyclopropylmethyl, eliciting higher activity than N-
substituents possessing the more polar oxygen or sulfur atom. We fixed this terminal N-
substituent to methyl, which is a one-carbon substituent. Since the structure of IKK-β

incorporates an important cysteine residue (Cys46) which exists in the vicinity of an
allosteric site located close to the catalytic ATP pocket (protein databank code: 3QA8) [44],
the observed irreversible allosteric inhibition might result from Michael addition of thiol
group to the double bond of the benzylidene moiety of compound 6. Because variable
substituents attached to the benzylidene moiety might elicit different electron push-pull,
which would affect the Michael addition of thiol group of cysteine, we prepared compounds
7 possessing variable substituents on the phenoxy ring attached to the benzylidene moiety.
Accordingly, we would like to report our promising results.
Fig. 2. Design of the novel thiazolidine-2,4-dione derivatives (7)
2. Results and discussion
2.1. Chemistry
As illustrated in Scheme 1, a straightforward synthesis of the target compounds (7) was
performed in six steps starting from 4-aminophenol (8a) or 3-aminophenol (8b). Thus, amide
coupling with chloroacetyl chloride yielded acetamide derivatives 9. Nucleophilic
displacement of chlorine atom of acetamide derivatives 9 with thiocyanate followed by
cyclization afforded the 2-iminothiazolidin-4-one derivatives 10 [45]. Acidic hydrolysis
afforded the desired thiazolidine-2,4-dione intermediates 11 [46]. O-Alkylation of the
phenolic hydroxyl group with 1,n-diiodoalkanes allowed installation of alkyl chain linker to
yield derivatives 12. N-Alkylation of 1-methylpiperazine with the prepared derivatives 12
provided the corresponding thiazolidine-2,4-dione derivatives 13. Lastly, Knoevenagel

condensation with various benzaldehyde derivatives afforded the designed compounds (7,
Table 1) [47].
Scheme 1. Reagents and reaction conditions: (a) chloroacetyl chloride, DCM, 0 ºC, 1.5 h, rt, 3 h; (b) KSCN,
acetone, 65 ºC, 7 h; (c) 2% HCl aqueous solution, 100 ºC, 7 h; (d) diiodoalkane, K2CO3, MeCN, 95 ºC, 7 h; (e)
1-methylpiperazine, K2CO3, MeCN, rt, 9 h; (f) appropriate aldehyde derivative, NaOAc, AcOH, 110 ºC, 16 h.
2.2. Biological evaluation
2.2.1. In vitro IKK-β induced phosphorylation modulation
As previously mentioned, all of the designed, synthesized and evaluated final compounds
have m- or p-hydroxyphenyl moiety at 3-position of the thiazolidine-2,4-dione scaffold. The
hydroxyphenyl moiety is O-alkylated with an alkyl chain tether to link it to N-
methylpiperazine moiety. The alkyl linker was varied in length (n = 3 or 5) while the
introduced substituents on the phenoxyphenyl moiety attached to the 5-exomethylene group
on the thiazolidine-2,4-dione ring have variable electronic characters as shown in Table 1. In
vitro evaluation of the IKK-β inhibitory activity of the synthesized 17 final compounds were
performed employing a 5-FAM (5-carboxyfluorescein) labeled IκBα-derived polypeptide (5-

FAMGRHDSGLDSMK-NH2; R7574, MDS Analytical Technologies) as a substrate for
IKK-β in an IMAP® based fluorescence assay [48]. To secure high sensitivity and accuracy
for the conducted in vitro assay, it was carried out as time-resolved Förster resonance energy
transfer assay (TR-FRET assay) [49]. As the percent inhibition and IC50 of irreversible
inhibitors increase by time, the measurements were recorded after two hours of interaction of
10 µM concentrations of the tested compound with IKK-β to permit the inactivation reaction
to approach completion.
Table 1. Results of IKK-β assay using the synthesized new ligands (7a−7q).
IKK-β
assay
IKK-β
assay
Linker
Position
Linker
Position
Cpd
n
R1
Cpd
n
R1
IC50 (µM)a
IC50 (µM)a
3
3
5
5
5
5
p
p
p
p
p
p
CN
F
0.20
0.67
0.60
0.70
0.43
0.70
5
5
5
5
5
5
m
m
m
m
m
m
F
Cl
6.06
2.61
2.67
2.73
2.43
8.07
7a
7b
7c
7d
7e
7f
7j
7k
7l
F
Br
Cl
CH3
OCH3
NH2
7m
7n
7o
Br
OCH3
5
p
CH3
0.87
5
5
m
m
NO2
CN
0.22
0.84
7g
7p
7q
5
5
p
p
NH2
2.61
1.71
7h
7i
NHCOCH3
a IC50 values (µM) exhibited by the newly synthesized compounds
As shown in Table 1, we first explored the biological activity of compounds whose linkers
were attached to the para position. Analogs 7a and 7b possessing the three carbons linker
were significantly active molecules. While compound 7b possessing the electronically
amphiphilic 4-fluoro substituent elicited an excellent submicromolar IC50 of 0.67 µM,
compound 7a possessing the electronegative 4-cyano group was more potent eliciting IC50 of
0.20 µM. Extending the length of the alkyl chain linker of compound 7b by two carbons
provided compound 7c with almost similar potency (Table 1). While variation of the
substituent to include the electronically amphiphilic 4-bromo substituent in derivative 7e

resulted in more potent compound than derivatives 7f and 7g which have the electron
donating 4-methoxy or 4-methyl moieties, compound 7d possessing 4-chloro substituent
elicited similar potency to compound 7f. However, more decrement of the potency was
elicited in derivatives 7h and 7i possessing the electron donating 4-amino or 4-acetamido
moieties, respectively. With IC50 value of 0.43 µM, the 4-bromo derivative 7e was the most
potent among compounds possessing a linker of 5 carbon atoms, however, it has the double
IC50 of compound 7a possessing 3 carbon atoms linker.
Next, we explored the outcome of displacing the linker from para to meta position via
preparation and evaluation of compounds 7j–q. Comparing the activities of meta and para
linked compounds possessing the same substituent on exomethylene moiety unveiled an
interesting activity trend. As Table 1 illustrates, derivative 7o possessing 4-amino moiety
showed relatively higher micromolar IC50 value of 8.07 µM while the corresponding
derivative 7h with para attached linker showed IC50 value of 2.61 µM, which was the least
potent among those compounds possessing the para attached linker. In addition, displacement
of the linker’s attachment point to meta position in compounds 7j, 7k and 7l, possessing the
electronically amphiphilic 4-halo substituents resulted in much less potent than the
corresponding derivatives having the linker attached to the para position. Thus, compounds
7j, 7k, 7l, 7m and 7n elicited micromolar IC50 values ranging between 2.61 to 6.06 µM.
Similarly, the meta-attached linker compounds 7m and 7n having the electron donating 4-
methyl and 4-methoxy, respectively, were less potent relative to the corresponding
compounds 7f and 7g possessing the para-attached linker. Thus, as shown in Table 1,
compounds 7m and 7n possessed micromolar IC50 values in comparison with the
submicromolar IC50 values of compounds 7f and 7g. On the other hand, replacement of the
electronically amphiphilic 4-halo; the electron donating 4-methoxy, 4-methyl and 4-amino

moieties with the electron withdrawing 4-nitro or 4-cyano yielded the potent compounds 7p
and 7q showing submicromolar IC50 values. In conclusion, these results suggest that
attaching the linker to the para position might generally be more favorable than attaching it
to the meta position. In addition, compounds bearing electron withdrawing substituents might
be more active than those with electron donating or amphiphilic substituents.
2.2.2. Mechanistic study of the induced IKK-β inhibition
Since we expect that these compounds to bind covalently to IKK-β, we selected the most
active inhibitors, as revealed in the biological assay against IKK-β, for investigation of the
elicited mode of IKK-β inhibition. Accordingly, the kinetic dose-response curves
(measurement intervals equivalent to 0, 20, 40 and 60 min) were generated using
approximately three-fold dilutions (30, 10, 3, 1, 0.3, 0.1, 0.03, 0.01, 0.003 and 0.001 µM) for
each of compounds 7a and 7p (Fig. 3). First, the validity of data was verified by calculating
the standard errors for tops, bottoms, HillSlope, spanning values for the curves and checking
the goodness of fitting; all of which showed excellent values and 95% confidence spanning
nearly the measured ranges. The kinetic measurements showed that increasing the time
boosts the elicited inhibitions; a characteristic of irreversible inhibition reaction. Table 2
summarizes the parameters of kinetic dose response curves. At zero time, the maximum
inhibition elicited by compound 7a was 60.9% but increased to 76.6% after 60 min. In
contrast, compound 7p was less effective at zero time showing a maximum inhibition of
53.2%, but after 60 min., it elicited a maximum inhibition percent of 89.2% rendering it was
more effective by time than compound 7a. The HillSlope of compound 7p was higher than
that of compound 7a indicating higher rate of the irreversible inactivation reaction of IKK-β
by compound 7p. Considering the fact that interaction of irreversible inhibitors with the
target protein involves an initial physical binding step followed by a second step of chemical

interaction to establish a covalent bond, these kinetic data might indicate compound 7a
possessing the para attached linker elicits a better physical binding phase but a lower rate of
chemical interaction relative to compound 7p. Following the published method for
calculation of Kinact values [50], nonlinear regression after calculating the Kobs for each
concentration of compounds 7a and 7p returned Kinact values of 0.013 and 0.027 (min-1),
respectively.
Fig. 3. A) Time-dependent dose-responses curves of compound 7a. B) Time-dependent dose-responses curves of
compound 7p. C) Nonlinear regression of calculated Kobs values against concentration of compound 7a. D)
Nonlinear regression of calculated Kobs values against concentration of compound 7p.
Table 2. Parameters of the time-dependent dose response curves for inhibition of IKK-β by compounds 7a and
7p measured by TR-FRET assay
Compound
7a
7p
Time intervals
0 min
-9.33
60.9
-6.24
0.9
20 min
-7.14
75
40 min
-4.47
76.7
60 min
-0.38
76.6
-6.69
1.56
77
0 min
-5.63
53.2
20 min
-3.68
73
40 min
-8.33
81.8
60 min
-3.88
89.2
Bottom
Top
Best-fit
values
-LogIC50
HillSlope
Span
-6.46
0.87
82.1
-6.6
-6.46
1.32
-6.61
2.43
76.7
-6.66
2.8
-6.66
3.47
1.29
70.2
81.2
58.8
90.1
93.1

Bottom
Top
1.29
2.21
0.05
0.09
2.89
1.54
2.09
0.05
0.08
2.99
1.65
1.85
0.04
0.15
2.68
1.21
1.25
0.03
0.15
1.84
1.81
2.27
0.07
0.25
3.14
1.33
1.38
0.03
0.36
1.98
1.48
1.49
0.03
0.37
2.14
1.11
1.12
0.02
0.41
1.6
Std. Error
LogIC50
HillSlope
Span
95% Confidence Intervals
Span
64.23 ~
76.12
75.99 ~
88.26
75.68 ~
86.70
73.18 ~
80.75
52.33 ~
65.23
72.58 ~
80.72
85.75 ~
94.54
89.77 ~
96.33
DFa
R2
26
26
26
26
26
26
26
26
Goodness of
Fit
0.99
0.99
0.98
0.99
0.96
0.99
0.99
0.99
a DF: Degrees of Freedom
2.2.3. Cellular assays for anti-inflammatory activity
As kinase assays showed a potential in vitro IKK-β inhibitory activity, a cellular level
biological investigation for a set of active derivatives on the viability and inflammatory
mediators has been explored. Among several biological responses stimulated by LPS in RAW
264.7 macrophages, the NF-κB signaling pathway is upregulated resulting in overproduction
of pro-inflammatory cytokines (TNF-α, IL-6) and nitric oxide [51, 52]. Therefore, the
production of these inflammatory mediators, as well as, the cell viability in LPS-treated RAW
264.7 macrophages were measured in presence of three doses (0.1, 1 and 10 µM) of
compounds 7a, 7j, 7k and 7p. The results are presented in the following sections.
2.2.3.1. Impact on RAW 264.7 macrophages Cells’ Viability
Because induced macrophages’ cells death, if any occurs, would lead to a reduction of the
measured production of cytokines (TNF-α and IL-6) in addition to nitric oxide regardless the
inhibition of IKK-β activity, cell viability assay was conducted to avoid inaccuracies arising
from this factor. The assay was conducted at concentrations used for measurement of TNF-α,
IL-6 and nitric oxide in presence and in absence of LPS. At 10 µM concentration, as
illustrated in Fig. 4, all of the tested compounds did not elicit cytotoxicity in absence of LPS.

Also in presence of LPS, compounds 7a, 7k and 7p did not show significant cytotoxicity at
the tested three concentrations. However, in presence of LPS, the 10 µM concentration of
compound 7j reduced the viability of macrophages eliciting IC80 value of 4.43 µM in
presence of 1 µg/ml of LPS (Table 3).
Fig. 4. Cell viability of RAW 264.7 macrophage cells: in absence of both of tested compound and LPS (cyan
color), in presence of LPS (green color), in presence of 6 µM of tested compounds (dark buff color) and in
presence of both of LPS and tested compounds (blue color). Red color indicates cytotoxicity of macrophages
cells (viability percent below 80%).
Table 3. Viability of RAW 264.7 macrophages and IC50 for reduction of production of NO, IL-6 and TNF-α.
IC80 a (µM)
Cell viability
>10
IC50 b (µM)
Compound
TNF-α
6.27
7.87
>10
NO
>10
>10
>10
>10
PGE2
9.83
>10
IL-6
>10
>10
>10
>10
7a
7j
4.43
>10
8.52
7k
>10
5.22
>10
7p
a IC80: defined as the concentration of the compound that decreases viability of cells to 80% in presence of 1

µg/ml of LPS.
b
IC50: defined as the concentration that decreases the measured amount of produced TNF-α NO, PGE2 or IL-6
to 50%.
2.2.3.2. Effect on LPS-induced TNF-α production in RAW 264.7 macrophages
The production of the pro-inflammatory cytokine TNF-α is a downstream of several LPS-
activated signaling pathways including NF-κB, MEK and ERK [53, 54]. Thus, inhibition of
IKK-β as a limiting step in the canonical NF-κB signaling pathway would result in reduction
of the inflammatory mediator TNF-α production levels. The results showed a dose dependent
decrease of TNF-α production in presence of compound 7p with 84% reduction relative to the
amount produced by LPS (Fig. 5). Similarly, the 10 µM concentration of compounds 7a and
7j exerted reduction of LPS-stimulated TNF-α production by 80% and 71%, respectively.
However, the 0.1 and 1 µM concentrations of both 7a and 7j did not result in a significant
reduction. Lastly, compound 7k was the least potent among evaluated compounds eliciting 44%
reduction of TNF-α production at 10 µM concentration. The determined IC50 values were
5.22, 6.27 and 7.87 for compounds 7p, 7a and 7j, respectively (Table 3). These results
indicate the anti-inflammatory properties of the prepared compounds.

Fig. 5. Levels of TNF-α produced in RAW 264.7 macrophage cells: green color indicates presence of only LPS;
blue color indicates presence of a candidate compound and LPS; **p<0.01, ***p<0.001.
2.2.3.3. Effect on LPS-induced nitric oxide production in RAW 264.7 macrophages
Inducible nitric oxide synthase (iNOS) is responsible for production of nitric oxide in
inflammatory conditions. The expression of iNOS is stimulated by several signaling
pathways including NF-κB, mitogen-activated protein kinases (p38, ERK1/2 and JNK) and
JAK/STAT [5, 55]. Thus, inhibition of NF-κB signaling pathway would result in reduction of
nitric oxide (NO) production, at least in part. To determine the extent of contribution of the
tested compounds to the reduction of NO production in LPS-stimulated RAW 264.7
macrophages, first, it was confirmed that the level of NO production in presence of 10 µM
concentration of only the candidate compound is unchanged from that of non-stimulated
macrophages. Second, the amounts of NO produced in LPS-stimulated macrophages were
measured in presence of increasing concentrations of tested compounds and referred to the
amounts produced after and before LPS stimulation. A concentration of 40 µM of the

selective inducible nitric oxide synthase inhibitor NIL was used as a standard for comparison
[5, 25, 56]. As shown in Fig. 6, all of compounds 7j, 7k and 7p showed a dose dependent
significant reduction of NO production. Even though the low concentrations of compound 7a
did not elicit significant reduction of NO production, the 10 µM concentration exerted
significant reduction close to, albeit higher, the reduction of NO production by 40 µM
concentration of the standard NIL. Similarly, the 10 µM concentration of compounds 7j and
7k elicited reduction of NO production approximating the reduction exhibited by 40 µM
concentration of the standard NIL. Even though compound 7p was the least effective
inhibitor of NO production, its ability was still significant at concentrations as low as 0.1 µM.
These data confirm the beneficial outcome of these compounds on reduction of NO
production through blocking the NF-κB signaling pathway. However, the measured high IC50
values might be due to presence of other LPS-activated signaling pathways which also
contribute to NO production.

Fig. 6. Levels of NO production in RAW 264.7 macrophage cells: yellow color indicates basal NO production;
dark buff color indicates NO production in non-induced cells in presence of candidate compound; green color
indicates NO production in LPS-induced cells; blue color indicates NO production in LPS-induced cells in
presence of candidate compounds; black color indicates NO production in LPS-induced cells in presence of
standard; **p<0.01, ***p<0.001.
2.2.3.4. Effect on LPS-induced PGE2 production in RAW 264.7 macrophages
Stimulation of macrophages by LPS results in activation of mitogen-activated protein kinase
(MAPK) and NF-κB signaling pathways. Both signaling pathways lead to induction of the
pro-inflammatory prostaglandin E2 (PGE2) production [57]. As a consequence, NF-κB
signaling pathway inhibitors would reduce PGE2 production resulting in alleviation of
inflammatory conditions. To verify the capability of the prepared IKK-β inhibitors to reduce
PGE2 production, LPS-induced RAW 264.7 macrophages model was used. As presented in
Fig. 7, incubation of macrophages with candidate compounds in absence of LPS did not show
significant increase in PGE2 production indicating no inflammation stimulatory activity of

these compounds. When macrophages were incubated with LPS, PGE2 production increased
substantially from 0.16 to 33.64 ng/mL which was reduced by the standard NS-398 [58] to
12.02 ng/mL. When macrophages were incubated with LPS in the presence of 10 µM
concentrations of candidate compounds, PGE2 production decreased to 13.87, 16.53 and
19.04 ng/mL by compounds 7k, 7a and 7p, respectively, while compound 7j did not affect
PGE2 production significantly. In contrast to compounds 7j, 7k and 7p whose 1 and 0.1 µM
concentrations did not reduce PGE2 production significantly, compound 7a exhibited
significant reduction of PGE2 production even at the low 0.1 µM concentration. The
calculated IC50 values for inhibition of PGE2 production by the most potent compounds 7k
and 7a were 8.52 and 9.83 µM, respectively. These data suggest potential anti-inflammatory
activity of compounds 7k and 7a.
Fig. 7. Levels of PGE2 production in LPS-induced RAW 264.7 macrophages: dark buff color indicates PGE2

production in non-induced cells in presence of candidate compound; green color indicates PGE2 production in
LPS-induced cells; blue color indicates PGE2 production in LPS-induced cells in presence of candidate
compounds; black color indicates PGE2 production in LPS-induced cells in presence of standard; **p<0.01,
***p<0.001.
2.2.3.5. Effect on LPS-induced IL-6 production in RAW 264.7 macrophages
Both of NF-κB and MAPK signaling pathways are recruited in the production of IL-6
subsequent to macrophages’ activation by LPS [59]. Accordingly, blockage of NF-κB
signaling via inhibition of IKK-β might contribute to reduction of the pro-inflammatory IL-6
production. To assess the ability of the tested compounds to reduce IL-6 production, the
amounts produced of IL-6 in LPS-stimulated RAW 264.7 macrophages were determined after
incubating the cells for 24 hours with LPS alone or LPS in presence of increasing
concentrations of each candidate compound. As illustrated in Fig. 8, compounds 7j and 7k
elicited a significant dose dependent reduction of IL-6 over the tested three concentrations
while compound 7a failed to show significant reduction. Although not fully comprehensible,
the reduction of IL-6 production by compound 7p was significant but the higher 10 µM dose
did not exert response higher than that of the 1 µM dose. The high IC50 values of all of the
evaluated compounds might be caused by presence of LPS-stimulated signaling pathways
other than NF-κB that can contribute to production of IL-6. However, these results show that
compounds 7j, 7k and 7p can contribute to the reduction of IL-6 production and, hence, assist
in reducing inflammation.

Fig. 8. Levels of IL-6 produced in RAW 264.7 macrophage cells: green color indicates presence of only LPS;
blue color indicates presence of a candidate compound and LPS; ***p<0.001.
In summary, the evaluation of the selected set of compounds at the cellular level revealed that
they possess anti-inflammatory properties with variable capacities to reduce the production of
the evaluated inflammatory mediators. While compound 7a possessed good anti-
inflammatory activity reducing the production of TNF-α, PGE2 and NO, it did not affect the
production of IL-6. On the other side, compound 7k was less potent than 7a in reducing TNF-
α production, but it reduced the production of IL-6, PGE2, and, NO. Although compound 7j
did not affect PGE2 production, meanwhile, compound 7p showed inhibition of PGE2
production, both of them showed some inhibitory activity for the production of other
evaluated inflammatory mediators.
2.2.4. Evaluation of plasma and microsomal stability

Upon in vivo administration, the compound stability is challenged by modification or
degradation triggered by plasma enzymes and/or proteins which might lead to poor in vivo
activity [60, 61]. The plasma instability might become a detrimental issue for the in vivo
efficacy, especially for irreversible inhibitors [62]. Therefore, ranking compounds for in vivo
studies should be guided by a preceded plasma stability study. Accordingly, the most potent
compounds in the conducted cellular assay were subjected to evaluation of their plasma
stability. In this regard, the amount remaining of compounds 7a and 7p after incubation with
rat plasma at 37 ºC were determined over a range of time intervals up to 2 hours (Fig. 9A).
The obtained curve showed high plasma stability for both compounds. With 83.3% remaining
after two hours, compound 7p was slightly more stable than compounds 7a whose remaining
amount was 73.6%.
Assessment of the metabolic stability of the promising candidates is important before
proceeding to in vivo evaluation because high metabolism rate would lead to a rapid
clearance of the administered molecule and accordingly unsatisfactory circulation time within
the bloodstream to elicit its action in vivo [60]. As hepatic microsomes are the major
contributor to metabolic inactivation of drugs within the body [63], the stability of the
candidate compounds 7a and 7p against human hepatic microsomes was assessed. In this
regard, the tested compounds, as well as, mibefradil as a standard drug were incubated in a
pool of human hepatic microsomes at 37 ºC to determine the percent remaining at time
intervals up to one hour. As presented in Fig. 9B, compound 7p was rapidly metabolized by
human hepatic microsomes with half life time around 20 min. On the other hand, compound
7a has a lower microsomal metabolism rate comparable to the calcium channel blocker drug
mibefradil. The half life time of compound 7a was very close to one hour. Collectively, the
performed plasma and microsomal stability studies present compound 7a as a potential

candidate for in vivo evaluation.
Fig. 9. (A) Plasma stability of compounds 7a and 7p versus time, (B) Human hepatic microsomal stability of
compounds 7a and 7p versus time, (C) In vivo evaluation of compound 7a protection against LPS-induced
septic shock.
2.2.5. In vivo protection against septic shock
Sepsis is a systemic inflammatory response that results in the fatal septic shock. The
pathogenesis of sepsis involves crucial roles for NO, PGE2, and TNF-α [64, 65]. As
compound 7a elicited in vitro inhibition of LPS-induced production of the inflammatory
mediators, as well as, reasonable plasma and microsomal stability, further in vivo assessment
was conducted. The protective effect of compound 7a against LPS-induced septic shock in
mice as an animal model was assessed. As shown in Fig. 9C, compound 7a exhibited an
efficient protection at both of the two administered doses. Intraperitoneal injection of LPS

resulted in 100% mortality within 16 hours, while pre-treatment with compound 7a resulted
in survival rate of 80% at 36 hours post injection.
2.3. In silico simulation study
IKK-β is composed of three domains which are: the kinase domain (KD, residues 16–307),
ubiquitin-like domain (ULD, residues 310–394), and scaffold dimerization domain (SDD,
410–666) 3QA8) [44]. The structure of the allosteric site allocated in the KD of IKK-β
incorporates a cysteine residue (Cys46) as a part of a short loop structure connecting a short
β-sheet structure (residues 41-44) with α-helix structure (residues 52–64). The literature
shows that kinase activity of IKK-β is sensitive to changes at this site; mutation of Lys44
which is the last residue in the β-sheet sequence, meanwhile, is one residue away from Cys46,
results in loss of IKK-β kinase activity [66]. Also, literature reports proved that the natural
product covalent allosteric IKK-β inhibitor binds to Cys46 at an allosteric site [38, 39].
Previous computational study showed that blocking solvent access to Trp58 is accompanied
with inhibition of kinase activity [38].
To get insights into the molecular interactions that trigger the elicited irreversible inhibition
of IKK-β by the synthesized compounds, a molecular simulation study was conducted using
the crystal structure of IKK-β (protein databank code: 3QA8). In this regard, covalent
docking of the active compounds 7a and 7p was conducted using CovDock workflow as
implemented in Schrödinger Suite. In such protocol, minimization of Ligand-protein complex
adduct is performed after covalent docking of the ligand. The top five poses were retrieved
and analyzed visually. After forming the ligand-IKK-β adduct, the top scoring pose of the
active compound 7a, possessing para linked N-methylpiperazine, showed the N-
methylpiperazine moiety directed towards the α-helix where Trp58 is buried (Fig 10A). This

binding mode is stabilized via favorable hydrophobic interactions with Tyr28, Arg47, Cys46
and Arg55, anion interaction with Glu49, in addition to hydrogen bonding interaction with
Pro52. Also, the second best pose of compound 7a presented more or less similar binding
mode which is stabilized by favorable hydrophobic interactions with Tyr28, Arg47 and Leu91,
cation interaction with Arg55, as well as hydrogen bonding interactions with Asp90, Gln48
and Arg55 (Fig. 10B). The scores of other poses of compound 7a were lower than these two
poses. In case of analog 7p bearing N-methylpiperazinyl-pentyl linked at the meta position,
the detected binding mode showed that the N-methylpiperazine moiety is directed towards the
α-helix in which Trp58 residue is buried which resembles the case for the active compounds
bearing p-linked N-methylpiperazinyl-alkyl moiety (Fig. 10C). This binding mode is
stabilized via a network of favorable interactions comparable to those of compounds 7a. In
lieu of these calculated binding modes, the inhibitory activity of this class of compound on
IKK-β would be understandable.

Fig. 10. The structures of covalently docked active compounds 7a and 7p: A) The best scoring binding mode of
compound 7a; B) The second best scoring binding mode of compound 7a; C) The calculated binding mode for
compound 7p.
3. Conclusion
Starting from in vitro micromolar active lead compound 6, a series of thiazolidine-2,4-dione
derivatives was developed as irreversible allosteric covalent inhibitors of IKK-β with
potential in vivo anti-inflammatory activity. In general, the analysis of structure activity
relationship revealed that compounds possessing the para-attached linker were the most

potent among the prepared compounds. Almost all of them elicited submicromolar IC50
values with the most potent IKK-β inhibitor compound 7a eliciting IC50 = 0.20 µM. In
comparison, most of the derivatives having the linker attached to the meta position were less
potent eliciting low micromolar IC50 values. However, when the 4-substitution on the
phenoxy ring was the electron withdrawing group nitro or cyano, the IC50 values went again
into the submicromolar range. Evaluation of the cellular activity of a set of active compounds
showed that compounds 7a and 7p exhibited significant anti-inflammatory activity in LPS-
treated macrophages through inhibition of production of inflammatory mediators. The
conducted plasma and microsomal stability assays predicted superiority of in vivo stability of
compound 7a over 7p. Accordingly, in vivo evaluation of compound 7a was conducted in
animal model of LPS-induced septic shock which confirmed its in vivo efficacy as it
increased the survival rates of treated animal groups. In silico simulation study provided a
reasonable explanation for the elicited IKK-β inhibitory activity. Collectively, this study
indicates that the irreversible covalent allosteric IKK-β inhibitor 7a could be a promising
anti-inflammatory investigational candidate for further evaluation.
4. Experimental
4.1. Chemistry
General: All reactions and manipulations were performed using standard Schlenk techniques.
All solvents and reagents have been purchased from commercial suppliers unless otherwise
mentioned and used without further purification. The NMR spectra were obtained on Bruker
1
Avance 400. H NMR spectra were referenced to tetramethylsilane (δ = 0.00 ppm) as an
internal standard and are reported as follows: chemical shift, multiplicity (br = broad, s =
singlet, d = doublet, t = triplet, dd = doublet of doublet, m = multiplet). Column
chromatography was performed on Merck Silica Gel 60 (230−400 mesh) and eluting solvents

for all of these chromatographic methods are noted as appropriated-mixed solvent with given
volume-to-volume ratios. TLC was carried out using glass sheets pre-coated with silica gel 60
F254 purchased by Merk. High-resolution spectra were performed on Waters ACQUITY
UPLC BEH C18 1.7µ−Q-TOF SYNAPT G2-Si High Definition Mass Spectrometry. HRMS
and HPLC purity charts, preparation of 4-(3,4,5-trifluorophenoxy)benzaldehyde reagent,
NMR charts as well as LC and MS conditions have been added to this article in the
supplementary file. Synthesis of intermediates 9a, 9b, 10a, 10b, 11a, 11b, 12a and 12b has
been reported in our recent publication [41].
4.1.1. 3-(3-(5-Iodopentyloxy)phenyl)thiazolidine-2,4-dione (12c). It has been synthesized
1
following the same procedure used for synthesis of compounds 12a and 12b [41]. H NMR
(400 MHz, DMSO-d6) δ: 1.48–1.53 (m, 2H), 1.71–1.76 (m, 2H), 1.79–1.86 (m, 2H), 3.31 (t, J
= 6.9 Hz, 2H), 3.96 (t, J = 6.4 Hz, 2H), 4.29 (s, 2H), 6.84–6.87 (m, 1H), 6.90 (t, J = 2.0 Hz,
1H), 7.01–7.04 (m, 1H), 7.40 (t, J = 8.1 Hz, 1H).
4.1.2. General procedure of intermediates 13.
1-Methylpiperazine (0.15 mL, 1.34 mmol) was added to a solution of the appropriate 12
derivative (0.67 mmol) and K2CO3 (180 mg, 1.34 mmol) in MeCN (3 mL). The reaction
mixture was stirred at room temperature for 9 h. The mixture was quenched by addition of
water and extracted with DCM. The organic layer was dried over anhydrous MgSO4, filtered
and evaporated under reduced pressure. The residue was purified by column chromatography
(SiO2, DCM/MeOH = 15/1 v/v including 1% of NH4OH).
3-(4-(3-(4-Methylpiperazin-1-yl)propoxy)phenyl)thiazolidine-2,4-dione (13a). Yield: 57%,

1H NMR (400 MHz, DMSO-d6) δ 1.84−1.88 (m, 2H), 2.13 (s, 3H), 2.32 (br, 8H), 2.40 (t, J =
7.1 Hz, 2H), 4.02 (t, J = 6.3 Hz, 2H), 4.27 (s, 2H), 7.02 (d, J = 8.8 Hz, 2H), 7.18 (d, J = 8.65
Hz, 2H). HRMS (ES+): m/z calculated for C17H23N3O3S: 350.1538 [M+H]+. Found 350.1540.
3-(4-(5-(4-Methylpiperazin-1-yl)pentyloxy)phenyl)thiazolidine-2,4-dione (13b). Yield:
43%, 1H NMR (400 MHz, DMSO-d6) δ 1.37−1.49 (m, 4H), 1.73 (q, J = 7.9 Hz, 2H), 2.13 (s,
3H), 2.25 (t, J = 6.7 Hz, 2H), 2.30 (br, 8H), 3.98 (t, J = 6.4 Hz, 2H), 4.27 (s, 2H), 7.02 (d, J =
8.9 Hz, 2H), 7.18 (d, J = 8.9 Hz, 2H).
3-(3-(5-(4-Methylpiperazin-1-yl)pentyloxy)phenyl)thiazolidine-2,4-dione (13c). Yield:
55%, 1H NMR (400 MHz, DMSO-d6) δ: 1.37–1.47 (m, 4H), 1.72 (t, J = 7.1 Hz, 2H), 2.13 (s,
3H), 2.23–2.33 (m, 10H), 3.95 (t, J = 6.5 Hz, 2H), 4.29 (s, 2H), 6.85 (d, J = 7.4 Hz, 1H), 6.90
(t, J = 2.0 Hz, 1H), 7.02 (dd, J = 2.3 Hz, 8.2 Hz, 1H), 7.39 (t, J = 8.1 Hz, 1H).
4.1.3. General procedure of target compounds (7a–q). NaOAc anhydrous (17.6 mg, 0.21
mmol), the appropriate aldehyde derivative (0.14 mmol) and the appropriate 13 analog (0.07
mmol) were dissolved in AcOH (3 mL) and stirred at 110 °C for 16 h. After cooling, the
mixture was neutralized to pH 8 by addition of NH4OH and extracted with DCM. The
organic layer was dried over anhydrous MgSO4, filtered and concentrated under reduced
pressure. The residue was purified by column chromatography (SiO2, DCM/MeOH = 15/1
v/v including 1% of NH4OH).
(Z)-4-(4-((3-(4-(3-(4-Methylpiperazin-1-yl)propoxy)phenyl)-2,4-dioxothiazolidin-5-
ylidene)methyl)phenoxy)benzonitrile (7a). Yield: 45%, mp: 247.7−248.2 °C, HPLC purity:
8.58 min, 93.77%, 1H NMR (400 MHz, DMSO-d6) δ 2.28 (br, 2H), 2.84 (s, 3H), 3.60−3.79