First Bispecific Inhibitors of the Epidermal Growth Factor Receptor Kinase and the NF-κB Activity As Novel Anticancer Agents

Article abstract of DOI:10.1021/acs.jmedchem.6b01774

The activation of the NF-κB transcription factor is a major adaptive response induced upon treatment with EGFR kinase inhibitors, leading to the emergence of resistance in nonsmall cell lung cancer and other tumor types. To suppress this survival mechanism, we developed new thiourea quinazoline derivatives that are dual inhibitors of both EGFR kinase and the NF-κB activity. Optimization of the hit compound, identified in a NF-κB reporter gene assay, led to compound 9b, exhibiting a cellular IC₅₀ for NF-κB inhibition of 0.3 μM while retaining a potent EGFR kinase inhibition (IC₅₀ = 60 nM). The dual inhibitors showed a higher potency than gefitinib to inhibit cell growth of EGFR-overexpressing tumor cell lines in vitro and in a xenograft model in vivo, while no signs of toxicity were observed. An investigation of the molecular mechanism of NF-κB suppression revealed that the dual inhibitors depleted the transcriptional coactivator CREB-binding protein from the NF-κB complex in the nucleus.

Full text of DOI:10.1021/acs.jmedchem.6b01774

Subscriber access provided by UNIV OF TASMANIA
Article
First bispecific inhibitors of the Epidermal Growth factor receptor
kinase and the NF-#B activity as novel anti-cancer agents
Mostafa M. Hamed, Sarah S Darwish, Jennifer Herrmann, Ashraf H. Abadi, and Matthias Engel
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01774 • Publication Date (Web): 14 Mar 2017
Downloaded from http://pubs.acs.org on March 14, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
First bispecific inhibitors of the Epidermal Growth
factor receptor kinase and the NFꢀκB activity as
novel antiꢀcancer agents
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
†
‡
‡
§
‡
∥
,
Mostafa M. Hamed, Sarah S. Darwish, Jennifer Herrmann, Ashraf Abadi, Matthias Engel *
†
Pharmaceutical and Medicinal Chemistry, Saarland University, Campus C2.3, and Helmholtz
Institute for Pharmaceutical Research Saarland (HIPS), Campus E8.1, Dꢀ66123 Saarbrücken,
Germany.
‡
Department of Pharmaceutical Chemistry, Faculty of Pharmacy and Biotechnology, German
University in Cairo, Cairo 11835, Egypt.
§
Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus E8.1, Dꢀ66123
Saarbrücken, Germany.
∥ Pharmaceutical and Medicinal Chemistry, Saarland University, Campus C2.3, Dꢀ66123
Saarbrücken, Germany.
ABSTRACT
The activation of the NFꢀκB transcription factor is a major adaptive response induced upon
treatment with EGFR kinase inhibitors, leading to the emergence of resistance in nonꢀsmall cell
lung cancer and other tumor types. To suppress this survival mechanism, we developed new
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
thiourea quinazoline derivatives that are dual inhibitors of both EGFR kinase and the NFꢀκB
activity. Optimization of the hit compound, identified in a NFꢀκB reporter gene assay, led to
compound 9b, exhibiting a cellular IC for NFꢀκB inhibition of 0.3 µM while retaining a potent
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
EGFR kinase inhibition (IC50 = 60 nM). The dual inhibitors showed a higher potency than
gefitinib to inhibit cell growth of EGFR–overexpressing tumor cell lines in vitro and in a
xenograft model in vivo, while no signs of toxicity were observed. An investigation of the
molecular mechanism of NFꢀκB suppression revealed that the dual inhibitors depleted the
transcriptional coꢀactivator CREBꢀbinding protein from the NFꢀκB complex in the nucleus.
Introduction
Inhibition of the epidermal growth factor (EGF) receptor kinase–mediated signaling is a wellꢀ
established strategy for the treatment of advanced stage nonꢀsmall cell lung cancer (NSCLC).
EGFR was found to be overexpressed in about 62 % of NSCLC cases, which correlated with a
1
poor prognosis. However, in most cases, the clinically used EGFR kinase inhibitors, such as
gefitinib and erlotinib, are only effective if the tumor cells harbor specific activating EGFR
mutations which appear to preserve the ligand–dependency of receptor activation but alter the
2
pattern of downstream signaling. These EGFR mutations include mainly small inꢀframe
2
deletions in exon 19, or the single point mutation L858R, and are found in ~10ꢀ50 % of NSCLC
1
patients, depending on the ethnicity and the tumor subtype. Of those, about 75 % show a
2, 3
response to the small molecule inhibitors compared to ∼10 % in the wildꢀtype case.
Hence,
only a minor proportion of lung cancer patients can actually profit from the treatment with EGFR
inhibitors.
ACS Paragon Plus Environment
2

Page 3 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In addition, tumors showing an initial response to treatment with EGFR inhibitors often
become resistant due to acquisition of a mutation in the ATP binding pocket of EGFR (T790M),
2
, 4
which mainly decreases the K for ATP, thus outꢀcompeting the binding of gefitinib.
m
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Alternatively, tumor cells might activate distinct proꢀsurvival signaling pathways, as exemplified
5
by the amplification of MET in lung cancers treated with EGFR inhibitors.
Tripleꢀnegative breast cancer was considered another potential indication for EGFR kinase
6
inhibitors, because they are not amenable to endocrine or ErbB2 (HER2)–directed therapies,
7
, 8
while they overexpress EGFR kinase, paralleling the loss of estrogen receptor–dependency.
However, inhibition of EGFR by gefitinib or erlotinib showed low efficacy in arresting tumor
9
cell growth at the cell line level, e. g. using tripleꢀnegative MDAꢀMBꢀ231 cells, and results
1
0
from clinical trials were disappointing. Nonetheless, cell line studies showed that besides
EGFR overexpression, the transcription factor NFꢀκB becomes constitutively active in ER–
negative breast cancer cells, thus ensuring cell survival in spite of EGFR inhibition. Indeed,
some evidence suggested that an additive or synergistic effect could be expected from a
simultaneous inhibition of both EGFR kinase and the NFꢀkB pathway in tripleꢀnegative breast
1
1, 12
cancer cells.
Notably, a crucial role of NFꢀκB in ensuring cell survival was also reported for
other tumor types that overexpress EGFR kinase, in particular lung tumors but also head and
1
3
neck squamous cell carcinomas (reviewed in ). Using lung cancer cell lines, a large siRNA
screen identified the NFꢀκB pathway activity as a key factor that determined the sensitivity
towards EGFR inhibitors. In the same study, knockdown of several components of the NFꢀκB
14
pathway enhanced cell death induced by EGFR inhibition. These findings were further
validated in EGFR–dependent tumor models, confirming that the activation of NFꢀκB signaling
conferred resistance to EGFR inhibitors and, conversely, that NFꢀκB knockdown enhanced
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
14
sensitivity to the latter. Furthermore, it was recently shown that in EGFR–mutant lung
adenocarcinoma from patients, initial EGFR inhibitor treatment led to a rapid engagement of NFꢀ
1
5
κB signaling, which promoted tumor cell survival and residual disease. In this study, direct
pharmacologic NFꢀκB inhibition prevented the EGFR inhibitor–induced adaptive response, thus
suppressing the emergence of EGFR inhibitor therapy resistance in several NSCLC models and
in a patient–derived xenograft in vivo. These findings demonstrated that NFꢀκB activation is a
critical adaptive survival mechanism engaged by EGFR oncogene inhibition and provided a
rationale for EGFR and NFꢀκB coꢀinhibition to eliminate residual disease and to extend the
duration of patient responses.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
While coꢀadministration of antiꢀtumor therapeutics is a common strategy in several current
cancer trials and has proven to be beneficial in some cases, toxic side effects could increase by
16
the number of different agents. Moreover, the individual pharmacokinetic properties render it
difficult to deliver effective amounts of multiple therapeutics to the tumor cells in a concerted
manner to achieve maximum efficacy. Therefore, it would be a major advantage to combine the
desired pharmacological activities in a single agent, in order to simultaneously suppress the
synergistic EGFR and NFꢀκB signaling pathways at the same time. In the following, we describe
the development of dual inhibitors of EGFR kinase and the NFꢀκB transcriptional activity based
on the 4ꢀaminophenylquinazoline scaffold, followed by an investigation of the mechanism of
NFꢀκB suppression. Eventually, we provide first evidence for the efficacy of the novel antiꢀ
cancer agents in vivo.
ACS Paragon Plus Environment
4

Page 5 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Results and Discussion
Identification of hits displaying dual EGFR and NFꢀκB inhibitory activity.
With respect to EGFR kinase inhibition, it was known from previous studies that the 4ꢀ
aminophenylquinazoline motif is both essential and sufficient to mediate strong inhibition of the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
7, 18
kinase in the nM range.
On the other hand, the 6ꢀ and 7ꢀpositions of the quinazoline scaffold
offered possibilities for substitutions without strongly compromising the EGFR–directed
potency, because these positions pointed towards the outside of the ATP binding pocket
(
compare e.g., PDB 2ITY). Of note, the quinazoline heterocycle was successfully used as a
scaffold for the synthesis of potent inhibitors for a range of enzymes besides protein kinases,
19
20
21
25
including endothelin converting enzyme, thymidylate synthase, trypanothione reductase,
2
2, 23
24
phosphodiesterases,
glucocerebrosidase, and G9aꢀlike protein lysine methyltransferase.
Hence, the quinazoline system can be considered a “privileged scaffold” according to Evans et
2
6
al. , thus potentially suitable to serve as an affinity anchor in inhibitors of diverse enzymes –
without evidence of promiscuous binding. Thus, there was some chance that new 4ꢀ
aminophenylquinazoline derivatives – which are likely to retain EGFR inhibitory properties –
would also inhibit one of the steps in the NFꢀκB activation cascade. Accordingly, a set of
appropriate derivatives was prepared by us, supplemented by compounds from previous
27, 28
studies,
and screened for inhibitory activity on the NFꢀκB activation pathway using a
luciferase–based reporter gene assay. The compounds selected for the screening included
4
derivatives with variable substitutions at the N ꢀposition of the aminoquinazoline nucleus, such
as haloanilines, alkylanilines, alkoxyanilines, sulfonamide containing anilines and alicyclic
amines (Chart 1, IꢀII). Another set of compounds featured at the 6ꢀposition different
combinations of linkers, potentially acting as a Hꢀbond donor/acceptor pair, as well as aliphatic
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
or (hetero)aromatic moieties which may be accommodated in potential hydrophobic binding
pockets of new target proteins. In this compound set, mꢀbromoaniline was kept constant at
position 4. The linker types connecting different substituents (R in Chart 1) to the quinazoline
nucleus comprised imine (III), amide (IV), amino alkyl amide (V) and thiourea (VI) groups
(Chart 1; the structures of all screened compounds are shown in Chart S1, Supporting
Information).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The reporter gene assay was performed using the lymphoma cell line U937. Due to its origin
2
9
from tissue macrophages, this cell line responds with a strong activation of the NFꢀκB pathway
after stimulation by LPS or TNFα. Inhibition of any of the essential components of the conserved
classical (canonical) NFꢀκB pathway would be expected to result in a decrease of the final
luciferase activity–based read out. Moreover, it was of importance that the U937 lymphoma cell
type lacks expression of EGFR, thus excluding any interference due to the intrinsic EGFR
inhibitory activity of the compounds.
Indeed, screening of the 4ꢀanilinoquinazoline derivatives at 10 µM led to the identification of
several compounds which suppressed the NFꢀκB reporter gene activity in U937 cells at variable
degrees (Table 1; IC s were calculated for the compounds showing more than 80 % inhibition).
5
0
The most potent hit was the benzylthiourea derivative VIa, exhibiting almost 100 % reduction of
the luciferase read out (Table 1). In comparison, the reference compound gefitinib showed a
considerably weaker inhibition of about 50 % at 10 µM, suggesting that the structural variations
had created a significant inhibitory activity on the NFꢀκB pathway. Importantly, VIa still
retained a nM potency with respect to EGFR inhibition, though it was about four times reduced
compared to gefitinib (Table 1). It is worth mentioning that 4ꢀphenethylaminoquinazoline
30, 31
derivatives were previously reported to inhibit the NFꢀκB activation in Jurkat cells,
ACS Paragon Plus Environment
6

Page 7 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
however, the substitution pattern and the obtained SAR were different; in particular, an ethylene
spacer between the quinazoline C(4) amino function and the phenyl ring was essential for NFꢀκB
1
7
inhibition, which would be incompatible with EGFR kinase inhibition. In contrast, the 4ꢀaniline
analog was inactive towards NFꢀκB, suggesting the involvement of another target. However, the
mechanism of NFꢀκB inhibition was not investigated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Screening of hit compound VIa against kinases directly involved in TNFα Receptor
signalling.
At this stage, it was straightforward to screen the most potent hit, VIa, against a panel of
32
kinases specifically involved in NFꢀκB activation in U937 cells, in order to test whether
selective inhibition of one of those kinases was responsible for the novel activity. Only one
kinase in the panel shown in Table 2, RIPKꢀ2, was weakly inhibited by VIa; however, with the
estimated IC above 10 µM at an ATP concentration of 5 µM in the assay, RIPKꢀ2 was unlikely
5
0
to be a target of this compound in cells. Thus, we could conclude that compound VIa did not
affect a kinase which is directly involved in TNFα receptor signaling. The results also showed
that VIa did not exhibit nonꢀselective kinase inhibition, which encouraged us to carry out an
optimization of the potency guided by the NFꢀκB reporter gene assay.
Chemistry
The identified hit compound VIa was subject to optimization by a targeted synthesis of
analogs. Our modifications centered on the substituents at the 4ꢀanilino ring, the side chain
attached to the thiourea linker and the thiourea linker itself (Schemes 1 and 2, Chart 2).
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Synthesis of the quinazoline nucleus was done by refluxing of 5ꢀnitroꢀ2ꢀaminobenzonitirile
with triethyl orthoformate in presence of drops of acetic anhydride to yield the formimidate
derivative 1. Cyclization to form the quinazoline nucleus took place by refluxing of 1 with
different anilines in acetic acid to yield the nitroquinazoline derivatives 2aꢀq. Reduction of the
nitro intermediates 2aꢀq to their amino derivatives 3aꢀq was done by refluxing the nitro
derivatives with stannous chloride in methanol under nitrogen atmosphere. The benzyl thiourea
derivatives 4aꢀ4p and VIa were obtained by stirring the aminoquinazoline derivatives 3aꢀq with
benzylisothiocyanate in DMF (Scheme 1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Reaction of compound 3q with thiophosgene yielded the isothiocyanate derivative 5 which
upon stirring with different amines in DMF gave the thiourea derivatives 6aꢀq and 7aꢀe (Scheme
4
2). The N ꢀ(3ꢀchlorophenyl) analogs of the later optimization campaign were synthesized in a
similar manner (Scheme 3). The urea derivatives 8aꢀb were obtained by stirring compound 3q
with different isocyanate derivatives in DMF (Scheme 2).
Optimization of hit VIa as dual inhibitors of both EGFR kinase and NFꢀκB activity.
The optimization of the hit compound included three parts. The first part was concerned with
the modifications of the substituents on the 4ꢀanilino ring while keeping the benzylthiourea at
position 6 of the quinazoline (Chart 2, 4aꢀp). The second part focused on modifications in the
side chain linked to the thiourea moiety while keeping the 3ꢀbromoaniline at position 4 of the
quinazoline (Chart 2, compound classes 6aꢀq and 7aꢀe). The last part was to confirm the
importance of the thiourea group by replacing it with the urea moiety (Chart 2, 8aꢀb).
First, benzyl thiourea derivatives (4aꢀp) according to route A were explored (Chart 2). It was
found that lipophilic substituents were highly preferred at the aniline ring, as the presence of
polar groups impaired the activity. This was clearly seen with polar substituents such as the
ACS Paragon Plus Environment
8

Page 9 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydroxy (4k and 4l), sulfonamide (4m), substituted sulfonamide (4n and 4o) or even the
bioisosteric pyridine (4p), which all exhibited markedly reduced NFꢀκB inhibitory activity
(Table 3). It was unlikely that this was only due to decreased cell permeability. Rather, the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
uniform reduction of activity with the more polar moieties suggested that the 4ꢀaminophenyl is
not only important for the potency toward EGFR kinase (see SAR discussion below) but also
seemed to interact with the unknown target(s) in the NFꢀκB pathway. With respect to the
lipophilic metaꢀsubstituents on the aniline ring, it was found that the halogens were most
favorable, with chlorine giving the highest NFꢀκB inhibitory activity (4c). This was followed by
bromine (VIa), ethyl (4f), methyl (4d) and finally the 2,3ꢀdimethyl (4e). Interestingly, the bromo
substituent was tolerated in the orthoꢀ, metaꢀ and paraꢀposition (4a, VIa and 4b, respectively),
with a slight preference for the paraꢀposition. However, this was only true for the NFꢀκB
suppression, whereas binding to EGFR strictly demanded the bromine to be in meta (VIa, Table
3, see discussion below). For the NFꢀκB inhibitory activity, the para position tolerated even
more bulky groups, such as tꢀbutyl (4h) and phenyl (4i), however, no further gain of activity
could be achieved compared with the halogens.
The potency of EGFR inhibition was more critically dependent on the position of the
substituents in the aniline ring. It was found that (i) the meta position was optimum for
substitutions and (ii) that the substituents must be hydrophobic. Hence, fortunately, the SAR
trend paralleled that of the NFꢀκB inhibition. Halogens provided the highest increase in potency,
in particular chlorine (4c), which was slightly superior to bromine in the hit compound (VIa).
Replacing the halogen by an alkyl group such as the methyl (4d) decreased the activity, which
further dropped with the polar hydroxyl group (4k). The aminophenyl binding pocket in EGFR
was completely filled by the metaꢀmethylated derivative 4d, whereas the more bulky ethyl group
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
4f) was less tolerated. Moreover, any substitutions in the orthoꢀposition as in 4a or 4e resulted
in a significantly decreased activity. Furthermore, all paraꢀsubstitutions on the aniline also
seemed to produce a steric clash with the binding pocket, irrespective to the nature of the
substituent (cf. 4b, 4g–j, 4l–o). Based on the sufficiently large overlap of the initial SAR, we
concluded that optimization of the NFꢀκB inhibitory activity should be possible without loss of
EGFR inhibition.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Next we tested whether the methylene spacer between the thiourea linker and the aromatic ring
was required. This was done by replacing the benzyl group as in VIa and 7c by phenyl, leading
to 6a and 6d, respectively (route B in Chart 2). As can be seen in Table 4, the nonꢀsubstituted
benzyl/phenyl pair VIa/6a showed comparable biological activities, whereas in the case of the pꢀ
chloro substituted matched pair 7c and 6d, the pꢀchorophenyl analog 6d was more potent than its
benzyl homolog against both targets. This result suggested that the methylene group could be
omitted without loss of activity and that the phenyl derivatives offered a higher potential for
optimization.
Furthermore, the importance of the thiourea group was explored by a direct comparison of the
thiourea derivatives VIa and 6d with their urea analogues 8a and 8b (route C, Chart 2). It
became evident that the thiourea moiety was required for the activity toward the NFꢀκB pathway
(Table 4) but not for EGFR inhibition. The urea motif gave more potent EGFR kinase inhibitors
in case of the benzyl substituent (cf. 8a vs. VIa), however, the urea and thiourea derivatives were
equipotent in case of the phenyl substituent (cf. 8b vs. 6d).
Finally, we investigated if the aromatic ring linked to the thiourea moiety was essential for the
activity. To this end, the aromatic ring was replaced by a methyl group (7a), a morpholinoꢀ4ꢀ
carboxamide derivative (7e) and an ethyl morpholine (7d). Since the suppression of the reporter
ACS Paragon Plus Environment
10

Page 11 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
gene activity was much stronger with the benzyl (VIa) and phenyl (6d) analogs, it was
concluded that the aromatic system significantly contributed to the NFꢀκB inhibitory activity
(Table 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Given that a pꢀchlorine substituent in 6d was well tolerated, several substituents were added to
the phenyl thiourea side chain in an attempt to further enhance the NFꢀκB inhibitory potency.
Firstly, we introduced several polar groups or heteroatoms in the phenyl ring (compounds 6e, 6f,
6
g and 7b), which all abolished the activity toward the NFꢀκB pathway in the U973 cells, except
the hydroxyl group (6e) (Table 4); however, since phenolic compounds are known to affect NFꢀ
33
κB signaling by various, often nonꢀspecific mechanism, this derivative was not further pursued.
Hence we decided to switch to lipophilic substituents (compounds 6bꢀ6d, 6hꢀ6q, Scheme 2); this
modification resulted in sharp SAR, as indicated by highly variable activities depending on the
size and the position of the substituents (compounds 6bꢀ6d, 6hꢀ6q, Table 4). Concerning the
EGFR kinase inhibition, diverse substituents of either lipophilic or hydrophilic nature were
tolerated, with the hydrophilic or heterocyclic moieties being the most potent (cf. 6e, 6f, 7b and
7
d). However, multiple and/or bulky lipophilic substituents on the phenyl ring (such as in 6i, 6m,
6oꢀ6q) decreased or abolished the activity. The latter SAR were conflicting with the
requirements for potent inhibition of NFꢀκB activation, thus it was not possible to optimize both
biological activities in parallel to the same degree. Altogether, compounds 6c and 6h showed the
best balance between EGFR inhibition and simultaneous suppression of the NFꢀκB activation in
the reporter gene assay (Table 4).
The reference compound gefitinib, which is in clinical use for the treatment of NSCLC, was
only moderately active against NFꢀκB in the reporter gene assay (Table 4), confirming that the
inhibition of EGFR kinase was not responsible for the suppression of the NFꢀκB activity in the
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
U937 cells. Moreover, several of the new compounds were considerably more potent in the
inhibition of NFꢀκB, despite of displaying lower EGFR kinase inhibition (cf. 6h, 6j and 6l in
Table 4).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Inhibition of MDAꢀMBꢀ231 breast cancer cell growth.
Next, we evaluated whether the dual inhibitory activity of the novel quinazoline derivatives
boosts the efficacy as potential antiꢀtumor agents compared with EGFR inhibition alone. The
breast cancer cell line MDAꢀMBꢀ231 was chosen because it was frequently used as a model for
9
, 11, 34
studying the effects of EGFR inhibition in tripleꢀnegative breast cancer cells.
Like many
clinical tripleꢀnegative tumors, MDAꢀMBꢀ231 cells are rather insensitive to the EGFR inhibitor
gefitinib (reported IC50 range for cell growth inhibition: 15ꢀ20 µM), in spite of the strong
9
, 35
overexpression of EGFR kinase.
In addition, this cell line displays constitutive activation of
11, 12
36
NFꢀκB,
which is a frequent feature found in most breast cancer tumors.
Hence, any superior efficacy of our novel compounds should be easily detectable using this
cell line. The results of the MTT assay are included in Tables 3 and 4; notably, two of the new
compounds inhibited MDAꢀMBꢀ231 cell growth more than 35 times stronger than the reference
compound gefitinib (cf. 6h, 6j). Our data further indicated that, as a general trend, the potency to
suppress MDAꢀMBꢀ231 cell growth was determined by the NFꢀκB inhibitory activity rather than
by the potency of EGFR kinase inhibition. This was obvious from the fact that, with the
exception of 6n, the most potent inhibitors of the NFꢀκB activation in U937 cells were also the
strongest inhibitors of the MDAꢀMBꢀ231 cell growth (cf. 6c, 6h, 6i, 6j, 6l and 6m in Table 4),
the more since among these, 6i, 6j and 6m showed only low potency against EGFR kinase. Vice
versa, several congeners showing a high activity toward EGFR kinase (VIa, 6a, 6b, 6d, 6e, 6f,
7a, 7b, 7d and 8a), that were among the least potent inhibitors of NFꢀκB activation, inhibited the
ACS Paragon Plus Environment
12

Page 13 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
MDAꢀMBꢀ231 cell growth with only moderate to poor IC s between 8.2 and >30 µM, similarly
5
0
to gefitinib. In conclusion, the suppression of the NFꢀκB activation was mainly responsible for
the tumor cell growth inhibition.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Effects on A549 lung cancer cells vs. nonꢀtumor cell lines.
The potency of the best compounds to inhibit the cell growth of the lung cancer cell line A549
was also tested. This assay was done to corroborate if the compounds with dual inhibitory
activity still offer an advantage towards a cell line which is intermediately sensitive to potent
EGFR inhibitors such as gefitinib. The results showed that the dual inhibitors are more potent
than gefitinib in inhibiting the growth of A549 cancer cells (Table 5). Importantly, when the
selected compounds were tested against nonꢀtumor cell lines (chinese hamster ovary (CHO)ꢀK1
and human umbilical vein endothelial cells (HUVEC)), significantly weaker effects of 6c and 6h
on the cell growth were noted, suggesting that the compounds exerted a rather tumor cell–
selective cytotoxicity (Table 5). Of note, the efficacy/toxicity window was larger for 6c and 6h
than for gefitinib with the cell lines tested.
Kinase selectivity profile of 6c.
Since the novel dual inhibitors were developed based on a kinase inhibitor scaffold, it was
straightforward to test the selectivity over other kinases and whether the suppression of the NFꢀ
κB activation was also due to inhibition of a kinase. To this end, an in vitro selectivity profiling
was performed against a panel of 106 protein kinases, including both tyrosine and Ser/Thr
kinases from each branch of the human kinome. For the profiling we selected 6c, which had
shown a slightly stronger EGFR kinase inhibition than 6h. The screening results are summarized
in Table S1 (Supporting Information). It was found that compound 6c exhibited an excellent
selectivity for the EGFR kinase, with only a weak inhibition of two other kinases, namely Mnk2
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
and Pimꢀ1. The IC s for these two kinases were determined to be 2.7 and 1.2 µM, respectively.
5
0
In the same assay campaign, the IC50 for EGFR kinase was 41 nM (Table S2, Supporting
Information). Interestingly, 6c appeared to be more selective than gefitinib, as the latter had
inhibited in a previously reported screening the kinases Mnk2, SIK2, Mnk1, LOK/STK10,
ErbB2/HER2, RIPK2, EPHA6, ErbB4/HER4 by 46 to 76 % at 0.5 µM (Supplementary Table S3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
7
in Ref. ). From these kinases, Mnk2, LOK, RIPK2 and ErbB4 were also present in the panel
screened with 6c, but only Mnk2 and ErbB4 were appreciable inhibited, by 78 % and 54 %,
respectively, however at a ten times higher concentration (5 µM). The cellꢀfree potency of 6c for
Mnk2 inhibition appeared too low to explain the efficacy of 6c in the cell–based NFꢀκB reporter
gene assay (IC50 = 1.9 µM). Moreover, gefitinib displayed a higher offꢀtarget inhibition of
3
7
Mnk2 than 6c, but was less potent in our reporter gene assay. Thus, Mnk2 could be excluded as
a potential target of the dual inhibitors in the NFꢀκB activation pathway. The inhibition of Pimꢀ1
by 6c was also rather weak, nevertheless we decided to check the potential role of Pimꢀ1 in the
NFꢀκB activation pathway using the potent Pimꢀ1 inhibitor SIMꢀ4a (IC for Pimꢀ1: 17 nM at
5
0
38
1
00 µM ATP in the cellꢀfree assay ). Testing SIMꢀ4a in the NFꢀκB reporter gene assay in
concentrations from 1 to 5 µM, we could not detect any inhibition (Table S3, Supporting
Information), ruling out Pimꢀ1 as a target of 6c in the NFꢀκB activation pathway.
In light of the remarkable selectivity of compound 6c, it was rather unlikely that the new
biological target was another protein kinase, although it could not be fully excluded since not the
complete kinome was screened. However, if it was a kinase, then it would be part of an unknown
NFꢀκB activation pathway that is also induced by TNFα, because all kinases identified as part of
32
the TNFα receptor complex had been included in the kinase screen, also all other kinases which
had previously been mentioned in literature to play a role in NFꢀκB activation.
ACS Paragon Plus Environment
14

Page 15 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Elucidation of the mechanism of action responsible for NFꢀκB suppression. 6c and 6h do
not inhibit the cellular proteasome.
Having developed novel dual inhibitors, we aimed at investigating the cellular mechanism of
action which was responsible for the observed suppression of NFꢀκB activation in the reporter
gene assay, assuming that it was not the inhibition of another kinase. The most obvious
biological activity to test was the potential inhibition of the cellular 26S proteasome, which is
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
9
essential for the degradation of the inhibitory protein IκB. The prototype of proteasome
inhibitors, bortezomib, inhibits mainly two of the three distinct proteolytic activities, thus
40
preventing the release of the NFꢀκB dimer.
For each of the three proteolytic activities, the trypsinꢀlike, the chymotrypsinꢀlike, and the
caspaseꢀlike, we used a specific fluorigenic peptide and total protein extract from MDAꢀMBꢀ231
41
cells as a source of proteasomal activities, basically as described. However, whereas
Bortezomib fully inhibited all three proteolytic activites at 1 µM, none of the test compounds
showed any inhibitory activity even at 50 µM (data not shown). Thus, the three main proteolytic
activities of the proteasome could be excluded as molecular targets.
6c and 6h do not inhibit translocation of NFꢀκB/p65 to the nucleus.
It was theoretically possible that the compounds caused the cytoplasmic retention of the NFꢀ
κB dimer through an alternate mechanism; hence we analyzed whether the translocation of the
RelA subunit (p65) of NFꢀκB was affected. To this end, we used a highꢀcontent screening system
employing a CHO cell line stably expressing a GFPꢀp65 fusion protein. A cytoplasmic retention
of this construct in the presence of the test compounds, as indicated by a diffuse cytoplasmic
fluorescence, would signify an inhibition of the upstream NFꢀκB activation. The system
automatically quantifies the ratio of cytoplasmic vs. nuclear fluorescence and provides
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
microphotographs of each well. The cells were first stimulated for 30 min by 25 ng/mL ILꢀ1β
after which the translocation of the GFPꢀNFκBꢀp65 fusion protein from the cytoplasm to the
nucleus was visualized. Bortezomib was used as a positive control, and gefitinib was also
included for comparison. While bortezomib clearly prevented the migration of the NFꢀκB
construct to the nucleus in a concentration–dependent manner, 6c, 6h and gefitinib were
completely inactive (Figure S1, Supporting Information). To make sure that the lack of activity
of the phenylthiourea derivatives was not due to solubility problems under the conditions used,
Pluronic Fꢀ127 (indicated by “P”) was added in some experiments to increase the solubility at
higher concentrations; however, this did not change the outcome (Figure S1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 6c does not influence DNA binding of NFꢀκB/p65.
Having ruled out that the dual inhibitors prevented the translocation of NFꢀκB/p65 to the
nucleus, it was tempting to speculate that the DNA binding activity of the NFꢀκB complex was
affected by the compounds. To investigate this possibility, the DNA binding capability of the
NFꢀκB species in nuclear protein extracts isolated from MDAꢀMB 231 cells was analyzed using
a commercial ELISA kit. This assay relied on measurement of the specific binding to the NFꢀκB
response element oligonucleotide immobilized in a 96 well plate. The experiment was performed
in two different ways: (i) MDAꢀMB 231 cells were incubated with 6c for 12 h, followed by
washing of the cells, isolation of the nuclear proteins and DNAꢀbinding analysis, and (ii) nuclear
extracts were prepared from untreated cells and the DNAꢀbinding quantified in the presence of
6c vs. DMSO. Method (i) would allow detecting any compound–induced, intrinsic downꢀ
regulation of the DNAꢀbinding activity, whereas variant (ii) would identify a direct inhibition of
DNAꢀbinding. The results of our experiments showed that 6c did not impair the DNAꢀbinding
ACS Paragon Plus Environment
16

Page 17 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity of the NFꢀκB complexes in the nuclear extract, neither directly nor via modulation of
regulatory mechanisms (Figure S2, Supporting Information).
Compound 6c reduces the amount of CREBꢀbinding protein (CBP) in the NFꢀκB
transcription factor complex.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
One of the few remaining possibilities was that the new inhibitors interfere with the NFꢀκB
activity at the level of transcriptional coꢀactivators. It is known that, in order to become fully
transcriptionally active in the nucleus, the NFꢀκB dimer recruits transcriptional coꢀactivators
4
2
such as the CREBꢀbinding protein (CBP)/p300, which functions as a histone acetyl transferase.
Evidence from literature suggested that at least in some cell types, a CBP knockdown can cause
a downꢀregulation of the NFꢀκB transcriptional activity without affecting nuclear translocation or
4
3, 44
DNAꢀbinding affinity of NFꢀκB.
in the DNAꢀbinding NFꢀκB complex isolated from MDAꢀMBꢀ231 cells that were treated with
c. Indeed, we found that compound 6c but not gefitinib reduced the amount of CBP in the NFꢀ
Hence, we decided to analyze the amount of CBP present
6
κB transcription complex in a concentration–dependent manner (Figure 1A). In the same
samples, the amount of RelA/p65 bound to the immobilized NFꢀκB response element
oligonucleotide was not changed by the compounds (Figure 1B). Thus, we provide first evidence
that our dual inhibitors downꢀregulate the transcriptional activity of NFꢀκB through depletion of
the coꢀactivator CBP from the transcription factor complex. However, further studies are needed
to elucidate the underlying mechanism.
Pharmacological disruption of the p65/CBP interaction might abolish the expression of several
tumor–relevant genes. For instance, it was shown that the cooperation of p65/RelA and CBP was
4
5
required for the expression of antiꢀapoptotic genes such as cIAP2 and XIAP in rat fibroblasts as
44
well as the proꢀinflammatory cytokines ILꢀ1β and ILꢀ6 and CSF2 in lung cells.
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Thus, it seems possible that our novel dual inhibitors also suppress the NFꢀkB–mediated ILꢀ6
expression in cancer cells. Of note, ILꢀ6 was identified by Blakely et al. as the NFꢀκB target gene
that was most significantly increased in the adaptive response to EGFR kinase inhibition by
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15
erlotinib, both at the mRNA and the protein level. This adaptive response mechanism was also
1
5
confirmed in resected NSCLC. On the other hand, the interaction of RelA with CBP is only
4
4, 46
required for the expression of a subset of NFꢀκB–dependent genes,
which might result in
lower toxicity than total inhibition of NFꢀκB activity, e. g. as mediated by IκB kinase
47ꢀ49
inhibitors.
Indeed, our in vitro and in vivo experiments suggested low toxicity against nonꢀ
tumor cells and tissues (see below). Also with respect to the therapeutic outcome, total
suppression of NFꢀκB activity is not desirable because it will also affect the antiꢀtumor activity
4
9, 50
of Tꢀlymphocytes in the tumor microenvironment.
Further studies are needed to investigate
whether the thiourea quinazolines will inhibit the CBP–dependent NFꢀκB signaling in tumor
cells sufficiently selectivity so that T cells of the M1 phentotype remain functional.
CBP is a coꢀactivator in a multitude of transcription factor complexes; total depletion of CBP,
in particular from the general transcription factors, would therefore lead to nonꢀselective
cytotoxicity, which was not observed with our dual inhibitors. This suggests a more selective
modulation of CBP association by our compounds, which remains to be elucidated. Interestingly,
CBP was already targeted previously by small molecules that inhibit the histone acetyltransferase
5
1
activity of the protein for the therapy of leukemia. These inhibitors also showed surprisingly
little cytotoxicity, although besides histones, more than 100 proteins from a multitude of
5
2
signaling pathways have been implicated as CBP acetylation substrates.
Optimization of pharmacokinetic (PK) properties.
ACS Paragon Plus Environment
18

Page 19 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
In order to obtain candidates for metabolic stability and in vivo PK studies, we aimed at further
optimizing the prototype dual inhibitors 6c and 6h. To this end, the most favourable compound
features as identified in the initial phase were combined in order to increase the potency toward
the NFꢀκB pathway. In addition, since this could result in predominantly lipophilic compounds,
we aimed at enhancing the PK properties and druglikeness in parallel. As described above, a
chlorine or bromine atom in the meta position of the 4ꢀanilino ring was favorable for the NFꢀκB
inhibitory activity (cf. 4c and VIa, respectively, Table 3). Hence we decided to uniformely
incorporate mꢀchloro due to the smaller size and more moderate impact on logP compared with
bromine. It was furthermore favorable for the potency to include a lipophilic aromatic system on
the thiourea group. Also considering metabolic stability issues, the chlorine as present in
compounds 6c and 6h was replaced by fluorine (9a and 9b). While the 3,4ꢀdifluorophenyl
moiety led to reduced activities both in the NFꢀκB reporter gene and the EGFR kinase assay
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(9a), the 3,5ꢀdifluoro substitution (9b) favorably increased the potency to suppress the activation
of NFꢀκB (IC = 0.3 µM, Table 6) compared with 6c and 6h (IC s: 1.9 and 1.0 µM,
5
0
50
respectively). In addition, the inhibitory potency toward EGFR kinase was still in the nanomolar
range (IC50 = 60.1 nM). The last set of compounds aimed at the enhancement of water solubility
through replacing the thiourea–linked phenyl by different Nꢀheterocycles, comprising pyridine
(
9c), thiazole (9d and 9e) and thiadiazole (9f and 9g). To maintain a suitable balance between
polarity and lipophilicity, all heterocyclic rings were decorated with a lipophilic substituent. The
ꢀchloropyridine (9c) and the electron–rich 5ꢀmembered heterocycles had a distinct influence on
3
the dual activity when compared with 6c: while the pyridine ring slightly decreased both
biological activities to a similar extent, the thiazole and thiadiazole rings reduced the potency
against EGFR kinase by about 5ꢀ6 fold (9dꢀ9g, Table 6). On the other hand, the latter
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compounds showed preserved or even enhanced activity against the NFꢀκB pathway, with 9d
and 9e exhibiting subꢀmicromolar IC s. Altogether, aromatic heterocycles, in particular thiazole,
50
were tolerated while enhancing the water solubility compared with the halogenated phenyl
analogs (cf. 9d vs. 6c and 9b, Table S4, Supporting Information).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The best compounds from this last optimization round, 9b, 9d and 9e were then tested for
growth inhibitory activity against MDAꢀMBꢀ231 cells. While 9b and 9d exhibited comparable
IC values (1.1 and 0.9 µM, respectively), 9e surprisingly lacked activity against this cell line
5
0
even at 20 µM. The reason for this is unclear, but it could be possible that 9e was unstable under
the prolonged cell culture conditions.
In summary, 9b showed a balanced inhibitory profile and was the most potent compound in
the NFꢀκB reporter gene assay (IC = 0.3 µM). Since the substitution pattern of the phenyl was
5
0
changed compared with 6c, we eventually verified that this did not influence the kinase
selectivity. It could be confirmed that the selectivity was retained or became even superior to 6c,
as indicated by the weaker inhibition of Pimꢀ1 and ErbB4 compared with 6c (Table S5).
The 4ꢀ(trifluoromethyl)thiazolyl analog 9d also exhibited a good NFꢀκB inhibitory activity
(IC = 0.6 µM) and efficiently inhibited MDAꢀMBꢀ231 cell growth, although the activity
5
0
toward purified EGFR kinase was decreased (IC = 137 nM). However, since the potency to
5
0
suppress MDAꢀMBꢀ231 cell growth was mainly determined by the NFꢀκB inhibitory activity (cf.
above), both 9b and 9d were selected as candidates for pharmacokinetic studies.
Pharmacokinetic evaluation in vitro and in vivo.
First, the in vitro metabolic stability of lead compounds 9b and 9d was evaluated using rat
liver microsomes. After incubation for defined periods, the remaining percentage of parent
compound was determined by LCꢀMS/MS. Halfꢀlife and intrinsic clearance were calculated for
ACS Paragon Plus Environment
20

Page 21 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the test compound and the two reference compounds diazepam and diphenhydramine (Table 7).
Although 9d was considerably less stable than 9b, reaching only the halfꢀlife and predicted
clearance of diphenhydramine, it was conceivable that this disadvantage might be compensated
by the higher water solubility (Table S4, Supporting Information). In contrast, 9b displayed
higher stability in the microsomal assay than the reference drug diazepam, however, there was
only a slight improvement of the water solubility compared with 6c (Table S4). Because of their
complementary properties, 9d and 9b were both included in the in vivo PK study. As shown in
Figure 2 and Table 8, 9b showed a good plasma halfꢀlife of about 7.8 h, and the compound was
not completely eliminated after 24 h. Despite the lower water solubility, the AUC of 9b was
more than 5ꢀtimes higher than that of 9d (Table S4). This was partially attributable to the more
rapid liver metabolism of 9d, however, the maximum plasma concentration (Cmax) reached by 9d
was lower than expected with a subcutaneous injection, suggesting that further factors impaired
the plasma availability. Finally, 9b was evaluated further in a xenograft tumor model.
Tumor xenograft study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MDAꢀMBꢀ231 cells were injected under the skin of nude mice and the treatment by
compounds or vehicle started after day 5 of the tumor cell inoculation. We decided to use a 10
mg/kg dose of 9b to minimize solubility problems, and 25 mg/kg of the reference drug gefitinib.
The higher dose was chosen to make sure that gefitinib would show at least some measureable
9
effect, since the MDAꢀMBꢀ231 cells were rather insensitive toward this drug (cf. Table 4). In a
previous study, s.c. application was demonstrated to be an efficient administration route for
5
3
gefitinib in a tumor xenograft model in nude mice. In addition, gefitinib exhibits higher water
solubility than our compounds (cf., Table S4) and was reported to achieve high and sustained
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
54
blood levels in mice, suggesting an at least comparable bioavailability of gefitinib to the
transplanted tumor cells in our model.
It can be seen from Figure 3 that the lower dose of 9b inhibited the tumor growth to the same
extent as gefitinib. During the study, tumor growth was unexpectedly slow, possibly impeding
the observation of more pronounced effects of the drugs. Especially in the first two weeks of
treatment, the delayed growth of some tumors caused high variations of the measured sizes. The
tumors then started growing stronger in the vehicle– but not in the compound–treated group, so
that significant differences were reached after day 21 (Figure 3).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Absence of adverse effects in vivo.
In both the pharmacokinetic study (durance 24 h) and the xenograft study (treatment over 28
days), no adverse effects of 9d and/or 9b were observed. No anomalous behavior was seen, and
the organs after killing of the animals appeared morphologically normal. The body weight was
not affected by 9b over the 28 days period. Thus, it can be concluded that the compounds were
nonꢀtoxic to mice and well tolerated at the administered dose.
ACS Paragon Plus Environment
22

Page 23 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Conclusion
Using a serendipity approach, thiourea quinazoline derivatives showing a dual inhibitory
activity towards the EGFR and the NFꢀκB activity were discovered. Optimization of the hit
compound VIa resulted in 9b and 9d as best compounds for NFꢀκB inhibition with cellular IC s
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5
0
in the submicromolar range. We provided in vitro and in vivo evidence that tumor cells which are
only weakly responsive towards EGFR inhibition can still be combatted by the novel NFꢀκB
inhibitory activity. Compared to sole EGFR kinase inhibitors, our dual inhibitors might show
higher efficacy in two scenarios: (i) The “inꢀbuilt” NFꢀκB inhibitory activity might sensitize
EGFR kinase–overexpressing, originally nonꢀresponsive tumors to EGFR kinase inhibition, and
(ii) it might suppress the occurrence of resistant tumor cells, that were initially sensitive toward
1
5
EGFR kinase inhibitors, but amplify NFꢀκB signaling as an adaptive response.
The modulation of the NFꢀκB transcriptional activity by the thiourea quinazoline derivatives
involves depletion of the coꢀactivator CBP, which is a novel mechanism requiring inꢀdepth
characterization in future studies. At any rate, the dual inhibitors exhibited a tumor cell–
selective cytotoxicity.
Based on the encouraging results from our simplified in vivo model, more challenging and
predictive tumor models can be envisaged in the future to test the efficacy of our dual inhibitors,
such as patient–derived cancer tissue xenografts, which preserve the intraꢀtumoral heterogeneity
5
5
of primary malignancies.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Experimental
Chemistry
Solvents and reagents were obtained from commercial suppliers and used as received. 1H and
3C NMR spectra were recorded on a Bruker DRX 500 spectrometer. Chemical shifts are
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
referenced to the residual protonated solvent signals. The purities of all biologically tested
compounds were determined by HPLC coupled with mass spectrometry and were higher than
95% in all cases. Mass spectrometric analysis (HPLCꢀESIꢀMS) was performed on a TSQ
quantum (Thermo Electron Corporation) instrument equipped with an ESI source and a triple
quadrupole mass detector (Thermo Finnigan). The MS detection was carried out at a spray
voltage of 4.2 kV, a nitrogen sheath gas pressure of 4.0 x 105 Pa, an auxiliary gas pressure of 1.0
x 105 Pa, a capillary temperature of 400 ºC, a capillary voltage of 35 V, and a source CID of 10
V. All samples were injected by an autosampler (Surveyor, Thermo Finnigan) with an injection
volume of 10 ꢁL. An RP C18 NUCLEODUR 100ꢀ3 (125 x 3 mm) column (MachereyꢀNagel)
was used as the stationary phase. The solvent system consisted of water containing 0.1% TFA
(A) and 0.1% TFA in acetonitrile (B). HPLCꢀMethod: flow rate 400 ꢁL/min. The percentage of
B started at an initial of 5%, was increased up to 100% during 16 min, kept at 100% for 2 min,
and flushed back to 5% in 2 min. High resolution precise mass spectra were determined for
compounds 6c, 6h, 9b, 9d and were recorded on ThermoFisher Scientific (TF, Dreieich,
Germany) Q Exactive Focus system equipped with heated electrospray ionization (HESI)ꢀII
source and Xcalibur (Version 4.0.27.19) software. Mass calibration was done prior to analysis
according to the manufacturer’s recommendations using external mass calibration. All samples
were constituted in methanol and directly injected onto the Q Exactive Focus using the integrated
syringe pump. All data acquisition was done in positive ion mode using voltage scans and the
ACS Paragon Plus Environment
24

Page 25 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
data collected in continuum mode. Melting points are uncorrected and were determined on Buchi
melting point apparatus (Bꢀ540). The IR spectra were measured on Nicolet 380 FTꢀIR
spectrometer.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
ꢀbenzylꢀ3ꢀ(4ꢀ((3ꢀbromophenyl)amino)quinazolinꢀ6ꢀyl)thiourea (VIa). Yield 63%; m.p.
1
197ꢀ198 °C; H NMR (500 MHz, DMSO) δ 9.91 (s, 1H), 9.83 (s, 1H), 8.64 (s, 1H), 8.47 (d, J =
1
.3 Hz, 1H), 8.42 (s, 1H), 8.25 (t, J = 1.9 Hz, 1H), 7.93 (ddd, J = 8.2, 1.9, 0.9 Hz, 1H), 7.85 (dd,
J = 8.9, 2.1 Hz, 1H), 7.78 (d, J = 8.9 Hz, 1H), 7.38 – 7.31 (m, 5H), 7.30 (ddd, J = 7.9, 1.9, 1.0
13
Hz, 1H), 7.25 (t, J = 7.1 Hz, 1H), 4.79 (d, J = 5.1 Hz, 2H). C NMR (126 MHz, DMSO) δ
1
81.56, 157.11, 153.66, 147.51, 141.07, 139.00, 136.98, 131.83, 131.80, 130.41, 128.21, 127.41,
26.83, 125.86, 123.89, 121.21, 120.43, 117.77, 115.26, 47.51. MS (+ESI): m/z = 464.09 (M +
1
H).
1
ꢀ(4ꢀ((3ꢀbromophenyl)amino)quinazolinꢀ6ꢀyl)ꢀ3ꢀ(3ꢀchlorophenyl)thiourea (6c). Yield
1
6
8
8
0%; m.p. 180ꢀ182°C; H NMR (500 MHz, DMSO) δ 10.19 (s, 1H), 10.06 (s, 1H), 9.84 (s, 1H),
.65 (s, 1H), 8.49 (d, J = 1.8 Hz, 1H), 8.23 (t, J = 1.8 Hz, 1H), 7.94 – 7.84 (m, 2H), 7.80 (d, J =
.8 Hz, 1H), 7.70 (t, J = 2.0 Hz, 1H), 7.43 (d, J = 9.0 Hz, 1H), 7.36 (td, J = 8.0, 4.1 Hz, 2H), 7.30
1
3
(d, J = 8.7 Hz, 1H), 7.20 (ddd, J = 7.9, 2.0, 1.0 Hz, 1H). C NMR (126 MHz, DMSO) δ 180.49,
1
57.15, 153.80, 147.66, 140.96, 140.89, 137.09, 132.49, 132.03, 130.40, 130.00, 128.10, 125.95,
+
1
24.34, 124.05, 123.52, 122.43, 121.18, 120.59, 118.36, 115.13. HRMS (+ESI) m/z: [M + H]
Calcd for C H BrClN S 483.99983; Found 483.99921.
2
1
16
5
1ꢀ(4ꢀ((3ꢀbromophenyl)amino)quinazolinꢀ6ꢀyl)ꢀ3ꢀ(3ꢀchloroꢀ4ꢀfluorophenyl)thiourea (6h).
1
Yield 63%; m.p. 206ꢀ207°C; H NMR (500 MHz, DMSO) δ 10.19 (s, 1H), 9.98 (s, 1H), 9.84 (s,
1
H), 8.65 (s, 1H), 8.49 (d, J = 2.1 Hz, 1H), 8.24 (t, J = 2.0 Hz, 1H), 7.91 (ddd, J = 8.1, 2.0, 1.0
Hz, 1H), 7.86 (dd, J = 8.9, 2.2 Hz, 1H), 7.80 (d, J = 8.9 Hz, 1H), 7.77 (dd, J = 6.8, 2.5 Hz, 1H),
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
.44 (ddd, J = 8.9, 4.6, 2.5 Hz, 1H), 7.41 (d, J = 9.0 Hz, 1H), 7.36 (dd, J = 14.2, 6.1 Hz, 1H),
1
3
7
.30 (ddd, J = 8.0, 1.9, 1.0 Hz, 1H). C NMR (126 MHz, DMSO) δ 180.77, 157.17, 154.42 (d,
1
JC-F = 244.6 Hz), 153.87, 147.75, 140.98, 137.00, 136.59, 132.07, 130.41, 128.21, 126.43,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
2
1
25.97, 125.20 (d, J = 7.2 Hz), 124.05, 121.21, 120.59, 118.71 (d, J = 18.6 Hz), 118.46,
C-F C-F
2
+
116.47 (d, J = 21.8 Hz), 115.18. HRMS (+ESI) m/z: [M + H] Calcd for C H BrClFN S
C-F
21 15
5
501.99041; Found 501.98968.
1
ꢀ(4ꢀ((3ꢀchlorophenyl)amino)quinazolinꢀ6ꢀyl)ꢀ3ꢀ(3,5ꢀdifluorophenyl)thiourea (9b). Yield
1
31%; m.p. 168.5ꢀ170.5°C; H NMR (500 MHz, DMSO) δ 10.36 (s, 1H), 10.25 (s, 1H), 9.87 (s,
1
H), 8.65 (s, 1H), 8.50 (d, J = 2.0 Hz, 1H), 8.11 (t, J = 1.9 Hz, 1H), 7.87 (dd, J = 8.9, 2.1 Hz,
H), 7.84 (dd, J = 8.3, 1.2 Hz, 1H), 7.81 (d, J = 8.8 Hz, 1H), 7.42 (t, J = 8.1 Hz, 1H), 7.36 (dd, J
1
1
3
= 9.3, 2.2 Hz, 2H), 7.17 (ddd, J = 8.0, 2.1, 0.9 Hz, 1H), 6.99 (tt, J = 9.3, 2.3 Hz, 1H). C NMR
(126 MHz, DMSO) δ 180.40, 162.03 (dd, J = 243.3, 15.3 Hz), 157.22, 153.92, 147.80, 142.17 (t,
J = 13.5 Hz), 140.84, 137.00, 132.77, 132.10, 130.13, 128.18, 123.12, 121.30, 120.24, 118.50,
+
115.18, 106.23 (dd, J = 27.9, 6.3 Hz), 99.49 (t, J = 26.0 Hz). HRMS (+ESI) m/z: [M + H] Calcd
for C H ClF N S 442.07048; Found 442.06946.
2
1
15
2
5
1ꢀ(4ꢀ((3ꢀchlorophenyl)amino)quinazolinꢀ6ꢀyl)ꢀ3ꢀ(4ꢀ(trifluoromethyl)thiazolꢀ2ꢀyl)thiourea
1
(9d). Yield 25%; m.p. 226ꢀ228°C; H NMR (500 MHz, DMSO) δ 12.48 (s, 1H), 10.35 (s, 1H),
9.91 (s, 1H), 8.67 (s, 1H), 8.58 (d, J = 2.1 Hz, 1H), 8.10 (s, 1H), 7.95 (dd, J = 8.9, 2.2 Hz, 1H),
7.86 (d, J = 1.0 Hz, 1H), 7.84 (d, J = 8.9 Hz, 2H), 7.42 (t, J = 8.1 Hz, 1H), 7.18 (ddd, J = 8.0,
13
2
1
1
.1, 0.9 Hz, 1H). C NMR (126 MHz, DMSO) δ 177.95, 162.58 (q, J = 4.9, 2.7 Hz), 157.28,
54.12, 147.97, 140.74, 137.42 (q, J = 36.6 Hz), 136.24, 132.75, 131.85, 130.10, 128.25, 123.21,
21.41, 120.93 (q, J = 269.9 Hz), 120.34, 118.71, 115.55 (q, J = 3.3 Hz), 115.15. HRMS (+ESI)
+
m/z: [M + H] Calcd for C H ClF N S 481.02837; Found 481.02740.
19
13
3
6 2
ACS Paragon Plus Environment
26

Page 27 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Biology
EGFR kinase phosphorylation assay. Phosphorylation assays were performed in a final
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
volume of 20 µL containing 8 mM MOPS (pH 7.0), 0.2 mM EDTA, 10 mM MnCl , 200 µM
2
substrate peptide, 0.25 mM DTT, 0.1 mg/mL BSA, 10 ng wildꢀtype EGFRꢀKinase (Cat. No.
32
4
0187, BPS Bioscience), 10 mM magnesium acetate, 100 µM γ–[ P]ATP, and inhibitors or
DMSO control (1.25% v/v). For IC50 curves with the wildꢀtype enzyme, the following
concentrations of the compounds (in nM) were tested in triplicates: 150, 100, 50, 25, 15, 10, 7.5,
5, 2.5. The assays were repeated at least once. Reactions were started by the addition of the
magnesium acetate/ATP mixture. After 30 min incubation at 30°C, 5 µL of each reaction was
spotted on phosphocellulose P81 paper (Whatman). The P81 paper was then washed 5 times with
50 mM phosphoric acid for 15 min, dried and exposed to a phosphorimager screen (Storm, GE
Healthcare), which was scanned and densitometrically analyzed the next day. The sequence of
the substrate peptide was derived from phospholipase Cꢀγ1 and had the sequence
56
“KHKKLAEGSAYEEV”, according to Fry et al..
Reporter Gene Assay. The NFꢀκB reporter gene assay was performed in U937 cells exactly
5
7
as previously described.
MDAꢀMBꢀ231 cell growth inhibition assay. Cells were seeded in 96ꢀwell standard assay
microplates at a density of 45,000 cells per well, then allowed to adhere overnight before
compound addition. After 24 hours, cells were treated with different concentrations of the
compounds. Cells were incubated for an additional 48 hours at 37 °C, after which 20 µL of MTT
reagent (prepared as 5 mg/mL PBS) were added and then incubated for additional 1 hour. After
that, 100 µL SDS (prepared as 10% in 0.01 N HCl) were added and incubated for at least 2 h at
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 58
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
7 °C to allow for cell lysis. Absorbance was then measured at a wavelength of 570nm in a plate
reader (PolarStar, BMG Labtech, Freiburg, Germany).
Cultivation of HUVEC, CHOꢀK1 and A549 cells for cytotoxicity assays. HUVEC cells
purchased from the American Type Culture Collection (Manassas, VA) were maintained and
seeded in Fꢀ12K nutrient mixture (Kaighn's modification; Life Technologies) with 10% FBS,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
00 ꢁg/mL heparin (SigmaꢀAldrich, St. Louis, MO), and 30 ꢁg/mL endothelial cell growth
3
supplement (SigmaꢀAldrich). For the MTT assay, 6 x 10 cells per well of 96ꢀwell plates were
used. CHOꢀK1 cells (ACCꢀ110) were obtained from the German Collection of Microorganisms
and Cell Cultures (Deutsche Sammlung für Mikroorganismen und Zellkulturen, DSMZ) and
3
were cultured under conditions recommended by the depositor. Cells were seeded at 6 x 10 cells
per well of 96ꢀwell plates in 180 ꢁL complete medium (F12K, 10% FBS). A549 cells obtained
from the “Deutsche Sammlung für Mikroorganismen und Zellkulturen” (DSMZ) (ACC 107)
were grown in Dulbecco’s modified Eagle’s medium (DMEM, Sigma) with 10 % fetal calf
3
serum (FCS, Sigma). Cells were also seeded at 6 x 10 cells per well of 96ꢀwell plates.
All cell media contained in addition penicillin G (final concentration 100 U/mL) and
streptomycin sulfate (final concentration 100 mg/mL), and were maintained at 37° C and 5%
CO in a humidified incubator. The cytotoxicity of the test compound was determined as
2
described above, under “MDAꢀMBꢀ231 cell growth inhibition assay”.
NFꢀκB (p65)ꢀDNA binding and p65/CREBꢀbinding protein ELISA. MDAꢀMBꢀ231 breast
cancer cells were grown in 6ꢀwell plates to 80% confluency in DMEM containing 10 % FCS
(low endotoxin quality, GIBCO, Thermo Fisher Scientific), and pen/strep mix. The cells were
incubated with the test compounds or DMSO as solvent control for 12 h. In some experiments,
TNFα was added at final concentration of 20 ng/mL (as indicated) and the incubation continued
ACS Paragon Plus Environment
28

Page 29 of 58
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
for 20 min at 37°C. The plates were then put on ice, washed twice with iceꢀcold PBS and
harvested in 1 mL PBS using a cell scraper (Sigma, cat. no. CLS3010). The suspension was
transferred to Eppendorf tubes on ice, and the cells were spun down at 500 g in a cooled
centrifuge (2° C). Nuclear protein extracts were immediately prepared using the Pierce NEꢀPER
Nuclear and Cytoplasmic Extraction Kit (Thermo Fisher Scientific, cat. no. PIꢀ78833) according
to the manufacturer´s instructions. Protein concentrations were determined using RotiꢀQuant
Bradford reagent (Carl Roth, cat. no. K015.3) against a calibration curve generated with BSA as
a standard. The absorbance values were measured at 595 nm in a BioPhotometer Plus
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(
Eppendorf, Germany). 5 µg of each nuclear extract sample was then analyzed using a NFꢀκB
p65) Transcription Factor Assay Kit (Cayman Chemicals, cat. no. 10007889). In brief, samples
(
were incubated with a specific double stranded DNA oligonucleotide containing the NFꢀκB
response element, which was immobilized on a 96ꢀwell plate. After washing, the NFꢀκBꢀ
containing complexes which bound to the immobilized DNA were detected using a specific
primary antibody directed against NFꢀκB (p65). In some experiments, CREBꢀbinding protein as
a component of the NFꢀκB transcription factor complex was detected instead, using antiꢀCBP
Antibody (Santa Cruz Biotechnology, cat. no. scꢀ583), at a 1:100 dilution. A secondary antibody
conjugated to HRP was added to provide a colorimetric readout at 450 nm. The absorbance
values were measured in a PolarStar plate reader (BMG Labtech, Germany). Specificity of the
binding was verified by adding 10 µL of the NFꢀκB response element oligonucleotide solution
(GGGACTTTCC, provided by the manufacturer), in a final volume of 100 µL, during the
incubation with nuclear extracts.
Metabolic Stability Assay. The assay was performed with liver microsomes from male
58
SpragueꢀDawley rats (BD Bioscience, Catalog # 452501) exactly as described previously.
ACS Paragon Plus Environment
29