Optimization of pyrrolizine-based Schiff bases with 4-thiazolidinone motif: Design, synthesis and investigation of cytotoxicity and anti-inflammatory potency

Article abstract of DOI:10.1016/j.ejmech.2019.111780

Two new series of pyrrolizine-5-carboxamides were synthesized and evaluated for their anticancer and anti-inflammatory activities. The new compounds exhibited potent cytotoxicity (IC₅₀ = 0.10–22.96 μM) against three cancer (MCF-7, A2780 and HT2

Full text of DOI:10.1016/j.ejmech.2019.111780

Journal Pre-proof
Optimization of pyrrolizine-based Schiff bases with 4-thiazolidinone motif: Design,
synthesis and investigation of cytotoxicity and anti-inflammatory potency
Ahmed M. Shawky, Mohammed A.S. Abourehab, Ashraf N. Abdalla, Ahmed M.
Gouda
PII:
S0223-5234(19)30932-8
DOI:
https://doi.org/10.1016/j.ejmech.2019.111780
EJMECH 111780
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 1 September 2019
Revised Date: 4 October 2019
Accepted Date: 9 October 2019
Please cite this article as: A.M. Shawky, M.A.S. Abourehab, A.N. Abdalla, A.M. Gouda, Optimization
of pyrrolizine-based Schiff bases with 4-thiazolidinone motif: Design, synthesis and investigation of
cytotoxicity and anti-inflammatory potency, European Journal of Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.ejmech.2019.111780.
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.

Optimization of pyrrolizine-based Schiff bases with 4-thiazolidinone
motif: design, synthesis and investigation of cytotoxicity and anti-
inflammatory potency
Ahmed M Shawky1, Mohammed A.S. Abourehab2,3, Ashraf N. Abdalla4,5, Ahmed M. Gouda6,7,*
1Science and Technology Unit (STU), Umm Al-Qura University, Makkah 21955, Saudi Arabia;
2Department of Pharmaceutics, Faculty of Pharmacy, Umm Al-Qura University, Makkah 21955,
Saudi Arabia;
3Department of Pharmaceutics, Faculty of Pharmacy, Minia University, Minia 61519, Egypt;
4Department of pharmacology, Faculty of Pharmacy, Umm Al-Qura University, Makkah 21955,
Saudi Arabia;
5Department of Pharmacology, Medicinal and Aromatic plants research institute, National center for
research, Khartoum 2404, Sudan.
6Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Umm Al-Qura University, Makkah
21955, Saudi Arabia;
7Medicinal chemistry department, Faculty of pharmacy, Beni-Suef University, Beni-Suef 62514,
Egypt;
__________________________________________________
*Correspondence: Ahmed M. Gouda: Department of Medicinal Chemistry, Faculty of Pharmacy,
Beni-Suef University, Beni-Suef 62514, Egypt.
Scopus ID: 26321547200
ORCID: 0000-0003-4527-8885
Tel.: (002)-01126897483
Fax: (002)-082-2162133
E-mail address: ahmed.gouda@pharm.bsu.edu.eg or amsaid@uqu.edu.sa
Page 1 of 55

Highlights
1- Two new series of pyrrolizines bearing Schiff bases or 4-thiazolidinone moiety were synthesized.
2- Compound 12 displayed potent cytotoxic activity (IC50 = 0.10 µM) against MCF-7 cells.
3- Compound 22 induced G1 cell cycle arrest and apoptosis in MCF-7 cells.
4- Compounds 12, 19 and 22 displayed inhibitory activity against CDK2 (IC50 = 0.58-1.27 µM).
5- Compounds 14 exhibited potent in vivo anti-inflammatory activity mediated by COX-2 inhibition
(IC50 = 0.64 µM).
6- Binding modes of the new compounds into COXs, CDK2, Aurora A, BRAF and EGFR enzymes
were investigated.
Graphical abstract
Key words:
Pyrrolizine; cytotoxicity; COX; Kinase; Apoptosis; Cell cycle analysis.
Page 2 of 55

Abstract
Two new series of pyrrolizine-5-carboxamides were synthesized and evaluated for their
anticancer and anti-inflammatory activities. The new compounds exhibited potent
cytotoxicity (IC50 = 0.10-22.96 µM) against three cancer (MCF-7, A2780 and HT29) cell
lines with selectivity index in the range of 1-258. Moreover, these compounds also exhibited
significant anti-inflammatory activity (18.13-44.51% inhibition of inflammation) mediated by
inhibition of COX-1/2 with preferential inhibition of COX-2. The study of SAR revealed
favorable cytotoxic outcomes of the aliphatic side chain and 4-thiazolidinone moiety at C6 of
the pyrrolizine nucleus, while anti-inflammatory activities was improved with the
(hetero)aromatic substituents. The IC50 values which inhibit COX-2 were higher than those
needed to inhibit the growth of cancer cell lines. Mechanistic studies also revealed inhibition
of multiple kinases by compounds 12, 19 and 22. Moreover, compounds 12, 14, 16 and 22
induced cell cycle arrest and apoptosis in MCF-7 cells. Docking studies revealed nice fitting
of the new compounds into COX-1/2. Additionally, compounds 12, 19 and 22 also exhibited
higher affinity for CDK2 than CAN508. To sum up, the above-mentioned data highlight
these compounds as promising anti-inflammatory and anticancer agents.
Page 3 of 55

1. Introduction
Combination therapy represents the cornerstone in the treatment of complex diseases [1].
The high efficacy of combination compared to cancer monotherapy is due to simultaneous
targeting of multiple enzymes or signaling pathways involved in cancer [2,3]. However, the
wide use of combined drugs was restricted by some limitations including the high cost, drug
toxicity and interactions [1,3,4]. These problems can be solved with the use of multi-target
anticancer agents [5]. A single multi-targeted agent can be used to avoid the pharmacokinetic
and drug interaction problems associated with combination therapy [5,6]. Several multi-
targeted agents are used effectively in treatment of cancers [7]. However, acquired resistance
was also developed for these drugs [8,9].
NSAIDs are commonly in management of cancer-related inflammation and pain [10].
Moreover, their chemopreventive effect attracted much attention in the last two decades. The
relationship between inflammation and cancer was also discussed in many reports [12,13].
The upregulation of cyclooxygenase 2 (COX-2) gen was confirmed in different types of solid
tumors [14,15]. Accordingly, targeting COX-2 enzyme could provide a new approach in
cancer prevention and therapy. This approach was supported by the high anticancer activity
of some selective COX-2 inhibitors [16]. Moreover, other types of COXs inhibitors also
exhibited anticancer activity against diverse types of cancer cells [17]. However, the exact
mechanism of anticancer activity of COXs inhibitors is still unclear [16]. In addition, their
use in prevention and treatment of cancers was restricted by many dangerous side effects
[18].
Recently, the diarylpyrrolizine derivatives were in our focus as potential therapeutic agents
with polypharmacological profile. Among these derivatives, licofelone 1 and compound 2 (Fig.
1) displayed both anti-inflammatory and anticancer activities [19-22]. The anticancer activity of
compounds 1 and 2 was mediated by induction of apoptosis in colon (HCA-7) and breast
(MCF-7) cancer cells, respectively [21,22]. Although licofelone 1 has dual COX/lipoxygenase
(LOX) inhibitory activity, the induction of apoptosis in HCA-7 cells was mediated by
inhibition of the epidermal growth factor receptor (EGFR) and its downstream kinases [22],
which was mediated by decrease in cell membrane fluidity rather than direct inhibition of the
enzyme.
Combination of licofelone 1 with several anticancer agents enhanced their anticancer efficacy
against different types of cancers [23-25]. Specifically, combining licofelone with the EGFR
Page 4 of 55

inhibitor gefitinib showed synergistic growth and progress suppression of pancreatic
carcinoma in mice [25]. Similar results were also observed when erlotinib was combined with
selective/nonselective COX inhibitors [26-28].
Rational design
The rational design of multi-target agents can be achieved by incorporating two or more
pharmacophores into one scaffold [29-30], which can target different enzymes/pathways.
Encouraged by these findings, we have designed two new scaffolds (A and B) to target both
oncogenic protein kinases (PKs) and COX enzymes, simultaneously. The first scaffold A (Fig. 1)
was designed by incorporating the pharmacophoric features of the COX inhibitor licofelone 1 and
those of the tyrosine kinase inhibitor CGP-59326 3 [31]. Structure activity relationship (SAR) of
both licofelone 1 and CGP-59326 3 was considered in the design of scaffold A. Structural
modifications of scaffold A included expansion of the pyrrolidine ring (n = 1, 2) and variation of
the side chain at C6 to include aliphatic, aromatic or heteroaromatic moieties, Fig. 1.
Fig. 1. Rational design and structural modifications of scaffold A.
Moreover, several mono/disubstituted 4-thiazolidinone-based derivatives (Fig. 2) exhibited
potent anticancer activity with IC50 values in the nanomolar rang [32-37]. Mechanistic studies
of these compounds revealed inhibition of diverse oncogenic kinases. Among these
derivatives, compounds 4 and 5 displayed potent inhibition of cyclin dependent kinase 2
(CDK2) and aurora A kinase, respectively [32,34]. In another attempt to target both
oncogenic protein kinases (PKs) and COX enzymes, simultaneously, we have designed
scaffold B by incorporating the 4-thiazolidinone ring and scaffold A. The SAR of scaffold B
Page 5 of 55

was studied by variation of substituents on phenyl ring and expansion of pyrrolidine ring (n =
2), Fig. 2.
Fig. 2. Rational design and structural modifications of scaffold B.
2. Results and discussion
2.1. Chemistry
Preparation of compounds 7, 9a-d and 10a-d was described in Scheme 1, following the
previous reports [38-41]. The starting materials 10a-d were selected out of 28 pyrrolizine
derivatives bearing substituents with different sigma (σ) constants on the phenyl ring.
Initially, the 28 pyrrolizine underwent preliminary evaluation of their binding affinity to
COXs, synthetic accessibility and drugs-likeness scores (Table S1 and S2, supplementary
data). In this study, compounds 10a-d showed considerable affinity for COX-1/2.
Considering the structural differences between the active site of COX-2 and COX-1 [42], it is
expected that affinity of the new compounds 11-18 to COX-2 will be optimized with the
increase in their molecular volume as compared to the starting materials. Compounds 11-18
were obtained by condensation of the pyrrolizine-5-carboxamide derivatives 10a with the
appropriate aldehydes in ethanol, Scheme 1. The new compounds were obtained in good
yield (61-75%), Table 1.
Page 6 of 55

Scheme 1. Reagents and reaction conditions: (a) (CH3)2SO4, benzene, CH2(CN)2; (b) ClCH2COCl,
gl. AcOH, CH3COONa; (c) acetone, K2CO3, reflux, 24 h; (d) appropriate aldehyde, EtOH, gl. AcOH,
stir 2h at 40-50 ○C, reflux, 6 h.
Table 1. Melting points and yield% of compounds 11-24, 28 and 29.
Comp.
No
11
m.p.
(○C)
Scaffold R1
R2
n
Yield%
A
A
A
A
A
A
A
A
A
A
A
B
B
Br
Br
Br
Br
Br
Br
Br
Br
H
C2H5
CH2CH(CH3)2
C6H5
4-OH-C6H4
4-Cl-C6H4
2-furanyl
2-thiophenyl
3-pyridinyl
C2H5
1
1
1
1
1
1
1
1
1
1
1
1
1
227-9
207-9
265-7
320-2
285-7
261-3
250-2
271-3
212-4
217-9
226-9
228-9
237-40
61%
64%
71%
73%
75%
69%
65%
71%
65%
59%
71%
58%
63%
12
13
14
15
16
17
18
19
20
21
22
23
CH3 C2H5
Cl
H
CH3
C2H5
-
-
Page 7 of 55

B
A
B
Cl
H
H
-
1
2
2
249-51
209-11
228-31
68%
54%
56%
24
28
29
C2H5
-
To optimize activity of the new compounds, a preliminary evaluation of their cytotoxic
activity was performed. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium
bromide (MTT) assay was used in this evaluation according to the previous report [43]. In
this evaluation, MCF-7 breast cancer cell line was selected representing the leading cancer
in incidence and mortality among women [44]. Cancer cells were treated with single dose
(5 µM) of the eight compounds 11-18 and growth% was determined after 48 h. Results
were represented in Fig. 3.
Fig. 3. Growth% of MCF-7 cells after single-dose treatment with compounds 11-18 (5 µM) for 48 h,
results were presented as mean ± S.D. (n = 6).
Results of the preliminary cytotoxic evaluation revealed the highest growth inhibitory activity
for compounds 11 and 12 which have aliphatic side chain at C6. Compounds 13-15, with the
aromatic rings in the side chain at C6, were slightly more active in inhibiting the growth of
MCF-7 cells compared to compounds 16-18 with the heteroaromatic rings. Among the eight
derivatives, compound 18 was the least active, while compound 11 displayed the highest
cytotoxicity. Based on these results (Fig. 3), three new derivatives 19-21 bearing the aliphatic
propylideneamino moiety at C6 were prepared by condensation of compounds 10b-d with
propionaldehyde in absolute ethanol, Scheme 2. Moreover, three 4-thiazolidinone-bearing
derivatives 22-24 were obtained from the reaction of compounds 10b-d with
Page 8 of 55

propionaldehyde and mercaptoacetic acid. The yield% of these compounds was presented in
Table 1.
Scheme 2. Reagents and reaction conditions: (a) propionaldehyde, EtOH, gl. AcOH, stir, 2h at 40-50
○
○C, reflux, 6 h; (b) propionaldehyde, mercaptoacetic acid, DCC, THF, 0 C; (c) benzaldehyde, gl.
AcOH, Na acetate, reflux, 12 h.
Compounds 22-24 were also refluxed with benzaldehyde to afford the condensed product Ia-
c, Scheme 2. Unfortunately, the expected products Ia-c were not formed.
Moreover, compound 10a was allowed to react with aromatic/heteroaromatic aldehydes in
order to prepare a new series of 2-(hetero)aryl-thiazolidin-4-one derivatives IIa-f, Scheme 3.
But also this reaction failed to produce the expected compounds IIa-f. This could be
attributed to the steric hinderance of the bulky (hetero)aromatic moieties. This was also in
concordance with the synthetic accessibility scores calculated for these compounds using
SwissADME
(developed
by
the
Swiss
Institute
of
Bioinformatics,
http://www.swissadme.ch/). The scores calculated for compounds IIa-f (Table S3,
supplementary data) were higher than that of compound 22.
Page 9 of 55

Scheme 3. Reagents and reaction conditions: (a) Ar-CHO (benzaldehyde, 4-hydroxy/chloro-
benzaldehyde, furan-2-carbaldehyde, thiophene-2-carbaldehyde, 3-pyridinecarbox-aldehyde),
mercaptoacetic acid, DCC, THF, 0 ○C.
Synthesis of the tetrahydroindolizines 28 and 29 was described in Scheme 4. The starting
materials 26 and 27 were synthesized according to the previous reports [40,45]. Compound
27 was reacted with propionaldehyde to afford the Schiff base 28, while compound 29 was
obtained from the reaction of compound 27 with propionaldehyde and mercaptoacetic acid.
The yield% of these compounds was presented in Table 1.
Scheme 4. Reagents and reaction conditions: (a) (CH3)2SO4, benzene, CH2(CN)2; (b) acetone,
○
K2CO3, reflux, 24 h; (c) propionaldehyde, EtOH, gl. AcOH, stir 2h at 40-50 C, reflux, 6 h; (d)
propionaldehyde, mercaptoacetic acid, DCC, THF, 0 ○C.
2.2. Biological evaluation
2.2.1. Cytotoxic activity
2.2.1.1. Cytotoxicity assay
To evaluate growth inhibitory activity of compounds 11-24, 28, 29 and lapatinib as reference
drug, MTT assay was used according to our previous report [43]. In this assay, three human
Page 10 of 55

cancer cell lines (breast MCF-7, ovarian A2780 and colon HT29) were used. The results of
the treatments with compounds 11-18 for 72h were summarized as IC50 values in Table 2.
The new compounds 11-18 displayed potent to moderate cytotoxic activity against the three
cancer cell lines with IC50 values in the range of 0.1-22.96 µM. Among these derivatives,
compounds 11 and 12 exhibited the highest cytotoxic activity (IC50 = 0.10-0.89 µM).
Compound 12 was slightly more active than compound 11 against both MCF-7 and HT29
cancer cell lines, while compound 11 exhibited higher activity against ovarian A2780 cells.
The two compounds were also more potent compared to the reference lapatinib (IC50 = 6.80-
12.67 µM) against the three cancer cell lines, Table 2.
Compounds 13-15 displayed potent to moderate cytotoxic activity against the three tested
cancer cell lines with IC50 values in the range of 0.47-14.48 µM. Among these three
derivatives, the 4-hydroxyphenyl analog 14 was the most active against MCF-7 and A2780
cancer cells, Table 2..
Table 2. Cytotoxic activity of compounds 11-18, and lapatinib against MCF-7, A2780, HT29
and MRC5 cell lines.
IC50 (µM)a,b
Comp.
No
R
MCF-7
A2780
HT29
MRC5
-CH2CH3
0.14±0.01
0.27±0.08
0.52±0.02
0.89±0.01 21.25±0.40
0.67±0.33 25.81±0.73
11
12
-CH2CH(CH3)2 0.10±0.05
0.50±0.15
4.16±1.27
2.97±0.14
9.00±0.13
13
14
15
16
17
0.49±0.02
6.03±3.35
0.47±0.04 14.48±4.44 25.03±0.31
8.30±2.02 7.79±1.84 13.85±1.04
6.89±2.84 12.23±1.27 2.04±0.58 15.68±0.40
9.41±1.82 19.30±2.13 3.65±0.79 23.10±2.03
Page 11 of 55

11.29±1.54 22.96±4.17 2.63±0.35 25.56±2.86
6.80± 1.20 10.40± 0.80 12.67±1.33 13.66±2.90
18
-
Lapatinib
a IC50 is the concentration of test compound which reduce cellular growth to 50% after treatment for 72 h.
b Results represent mean IC50 value ± S.D. (n = 3).
Compounds 16-18 with the heteroaromatic rings in the side chain at C6 were less active
against both MCF-7 and A2780 cancer cell lines than compounds 13-15, meanwhile, they
were more active against HT29 cells (IC50 = 2.04-3.65 µM), Table 2.
Cytotoxic activity of the four propylideneamino-bearing derivatives 19-21 and 28 expressed
as IC50 values were presented in Table 3. The four compounds showed potent to moderate
cytotoxic activity (IC50 = 0.14-15.60 µM) against the three-cancer cell lines. Compound 19
was the most active against MCF-7 cells, while compounds 21 and 28 were the most active
against A2780 and HT29 cells, respectively.
Table 3. Cytotoxic activity of compounds 19-21, 28 and lapatinib against MCF-7, A2780,
HT29 and MRC5 cell lines.
IC50 (µM)a,b
Comp.
R
n
No
19
20
MCF-7
A2780
HT29
MRC5
H
1
1
0.57±0.06
15.60±0.99
0.20±0.01
3.17±0.23
CH3
0.84±0.14
3.88±1.08
0.61±0.08
0.56±0.01
Cl
H
-
1
2
-
0.74±0.04
1.43±0.37
0.58±0.08
1.02±0.06
0.14±0.01
0.17±0.02
1.53±0.16
21
15.56±5.64
28
6.80± 1.20 10.40± 0.80
12.67±1.33 13.66±2.90
Lapatinib
a IC50 is the concentration of test compound which reduce cellular growth to 50% after treatment for 72 h.
b Results represent mean IC50 value ± S.D. (n = 3).
Page 12 of 55

Among the three cancer cell lines, HT29 was the most sensitive to the four compounds 19-21
and 28, Table 3. On the other hand, A2780 cells was the least sensitive to these compounds.
The three pyrrolizines 19-21 displayed higher activity against the MCF-7 cells compared to
the tetrahydroindolizine 28. Moreover, compound 19 was more active against the three
cancer cell lines compared to its corresponding analog 28.
The 4-thiazolidinone-bearing compounds 22-24 and 29 were also evaluated for their
cytotoxicity by the MTT assay. The results revealed potent cytotoxic activity with IC50 values
in the range of 0.11-4.24 µM, against the three cancer cell lines, Table 4. It was clear that
these compounds have better cytotoxic activity profile compared to their opened analogs 19-
21 and 28. Among the four thiazolidinone-bearing derivatives, compound 22 exhibited the
highest cytotoxic activity against MCF-7 cells, while compound 23 was the most active
against both A2780 and HT29 cells.
Table 4. Cytotoxic activity of compounds 22-24, 29 and lapatinib against MCF-7, A2780,
HT29 and MRC5 cell lines.
IC50 (µM)a,b
Comp. No
R
n
MCF-7
A2780
HT29
MRC5
H
1
1
0.16±0.08
0.22±0.07
0.60±0.10
0.63±0.09
22
23
CH3
0.47±0.52
0.11±0.01
0.12±0.03
1.70±0.78
Cl
H
1
2
-
0.21±0.04
0.84±0.03
6.80± 1.20
0.31±0.06
4.24±0.75
10.40± 0.80
0.20±0.01
2.06±0.07
12.67±1.33
0.91±0.10
11.10±0.62
13.66±2.90
24
29
-
Lapatinib
a IC50 is the concentration of test compound which reduce cellular growth to 50% after treatment for 72 h.
b Results represent mean IC50 value ± S.D. (n = 3).
2.2.1.2. Evaluation of cytotoxic selectivity
Page 13 of 55

Selectivity toward cancerous cells is one of the major problems which usually appear during
the discovery and development of new anticancer agents. To assess the selectivity of our
newly synthesized compounds, their cytotoxic activity was evaluated against normal human
foetal lung fibroblast (MRC5). MTT assay was used in this evaluation and results were
expressed as IC50 values, Tables 2-4.
Compounds 11, 12, 14-18 displayed IC50 values in the range of 13.85-25.81 µM against
normal MRC5 cells, indicating less toxicity compared with lapatinib (IC50 values = 13.66
µM). Among these compounds, compound 12 was the least toxic for the normal MRC5 cells,
Table 2.
Selectivity index (SI) of the new compounds was calculated to compare their cytotoxic
selectivity. The value of SI of each compound was obtained by dividing IC50 value of the
compound against the normal MRC5, by IC50 value of this compounds against the specific
cancer cell line.
The Schiff bases 11-18 showed SIs in the range of 1-258, Fig. 4. Among these, compounds
11 and 12 displayed the highest cytotoxic selectivity. Compound 11 was 24-152 times more
selective for the three cancer cell lines. Moreover, compound 12 was the most selective,
showing SI in the range of 39-258 toward the three cancer cell lines.
Compounds 13-15 (with the aromatic rings at C6) exhibited SI of 2-52 for the three cancer cell
lines, being less selective than the aliphatic analogs 11 and 12. On the other hand, compounds
16-18 were the least selective (SI = 1-10) among the eight pyrrolizines 11-18. Except for
compound 16, all other compounds were more selective than lapatinib for the three cancer cell
lines, Fig. 4.
Page 14 of 55

Fig. 4. Selectively index of compounds 11-18 and lapatinib for MCF-7, A2780 and HT29 cancer
cells.
Compounds 19-24, 28 and 29 were also evaluated for their selectivity against MRC5 cells
using MTT assay. Compounds 19-24, 28 and 29 displayed IC50 values in the range of 0.17-
11.1 µM against the normal MRC5 cells. Among these, the tetrahydroindolizine 29 was the
least toxic toward MRC5 cells, Tables 3 and 4.
The SIs of compounds 19-24, 28 and 29 toward the three cancer cell lines was outlined in
Fig. 5. Generally, the 4-thiazolidinone-bearing derivatives 22-24 and 29 displayed SIs in the
range of 3-15 toward MCF-7 and A2780 cells. They were more selective compared to their
opened analogs 19-21 and 28. Among these, compounds 19, 23 and 29 exhibited the highest
selectivity toward HT29, A2780 and MCF-7 cells, respectively.
Fig. 5. Selectively index of compounds 19-24, 28, 29 and lapatinib for MCF-7, A2780 and HT29 cancer
cells.
2.2.1.3. Structure activity relationship (SAR)
The SAR of cytotoxic activity of compounds 11-18, 19-21 and 28 was outlined in Fig. 6.
Compounds 11 and 12 with the aliphatic side chain (R2 = ethyl, isobutyl) were the most active
as cytotoxic agents against the three (MCF-7, A2780 and HT29) cancer cell lines. Replacement
of bromo (Br) in compound 11 by H, electron donating (CH3) or electron withdrawing (Cl)
groups resulted in a strong drop in cytotoxic activity against A2780 cells. Replacement of the
aliphatic side chain (R2) in compounds 11 and 12 with (hetero)aromatic (phenyl, 4-
(hydroxy/chloro)phenyl, 2-furanyl, 2-thiophenyl or 3-pyridinyl) rings resulted in decrease in
Page 15 of 55

cytotoxicity against the three cancer cell lines. In addition, expansion of the pyrrolidine ring
in compound 19 to piperidine decreased the cytotoxic activity by nearly ~50%.
Fig. 6. SAR of cytotoxicity of compounds 11-21 and 28 against MCF-7, A2780 and HT29 cells.
The SAR study of compounds 22-24 and 29 was presented in Fig. 7. Cyclization of the
propylideneamino side chains in compounds 19-21 and 28 into 4-thiazolidinone ring resulted
in improvement of cytotoxicity against both MCF-7 and A2780 cells. Cytotoxicity of
compound 22 against HT29 cells was also improved with the 4-methyl (+σ) and 4-chloro (-σ)
substituents. On the other hand, expansion of the pyrrolidine ring in compound 22 to
piperidine resulted in a noticeable decrease in cytotoxicity against the three cancer cell lines.
Fig. 7. SAR of cytotoxicity activity of compounds 22-24 and 29 against MCF-7, A2780 and HT29 cells.
Page 16 of 55

2.2.2. Anti-inflammatory activity
2.2.2.1. In vitro COX inhibitory activity
Licofelone 1 displayed in vivo anti-inflammatory activity mediated at least partially by
inhibition of COXs enzymes. In this work, our compounds 11-24, 28 and 29 were evaluated
for their in vitro inhibitory activity against COXs using COX inhibitor screening assay kit
according to the previous report [46]. The results were presented as IC50 values, Table 5.
In general, most of the new compounds inhibited the two COXs with preferential inhibition
of COX-2. Compounds 11-24, 28 and 29 displayed IC50 values in the range of 0.64-56.1 µM
against COX-2. Among these, compound 14 was the most potent as COX-2 inhibitor, while
compound 16 was the most active inhibitor of COX-1. Compounds 11 and 29 displayed
nearly equal inhibitory activity against the two COXs (SI = ~1). Although compounds 11 and
12 have nearly equal IC50 values against COX-2 activity, but compound 12 was more
selective toward COX-2, Table 5.
Table 5. COX inhibitory activity of compounds 11-24, 28, 29, celecoxib and indomethacin.
IC50 (µM) a
COX-2
14.15±0.85 13.31±0.77
SI b
COX-1
Comp. No
1.06
>7
>67
112.45
>13
11
12
13
14
>100
13.49±0.63
1.49±0.05
0.64±0.02
7.53±0.34
>100
71.97±2.9
>100
15
13.16±0.64 1.65±0.04
7.98
>10
>10
>1.89
2.68
3.68
4.05
>80
16
17
18
19
20
21
22
23
>100
>100
>100
9.25±0.44
9.84±0.39
52.97±2.71
27.69±1.95 10.33±0.73
15.62±0.71 4.24±0.11
43.72±2.9
>100
>100
10.79±0.67
1.25±0.03
1.15±0.02
7.14±0.25
>86
24
28
29
27.22±2.1
3.81
1.17
111.32
0.02
65.39±3.61 56.10±2.82
27.83±14.2 0.25±0.12
Celecoxib
Indomethacin
1.56±0.03
68.45±4.1
a Values represent the mean of four determinations ± SEM, (n = 3).
b selectivity index (SI) for COX-2 = IC50 against COX-1/IC50 against COX-2.
Page 17 of 55

On the other hand, the presence of a second aromatic ring in the side chain at C6 of the
pyrrolizine scaffold enhanced COX-2 inhibitory activity and selectivity. Compounds 13-15
inhibited activity of COX-2 with IC50 values in the range of 0.64-7.53 µM. Among these,
compound 14 was the most active and selective as COX-2 inhibitor. Replacement of the
(un)substituted phenyl in compounds 13-15 with heteroaryl ring resulted in remarkable
decrease in COX-2 inhibitory activity and selectivity. Compounds 16-18 exhibited weaker
inhibition in activity of COX-2 compared to compounds 13-15 with pronounce decrease in
selectivity, Table 5.
Compounds 19-21 displayed inhibitory activity against COX-2 with IC50 values in the range
of 4.24-52.97 µM. Among these, compound 21 was the most potent as COX-2 inhibitor.
Cyclization of the 6-propylideneamino side chain in these compounds into thiazolidinone
ring improved COX-2 inhibitory activity and selectivity. On the other hand, the 4-
thiazolidinone-bearing indolizine 29 was less active/selective compared to its opened analog
28, Table 5.
2.2.2.2. In vivo Anti-inflammatory activity
Based on the results of COXs inhibitory assay, compounds 12, 14, 16, 21 and 24 (the most
active COX-2 inhibitors) were selected for evaluation of their in vivo anti-inflammatory
activity. The reduction in the carrageenan-induced inflammation was used as a measure of
their anti-inflammatory activity [47]. Test samples, celecoxib and indomethacin were given
orally. The experimental procedures were completed according to our previous report [48].
The results presented in Table 6, revealed the ability of the tested compounds to inhibit
inflammation with inhibition% in the range of 18.13-44.51%, at 2-4 h after injection of
carrageenan.
Table 6. Mean change in edema thickness and anti-inflammatory activity of compounds 12, 14,
16, 21, 24, celecoxib and indomethacin.
Change in edema ± SEM
(% inhibition of edema)
Compound
2h
4h
0.97±0.06 (-)
1.11±0.07 (-)
Control
12
0.79±0.09 (18.13)
0.69±0.05 (28.15)
0.78±0.13 (18.83)
0.78±0.07 (19.17)
0.77±0.11 (30.82)**
0.62±0.05 (44.51)***
0.71±0.07 (35.79)**
0.84±0.08 (24.06)**
14
16
21
Page 18 of 55

0.76±0.06 (21.76)
0.75±0.07 (32.48)**
24
0.61±0.03 (36.96)* 0.36±0.02 (67.37)***
0.65±0.03 (32.30)* 0.48±0.04 (56.39)***
Celecoxib
Indomethacin
Data expressed as mean ± SEM, (n = 6); data were analyzed by One way ANOVA followed by
student-Newman-Keuls multiple comparison test, n = 6; *P < 0.05 compared to control, **P < 0.01
compared to control, ***P < 0.001 compared to control.
Compound 14 was the most potent as anti-inflammatory agent. It displayed 44.51% inhibition
of inflammation (4 h post-carrageenan) compared to 67.37% and 56.39% for celecoxib and
indomethacin, respectively. The (hetero)aromatic-bearing derivatives 14 and 16 exhibited
higher anti-inflammatory activity as compared with compound 12 (4h post-carrageenan).
Moreover, the thiazolidinone-bearing derivative 24 displayed slightly higher anti-
inflammatory activity compared to its corresponding Schiff base 21. However, all of the test
compounds were less active than celecoxib and indomethacin. Considering the results of
COXs inhibition (Table 5), the anti-inflammatory activities of the tested compounds were
less than expected. These results can be due to pharmacokinetic issues.
2.2.3. Kinase inhibitory activity
2.2.3.1. Kinases profiling test
Currently, protein kinases are one of the most promising targets in development of new
anticancer agents. Based on their rational design, compounds 12, 19 and 22, the most active
as cytotoxic agents were evaluated for their inhibitory activity against 20 kinases of diverse
types and families. The profiling test was performed according to our previous report [21].
Results were summarized in Table 7.
Table 7. kinases inhibitory activity of compounds 12, 19, 22 and Imatinib.
Compounds
Kinase
12
19
22
Imatinib
14%
ALK1
-5%
4%
-11%
AMPK
(A1/B1/G1)
ASK1
Aurora A
BLK
BRAF
12%
2%
18%
-4%
26%
1%
7%
60%
-4%
6%
5%
-38%
-3%
13%
-2%
0%
-14%
-8%
-32%
-15%
-34%
-23%
-3%
-11%
26%
36%
-23%
-16%
CDK2/cyclin A1
CK1 Alpha 1
Page 19 of 55

33%
-10%
-5%
38%
10%
4%
32%
-9%
-12%
-23%
10%
5%
DYRK3
EGFR
EPHA1
FLT1
GRK1
GSK3 alpha
MSK1
NEK1
p38 Alpha
PDK1
PRKG1
SGK1
-25%
40%
25%
1%
-10%
13%
22%
-10%
2%
-1%
49%
-2%
17%
34%
4%
18%
1%
-10%
-24%
-22%
-1%
2%
-15%
11%
-3%
29%
-14%
33%
28%
3%
-11%
13%
8%
7%
3%
-14%
The negative values indicate inhibition in kinase activity, while the positive values indicate the
activation of the enzyme.
Compounds 12, 19 and 20 displayed inhibitory activity for 17 of the selected kinases with
inhibition% in the range of 1-34%. The three test compounds displayed the highest inhibitory
activity against CDK2/cyclin A1 with inhibition% in the range of 4-34%, compared to only
2% inhibition for imatinib. Moreover, they displayed inhibition of GSK3-α with inhibition%
in the range of 3-24%. Moreover, compound 12 reduced activity of NEK1 and , dual
specificity tyrosine phosphorylation-regulated kinase 3 (DYRK3) by 14 and 25%,
respectively, Table 7.
Compound 19 showed inhibitory activity against the largest number of kinases (16 kinases),
where the highest inhibitory activity was observed against CDK2 (34%) and B
lymphocyte kinase, BLK (32%), Table 7.
Moreover, compounds 19 and 22 simultaneously inhibited eight kinases including apoptosis
signal-regulating kinase 1 (ASK1), Aurora A, CDK2, CK1 Alpha 1, GRK1, GSK3 alpha,
MSK1 and PDK1 with inhibition% in the range of 3-34%. They reduced activity of PDK1
and CK1 Alpha 1 kinases by 11-23%. Compound 19 also exhibited 14% inhibition of the
activity of SGK1. The overexpression of PDK1, CK1 Alpha 1 and SGK1 in many types of
human tumors [49-51] make them valuable targets in cancer therapy. Accordingly, even
small inhibition may contribute partially to the mechanism of action of compounds.
In general, the results in Table 7 revealed an improvement in kinase inhibitory activity of
new compounds as compared to our previously reported pyrrolizine 2 [21]. Compound 2
exhibited inhibitory activity against six kinases (inhibition% = 7-20%), while compounds 12,
19 and 22 inhibited seventeen kinases (inhibition% = 1-34%).
Page 20 of 55

2.2.3.2. CDK2 inhibitory activity
In kinases profiling test, compounds 12, 19 and 22 displayed the highest inhibitory activity
against CDK2. Accordingly, the three compounds underwent further study to determine their
IC50 values against CDK2. In this study, kinase inhibitory activity of compounds 12, 19 and
22 was determined using Adenosine diphosphate (ADP)-Glo kinase assay kit (Promega)
according to the previous report [52]. The results revealed potent CDK2 inhibitory activity
for the three compounds, Table 8. They exhibited IC50 values in the (sub)micromolar range
against CDK2, compared to imatinib (IC50 >100 µM), where compound 19 was the most
active (IC50 = 0.58 µM).
Table 8. Inhibition of CDK2 by compounds 12, 19 and 22.
IC50 (µM)a ±SEM
Comp. No
1.27±0.066
0.58±0.029
0.63±0.031
>100
12
19
22
Imatinib
a Average of three determinations; IC50, concentration which decrease kinase activity to 50%.
2.2.4. Cell cycle analysis
To investigate the mechanism of the new compounds, cell cycle changes in MCF-7 cells treated
with the new compounds were analysed. Compounds 12, 14, 16 (representing scaffold A) and
22 (representing scaffold B) were selected for this study based on their cytotoxic activity
against MCF-7 cells. Flow cytometry was used to analyse propidium iodide (PI)-stained cells
according to our previous report [21]. The results of 24 h treatment with the test compounds
were outlined in Table 9. More than twofold increase in the number of cells in the S-phase was
observed after treatment with 1 µM of compounds 12, 14 and 16. Moreover, this effect was
maintained even at higher doses. These results indicated that the three compounds considerably
induced cell cycle arrest in the S-phase. On the other hand, compound 22 increased subG1
population, and G1-phase (Figure S114, supplementary data). It was clear from these results
that the cell growth arrest caused by compounds 12, 14 and 16 differ from that caused by the
thiazolidinone-bearing derivative 22.
Table 9. Cell cycle analysis of MCF-7 cells treated with compound 12, 14, 16 and 22.
Page 21 of 55

Cell Cycle Stage (%)
G1% S%
Comp. Conc.
No
(µM)
Sub G1%
0.6±0.2
1.0±0.1
0.9±0.2
0.7±0.2
0.9±0.1
0.4±0.1
0.4±0.1
0.8±0.2
0.7±0.2
1.1±0.4
0.4±0.1
0.4±0.1
3.0±0.8
2.4±0.3
2.4±0.2
8.9±3.2
G2/M%
0 µM
1µM
5µM
10µM
0 µM
1µM
5µM
10µM
0 µM
1µM
5µM
10µM
0 µM
1µM
5µM
10µM
54.6±5.5 17.3±2.5 26.8±4.1
30.8±5.1 55.6±4.1 12.4±1.4
30.2±5.3 55.1±3.9 14.4±0.9
29.6±3.9 55.2±3.4 15.1±2.6
55.1±4.7 18.2±2.2 24.9±3.6
31.7±6.5 54.1±1.1 14.2±4.5
28.9±0.4 49.7±3.8 22.0±2.8
27.9±3.1 51.9±3.3 19.8±1.1
53.9±7.3 17.8±4.1 25.6±2.1
24.5±0.7 53.7±3.5 22.1±0.9
25.5±1.9 53.6±2.9 21.8±0.7
25±0.4 51.5±2.1 23.0±1.6
49.6±2.6 28.5±1.6 17.2±1.5
56.4±1.6 20.6±1.5 15.6±2.6
59.4±3.5 17.8±1.6 15.8±1.4
59.1±1.9 14.2±2.4 14.1±0.3
12
14
16
22
Data shown in mean % ± SD (n = 2), treatment for 24 h, experiment was repeated 3x.
2.2.5. Apoptosis assay
The programed cell death mechanism (apoptosis) was reported in several cancer cell lines
following treatment with COXs inhibitors [53]. Moreover, the anticancer activity of the
diarylpyrrolizines 1 and 2 was mediated by induction of apoptosis in colon and breast cancer
cells, respectively [20-22]. Accordingly, compounds 12, 14, 16 and 22 were evaluated for
their possible apoptosis inducing activities. Annexin V fluorescein isothiocyanate
(FITC)/Propidium Iodide (PI) staining assay was used to determine early and late apoptosis.
The assay was performed by treating MCF-7 cells with test compounds according to our
previous report [43]. The cells treated with compound 12 (1 µM) showed about twofold
increase in apoptotic events (C2: late apoptosis and C4: early apoptosis) compared to control;
which increased to more than twofold at 5 and 10 µM. Each of compounds 14 and 16 induced
twofold more apoptosis at 1 µM compared to control, while the apoptotic events decreased to
onefold at 5 and 10 µM, Fig. 8.
Page 22 of 55

Fig. 8. Annexin V phases of MCF-7 treated with compound 12, 14 and 16 at 0, 1, 5 and 10 µM (24 h,
x axis: Annexin V; y axis: PI). C1: necrotic; C2: late apoptosis; C3: live cells; C4: early apoptosis.
Data shown is mean % cell number ± SD (n= 3). Experiment was repeated 3 times.
The results of Annexin V-FITC/PI staining assay for compound 22 (Figure S115,
supplementary data). The results revealed significant dose dependent increase in apoptotic
events (0%, 19%, 21% and 30% at 0, 1, 5 and 10 µM, respectively). The results were
presented
2.3. Molecular docking study
The new compounds displayed weak to strong inhibitory activity against COXs/protein
kinases in vitro. Docking studies usually provide valuable information about the interactions
between ligand and relevant protein which can be used in future optimization of these
ligands. In this work, molecular docking studies were performed to explore different types of
interaction between the new compounds and COXs/protein kinases. Crystal structure of
COXs
and
protein
kinases
were
obtained
from
protein
data
bank
(http://www.rcsb.org/pdb/home/home.do). The study was done using AutoDock 4.2 [54].
Visualization of the binding interactions of the new compounds in the active sites of these
Page 23 of 55

proteins was performed by Discovery studio visualizer [55]. The binding free energies and
inhibition constants of the new compounds into these enzymes were also evaluated.
2.3.1. Docking study into COX-1 enzyme
In this study, COX-1 crystal structure complexed with ibuprofen (pdb code: 1EQG) [56] was
used according to our previous report [57,58]. To validate docking procedures, the native
ligand (ibuprofen) was re-docked in COX-1. The results revealed superimposition of the
redocked ibuprofen above the co-crystallized ligand with RMSD of 0.87 Å.
The new compounds displayed binding free energy in the range of -7.78 to -9.21 kcal/mol for
COX-1, which was higher than licofelone 1 (∆Gb = -6.83 kcal/mol). Compounds 11-13, 15-
17 and 29 exhibited higher affinity to COX-1 than ibuprofen (∆Gb = -8.43 kcal/mol for).
Moreover, the new compounds displayed inhibition constant in the range of 0.177-1.97 µM
compared to 0.664 µM for ibuprofen. Compounds 11 and 28 did not show any conventional
hydrogen bonding with the COX-1. However, the remaining compounds displayed up to 5
hydrogen bonds with amino acids in COX-1, compared to three hydrogen bonds for
ibuprofen. Like ibuprofen, compounds 15 and 16 displayed three hydrogen bonds with the
same amino acids (ARG120 and TYR355) in COX-1, Table 10.
Among the twelve Schiff bases, compound 16 (∆Gb = -9.21 kcal/mol) displayed the highest
binding free energy to COX-1, while compound 29 (∆Gb = -8.74 kcal/mol) showed the best
affinity for COX-1 among the four thiazolidinone-bearing derivatives 22-24 and 29.
Compounds 22 and 29 displayed slightly higher binding affinity for COX-1 than their
corresponding analogs 19 and 28, Table 10.
Table 10. Results of the docking study of compounds 11-24, 28, 29, licofelone 1 and
ibuprofen into COX-1 (pdb: 1EQG) [56].
Lengthd
Atoms in H-bonding
a
Ligand
Ki b
HBsc
∆
Gb
(Å)
In ligand
-
Amino acid
-
e
-8.62
-8.94
-8.67
0.484 µM
0.278 µM
0.438 µM
-
-
11
12
13
1
2
7-CN
OH of SER530
NH of ARG120
OH of TYR355
NH2 of ARG120
OH of TYR385
NH2 of ARG120
OH of TYR355
2.55
2.28
2.57
2.28
2.32
2.21
2.43
CONH
CONH
CONH
CONH
CONH
CONH
-8.33
-9.11
0.777 µM
0.211 µM
2
3
14
15
Page 24 of 55

CONH
NH of ARG120
3.03
1.68
2.13
2.90
2.13
2.39
2.12
2.29
2.73
2.63
2.33
1.83
2.23
2.53
2.85
2.94
2.37
1.92
1.95
2.06
2.94
1.84
2.62
2.09
2.12
2.96
-
-9.21
0.177 µM
3
Furanyl O NH2 of ARG120
Furanyl O OH of TYR355
16
N=CH
CONH
CONH
CONH
CONH
CN
OH of TYR355
OH of TYR385
NH2 of ARG120
NH2 of ARG120
OH of TYR355
NH2 of ARG83
NH of ARG83
NH2 of ARG120
NH2 of ARG120
OH of TYR355
OH of SER530
NH of ARG83
NH2 of ARG83
NH2 of ARG120
NH2 of ARG120
OH of TYR355
NH2 of ARG83
NH of ARG83
NH2 of ARG83
NH of ARG83
-8.74
-7.94
-7.78
0.389 µM
1.520 µM
1.970 µM
2
2
5
17
18
19
CN
N=CH
CONH
CONH
CN
CN
CN
N=CH
CONH
CONH
CN
CN
CN
-8.33
-8.26
0.782 µM
0.879 µM
1
5
20
21
-8.04
-8.11
-8.14
1.270 µM
1.140 µM
1.070 µM
2
2
2
22
23
24
CN
PhNHCO NH2 of ARG83
PhNHCO OH of TYR355
S
-
NH of ARG83
-
e
-8.26
-8.74
0.877 µM
0.389 µM
-
28
29
2
PhNHCO NH2 of ARG83
1.89
2.29
2.49
1.71
1.79
1.83
PhNH
COOH
C=O
CO of GLU524
NH2 of ARG83
NH of ARG120
NH2 of ARG120
OH of TYR355
-6.83
-8.43
9.870 µM
0.664 µM
1
3
1
Ibu.
COOH
COOH
a Binding free energy (kcal/mol); b Inhibition constant (µM); cHBs, number of hydrogen bonds; d length
in angstrom (Å); e No hydrogen bond detected
2.3.2. Docking study into COX-2 enzyme
The binding modes of the new compounds into COX-2 (pdb code: 1CX2) [59] were evaluated
in a docking study using AutoDock 4.2, according to the previous report [57,58]. SC-558 was
initially re-docked into the active site of COX-2 to validate docking procedures. The result
revealed superimposition of the redocked ligand above native SC-558 with RMSD of 1.56 Å.
The new compounds displayed high binding affinity to COX-2 (∆Gb = -8.92 to -11.86
kcal/mol), Table 11.
Page 25 of 55

Table 11. Docking results of compounds 11-24, 28, 29, licofelone 1 and SC-588 into COX-2 (pdb:
1CX2)[59].
Lengthd
Atoms in H-bonding
a
Ligand
Ki b
HBsc
∆
Gb
(Å)
In ligand
CONH
CONH
CONH
-
In protein
-9.45
-10.17
119.07 nM
35.34 nM
1
2
OH of TYR355
OH of TYR355
NH2 of ARG120
-
1.81
1.76
3.38
-
11
12
e
-11.86
-11.47
2.03 nM
3.94 nM
-
13
14
2
2
2
2
3
7-CN
OH of SER530
NH2 of ARG513
OH of SER530
NH2 of ARG120
NH of ALA527
NH2 of ARG513
OH of SER530
NH2 of ARG513
OH of SER530
NH of GLY526
NH2 of ARG513
CO of VAL523
OH of TYR355
-
3.55
3.82
3.49
2.95
3.23
3.79
3.41
3.91
3.52
3.61
3.84
1.93
2.59
-
CONH
7-CN
CONH
Furanyl O
CONH
7-CN
-10.93
-10.63
-11.23
-11.33
9.7 nM
15
16
17
18
16.12 nM
5.88 nM
4.96 nM
CONH
7-CN
Pyridinyl N
CONH
CONH
6-N=CH
-
-9.2
179.06 nM
2
19
e
-8.92
-9.14
-9.74
-9.92
-10.05
-9.13
-10.14
289.37 nM
198.07 nM
72.18 nM
53.54 nM
43.29 nM
203.67 nM
36.98 nM
-
20
21
22
23
24
28
29
e
-
-
-
-
1
1
1
1
3
CONH
CN
CO of LEU352
OH of SER530
OH of SER530
CO of LEU352
CO of VAL523
NH of ARG120
OH of TYR355
-
2.53
2.65
2.72
2.16
2.11
2.06
2.57
CN
Ph CONH
Ph CONH
CN
CN
-
e
-8.75
-10.74
383.87 nM
13.38 nM
-
-
1
4
O of SO2
H1 of NH2
H2 of NH2
CF3
NH of HIS90
CO of LEU352
CO of GLN192
NH of ARG120
2.03
2.11
2.09
2.46
SC-588
a Binding free energy (kcal/mol); b Inhibition constant (nM); cHBs, number of hydrogen bonds; d length
in angstrom (Å); e No hydrogen bond detected.
Moreover, these compounds exhibited inhibition constant in the range of 2.03-289.37 nM
against COX-2 compared to 13.38 nM for SC-558, Table 11. Compound 13 displayed the
highest binding affinity to COX-2, while compound 20 showed the lowest affinity.
Compounds 13, 20 and 21 did not show any conventional hydrogen bonds with COX-2
although the remaining compounds displayed 1-3 hydrogen bonds.
Compounds 13-18 with the (hetero)aromatic rings in the side chain at C6 displayed higher
binding affinity than those bearing aliphatic side chain (11 and 12). In addition, the 4-
Page 26 of 55

thiazolidinone-bearing derivatives 22-24 and 29 showed also higher affinity for COX-2
than their corresponding Schiff bases 19-21 and 28, Table 11.
To visualize different types of bonding interactions between the new compounds and COX-
1/2, compounds 13 and 24 were selected based on their high affinities to COX-2.
Compound 13 exhibited inhibition constants (Ki) of 438 and 2.03 nM for COX-1 and COX-
2, respectively, indicating high selectivity for COX-2, Table 10 and 11. These results were
in accordance with the results of COXs inhibition assay, Table 5.
Several interactions of hydrophilic and hydrophobic types were observed between
compound 13 and COX-1. Among these interactions, two conventional hydrogen bonds
were detected with ARG120 and TYR355. In addition, compound 13 formed up to fourteen
hydrophobic interactions with the hydrophobic residues in the active site of COX-1, Fig. 9.
No conventional hydrogen bonds were observed between compound 13 and COX-2.
Instead, other types of interactions including carbon hydrogen bond, halogen, pi-cation, pi-
sulfur and amide-pi stacked interactions were observed. Compound 13 displayed also pi-
sigma, alkyl and pi-alkyl hydrophobic interactions with the hydrophobic residues in COX-
2, (Figure S116, supplementary data).
Moreover, compound 24 displayed inhibition constants of 1070 and 43.29 nM for COX-1 and
COX-2, respectively, indicating high potential selectivity for COX-2. These findings are
matched with the results of in vitro COXs inhibition assay, where compound 24 showed SI
>86 for COX-2, Table 5. Investigation of the binding mode of compound 24 revealed two
conventional hydrogen bonds with COX-1 and only one conventional hydrogen bond with
COX-2, Fig. 9. The higher affinity to COX-2 could be attributed to the presence of pi-cation
interaction with ARG513, and two pi-sulfur interactions with TYR385 and TRP387 amino
acids, (Figure S117, supplementary data).
Page 27 of 55

Fig. 9. Binding modes/interactions of compound 13 and 24 (shown as stick, colored by element) into
the active site of COXs: A) 3D binding mode of compound 13 into COX-1 ; B) 3D binding mode of
compound 13 into COX-2; C) 3D binding mode of compound 24 into COX-1 ; D) 3D binding mode
of compound 24 into COX-2; hydrogen bonds represented by green dotted line; receptor surface
shown as H-bond donor (red)/acceptor (green); hydrogen atoms were omitted for clarity.
Interestingly, the propylideneamino-bearing derivatives (11, 19-21 and 28) displayed high
binding affinities to COX-1/2, although no hydrogen bonding interaction was observed. That
was due to formation of other types of interaction such as carbon-hydrogen and hydrophobic
interactions. Among these derivatives, compound 11 exhibited two carbon hydrogen bonds
with COX-1, while compound 28 showed only one bond. The two compounds displayed
different types of hydrophobic interactions including pi-sigma, alkyl and pi-alkyl types,
(Figure S118, supplementary data).
Similarly, compounds 20 and 21 showed good affinity to COX-2 without forming any conventional
hydrogen bonds with this enzyme. Compound 20 displayed two carbon hydrogen bonds with
VAL523 and SER530, while compound 21 formed only one carbon hydrogen bond with VAL523.
The two compounds also displayed different types of hydrophobic interactions, (Figure S118,
supplementary data).
Page 28 of 55

In general, all the new compounds displayed higher affinity for COX-2 than COX-1. These
results are matched with the results of the in vitro COX inhibition assay. The conventional
hydrogen bonds formed between the new compounds and COX-1/2 enzymes contributed to
their overall binding affinity. However, the hydrogen bonds formed with COX-2 are longer and
weaker than those formed with COX-1, Tables 10 and 11. These results indicated that the new
compounds interact mainly by hydrophobic interactions with COX-1/2. These interactions
contribute significantly to the overall bonding affinity of these compounds.
2.3.3. Docking study into protein kinases
Targeting oncogenic protein kinases with inhibitors has proven success in treatment of
different types of cancers [60]. Understanding the nature of the interactions between these
inhibitors provided a useful tool in the design new of potent kinase inhibitors [43]. In this
work, a comparative docking study of compounds 12, 19 and 22 into PKs was performed by
AutoDock 4.2. Four kinases including CDK2, Aurora A, BRAF, and EGFR were selected for
this study based on the results of kinase inhibition assay, Tables 7 and 8. The crystal
structure of CDK2 (pdb code: 3TNW) [61], Aurora A (pdb code: 3E5A) [62], BRAF (pdb
code: 4RZV) [63] and EGFR (pdb code: 1M17) [64] were obtained from protein data bank.
The native ligands were redocked into the active sites of their corresponding PKs to validate
docking procedures. The results of the study were summarized in Table 12.
Table 12. Docking results of compounds 12, 19, 22 into CDK2, Aurora A, BRAF and EGFR
kinases in comparison to their native ligands.
PK
(pdb)
CDK2
(3TNW)
PK
(pdb)
BRAF
b
b
a
a
Ligand
12
Ki
Ligand
Ki
∆Gb
∆Gb
-8.04 1.290 µM
-9.29
-8.69
-9.18
154.4 nM
424.84 nM
186.6 nM
12
19
22
-7.63 2.550 µM (4RZV)
-8.35 0.754 µM
19
22
CAN508 c
Vemurafenib
-6.47 18.20 µM
-12.77 432.93 nM
-8.81 346.04 nM
-7.48 3.27 µM
-8.32 801.79 nM
-8.49 602.62 nM
-8.51
-7.31
-8.12
-7.39
577.35 nM
4.37 µM
1.12 µM
3.84 µM
Aurora A 12
EGFR
(1M17)
12
19
22
Erlotinib
(3E5A)
19
22
VX6 d
aBinding free energy (kcal/mol).
bInhibition constant.
cCAN508, (E)-4-((3,5-diamino-1H-pyrazol-4-yl)diazenyl)phenol.
dVX6, N-(4-((4-((3-methyl-1H-pyrazol-5-yl)amino)-6-(4-methylpiperazin-1-yl)pyrimidin-2-yl)thio)phenyl)cyclopropanecarboxamide.
Page 29 of 55